Discussion:
Apple, Windows, Linux
(too old to reply)
Alexander Ausserstorfer
2009-10-31 07:22:09 UTC
Permalink
Working with Apple, Windows and Linux at the same time I made these
experiences:

1. ) Apple seems to be the most stable system of the three above. I
don't like the keyboard. The mouse is too small and the screen too
big. The desktop is confusing because you will find no menu bar in the
programm windows. The link isn't logical. Also, Apple computers aren't
always responsible like Windows or Linux. The system is very close to
the user as the other two are.

2.) Windows seems to be more robust as Linux but not so stable as
Apple's OS (I tried different systems in both worlds).

3.) Linux: With KDE you will get almost the same desktop as with
Windows. Programs often crash down (also Internet). Trying Open Office
I need four times to get the work done because of crashs but the
fourth resulted in a corrupt document file and all the work and time
were lost.

What is so good on these systems? I will keep both Acorn computers and
invest further in them so long as possible.

Alex'
Ben Shimmin
2009-10-31 12:21:32 UTC
Permalink
Alexander Ausserstorfer <***@chiemgau-net.de>:
> Working with Apple, Windows and Linux at the same time I made these
> experiences:
>
> 1. ) Apple seems to be the most stable system of the three above. I
> don't like the keyboard. The mouse is too small and the screen too
> big. The desktop is confusing because you will find no menu bar in the
> programm windows. The link isn't logical. Also, Apple computers aren't
> always responsible like Windows or Linux. The system is very close to
> the user as the other two are.

To address some of your concerns:

1) Get a different keyboard.

2) Get a bigger mouse.

3) Get a smaller screen.

4) The menu bar is in the same place all the time, and changes depending
on the application currently in focus.

5) Computers are usually fairly responsible, it's the users you have to
worry about.

6) I'm not sure if your last sentence is positive or negative, but,
assuming the latter, try moving the computer further back on the
desk, or siting further away.

(Points 1, 2, 3, 5, and 6 apply to Windows and Linux too.)

Hope this helps!

b.

--
<***@bas.me.uk> <URL:http://bas.me.uk/>
`...the Plain English Campaign [...] says some officials only use Latin to
make themselves feel important. A Campaign spokesman said the ban might
stop people confusing the Latin abbreviation e.g. with the word "egg".'
Tim Hill
2009-10-31 13:37:55 UTC
Permalink
In article <***@rialto.bas.me.uk>,
Ben Shimmin <***@llamaselector.com> wrote:
> Alexander Ausserstorfer <***@chiemgau-net.de>:
> > Working with Apple, Windows and Linux at the same time I made these
> > experiences:
> >
> > 1. ) Apple seems to be the most stable system of the three above. I
> > don't like the keyboard. The mouse is too small and the screen too
> > big. The desktop is confusing because you will find no menu bar in the
> > programm windows. The link isn't logical. Also, Apple computers aren't
> > always responsible like Windows or Linux. The system is very close to
> > the user as the other two are.

> To address some of your concerns:

> 1) Get a different keyboard.

He who has all even advertises a left-handed model. Under my fingers the
rubbery keys on modern Macs remind me of the Spectrum but feel more like
the Cambridge Z88 on which I used to enjoy typing stuff.

> 2) Get a bigger mouse.

In my hands an Apple mouse feels more like a vole too.

Get optical, of course. Potential Apple buyers shouldn't be put off by
our local dealer who tries to demonstrate Apple opticals on a white
melamine surface. I suggested that an investment in a better surface will
sell more Macs!

> 3) Get a smaller screen.

Someone bought the wrong one. In my experience screens can't be too big.
Or is this a reference to the huge border at the bottom of the iMac
screens to accommodate the hardware ?

> 4) The menu bar is in the same place all the time, and changes depending
> on the application currently in focus.

You just get used to it, just like the differences in window furniture.

> 5) Computers are usually fairly responsible, it's the users you have to
> worry about.

I didn't understand the original assertion. In my eyes, all manufacturers
of gadgets and computers are all completely irresponsible; using our
scarce resources to make stuff which chews more energy. We're
irresponsible for fuelling that industrybbc!!

> 6) I'm not sure if your last sentence is positive or negative, but,
> assuming the latter, try moving the computer further back on the
> desk, or siting further away.

This applies to point three too.

> (Points 1, 2, 3, 5, and 6 apply to Windows and Linux too.)

Not limited to W.. or L..

--
Have an ethical share in cheaper telecoms: http://tinyurl.com/phone-coop
Genuine spam-proof addresses for Usenet: http://www.invalid.org.uk/
Email address for replies: substitute postmaster@ for tim@

... "All that lives must die, passing through nature to eternity" Hamlet, Act i, Sc.2
Tim Hill
2009-10-31 13:50:43 UTC
Permalink
In article <***@invalid.org.uk>,
Tim Hill <***@invalid.org.uk> wrote:
> irresponsible for fuelling that industrybbc!!
^^^
Typo delete xxx
Ben Shimmin
2009-10-31 14:22:41 UTC
Permalink
Tim Hill <***@invalid.org.uk>:
> In article <***@rialto.bas.me.uk>,
> Ben Shimmin <***@llamaselector.com> wrote:
>> Alexander Ausserstorfer <***@chiemgau-net.de>:
>> > 1. ) Apple seems to be the most stable system of the three above. I
>> > don't like the keyboard. The mouse is too small and the screen too
>> > big. The desktop is confusing because you will find no menu bar in the
>> > programm windows. The link isn't logical. Also, Apple computers aren't
>> > always responsible like Windows or Linux. The system is very close to
>> > the user as the other two are.
>
>> To address some of your concerns:
>
>> 1) Get a different keyboard.
>
> He who has all even advertises a left-handed model. Under my fingers the
> rubbery keys on modern Macs remind me of the Spectrum but feel more like
> the Cambridge Z88 on which I used to enjoy typing stuff.

My aluminium Apple keyboard (with the `chiclet' keys, which are indeed
visually very reminiscent of the rubber Spectrum ones) is just about
the nicest I've ever used, I think. The keyboard on my MacBook Pro is
very good too, as laptop keyboards go.

>> 2) Get a bigger mouse.
>
> In my hands an Apple mouse feels more like a vole too.
>
> Get optical, of course. Potential Apple buyers shouldn't be put off by
> our local dealer who tries to demonstrate Apple opticals on a white
> melamine surface. I suggested that an investment in a better surface will
> sell more Macs!

Apple's mice have always been... controversial. The new Magic Mouse,
for example, is actually less functional than the previous Mighty Mouse
(new called `Apple Mouse', for legal reasons). But it looks pretty!

There is, of course, nothing to stop you using your own peripherals.

>> 3) Get a smaller screen.
>
> Someone bought the wrong one. In my experience screens can't be too big.
> Or is this a reference to the huge border at the bottom of the iMac
> screens to accommodate the hardware ?

I have occasionally thought that, in pictures, the iMacs do look like
they have a rather over-prominent jaw; however, in the flesh, they look
great. The new 27" one does actually look proportionally better, I
think, even in the pictures. But it is enormous.

b.

--
<***@bas.me.uk> <URL:http://bas.me.uk/>
`...the Plain English Campaign [...] says some officials only use Latin to
make themselves feel important. A Campaign spokesman said the ban might
stop people confusing the Latin abbreviation e.g. with the word "egg".'
Alexander Ausserstorfer
2009-11-07 04:52:44 UTC
Permalink
In der Nachricht
<***@rialto.bas.me.uk>
Ben Shimmin <***@llamaselector.com>
hat geschrieben:

[snip]

> Apple's mice have always been... controversial. The new Magic Mouse,
> for example, is actually less functional than the previous Mighty Mouse
> (new called `Apple Mouse', for legal reasons). But it looks pretty!
>
> There is, of course, nothing to stop you using your own peripherals.

Less income?

[snip]

Alex'

--
E-Mail: bavariasound [ at ] chiemgau-net [ point] de
PGP (key): http://home.chiemgau-net.de/ausserstorfer/kontakt/key
Web: http://home.chiemgau-net.de/ausserstorfer/
Chris Evans
2009-11-02 13:35:47 UTC
Permalink
In article <***@invalid.org.uk>, Tim Hill
<URL:mailto:***@invalid.org.uk> wrote:
> In article <***@rialto.bas.me.uk>,
> Ben Shimmin <***@llamaselector.com> wrote:
> > Alexander Ausserstorfer <***@chiemgau-net.de>:

> > 5) Computers are usually fairly responsible, it's the users you have to
> > worry about.
>
> I didn't understand the original assertion. In my eyes, all manufacturers
> of gadgets and computers are all completely irresponsible; using our
> scarce resources to make stuff which chews more energy. We're
> irresponsible for fuelling that industrybbc!!

I wonder if he meant responsive?

Chris Evans

--
CJE Micro's / 4D 'RISC OS Specialists'
Telephone: 01903 523222 Fax: 01903 523679
***@cjemicros.co.uk http://www.cjemicros.co.uk/
78 Brighton Road, Worthing, West Sussex, BN11 2EN
The most beautiful thing anyone can wear, is a smile!
Alexander Ausserstorfer
2009-11-07 04:50:44 UTC
Permalink
In der Nachricht
<***@invalid.org.uk>
Tim Hill <***@invalid.org.uk>
hat geschrieben:

> In article <***@rialto.bas.me.uk>,
> Ben Shimmin <***@llamaselector.com> wrote:
> > Alexander Ausserstorfer <***@chiemgau-net.de>:
> > > Working with Apple, Windows and Linux at the same time I made these
> > > experiences:
> > >
> > > 1. ) Apple seems to be the most stable system of the three above. I
> > > don't like the keyboard. The mouse is too small and the screen too
> > > big. The desktop is confusing because you will find no menu bar in the
> > > programm windows. The link isn't logical. Also, Apple computers aren't
> > > always responsible like Windows or Linux. The system is very close to
> > > the user as the other two are.
>
> > To address some of your concerns:
>
> > 1) Get a different keyboard.
>
> He who has all even advertises a left-handed model. Under my fingers the
> rubbery keys on modern Macs remind me of the Spectrum but feel more like
> the Cambridge Z88 on which I used to enjoy typing stuff.
>
> > 2) Get a bigger mouse.
>
> In my hands an Apple mouse feels more like a vole too.
>
> Get optical, of course. Potential Apple buyers shouldn't be put off by
> our local dealer who tries to demonstrate Apple opticals on a white
> melamine surface. I suggested that an investment in a better surface will
> sell more Macs!

Here by us, Apple computers are three times more expensive as other
current machines. I'm not an Apple owner. The machines I'm working with,
are the ones of our organisation. For the money I have to give for a new
Apple computer, I can surely demand an excellent keyboard which comes
with the machine.

I often tried different keyboards, also on my Acorn computers. I stick
to the original keyboard of Acorn computers since an age of fourteen and
since I learnt to type with ten fingers. All other were garbage.

> > 3) Get a smaller screen.
>
> Someone bought the wrong one. In my experience screens can't be too big.
> Or is this a reference to the huge border at the bottom of the iMac
> screens to accommodate the hardware ?

I meant the resolution of the desktop is too large. I need too much time
to move the mouse around. There are also too much information on screen.
It overexcites my senses. 800 x 600 would be good.

> > 4) The menu bar is in the same place all the time, and changes depending
> > on the application currently in focus.
>
> You just get used to it, just like the differences in window furniture.

You will lose a lot of time using this technique. Here, RISC OS is more
clever. The menus are appearing where you are already with the mouse.

> > 5) Computers are usually fairly responsible, it's the users you have to
> > worry about.
>
> I didn't understand the original assertion. In my eyes, all manufacturers
> of gadgets and computers are all completely irresponsible; using our
> scarce resources to make stuff which chews more energy. We're
> irresponsible for fuelling that industrybbc!!

Sorry, I meant the machine doesn't listen to the user all the time. It
may be that the reaction to the users action will have a delay.

> > 6) I'm not sure if your last sentence is positive or negative, but,
> > assuming the latter, try moving the computer further back on the
> > desk, or siting further away.

Misinterpretation but that's okey. My intention was that RISC OS is much
more open and flexible (for the user) as MacOS. That means: no log-in.
The user can examine and chance almost every bit in the system without
any problem or obstacle. I cannot find the often praised freedom by
Americans in their computers.

> This applies to point three too.
>
> > (Points 1, 2, 3, 5, and 6 apply to Windows and Linux too.)
>
> Not limited to W.. or L..


--
E-Mail: bavariasound [ at ] chiemgau-net [ point] de
PGP (key): http://home.chiemgau-net.de/ausserstorfer/kontakt/key
Web: http://home.chiemgau-net.de/ausserstorfer/
Ollie Clark
2009-11-09 10:40:38 UTC
Permalink
Alexander Ausserstorfer wrote:
>
> I meant the resolution of the desktop is too large. I need too much time
> to move the mouse around. There are also too much information on screen.
> It overexcites my senses. 800 x 600 would be good.

I'm pretty sure that like all other operating systems, Mac OS lets you
change the resolution. Problem solved. I can't imagine what sort of hell
it would be working at 800x600 these days but each to their own.

Cheers,

Ollie
Ben Shimmin
2009-11-09 13:09:30 UTC
Permalink
Ollie Clark <***@ollieclark.com>:
> Alexander Ausserstorfer wrote:
>> I meant the resolution of the desktop is too large. I need too much time
>> to move the mouse around. There are also too much information on screen.
>> It overexcites my senses. 800 x 600 would be good.
>
> I'm pretty sure that like all other operating systems, Mac OS lets you
> change the resolution. Problem solved. I can't imagine what sort of hell
> it would be working at 800x600 these days but each to their own.

I can confirm that you certainly can set a Mac's resolution to 800x600,
or even lower than that; however, I would say that Mac OS X isn't
particularly well-suited to running at that sort of resolution, not
helped by the fact that Apple still haven't finished resolution
independence (which would let you scale all the window furniture down
in size to give you a little more space to work in).

I can't say I've worked in anything less than 1024x768 for, oh, about
fifteen years, but there we go (currently 1680x1050 + 1440x900 at home,
1280x1024 + 1280x1024 + 1440x900 at work).

b.

--
<***@bas.me.uk> <URL:http://bas.me.uk/>
`...the Plain English Campaign [...] says some officials only use Latin to
make themselves feel important. A Campaign spokesman said the ban might
stop people confusing the Latin abbreviation e.g. with the word "egg".'
Mr John FO Evans
2009-11-10 10:17:51 UTC
Permalink
In article <***@rialto.bas.me.uk>, Ben Shimmin
<***@llamaselector.com> wrote:
> > I'm pretty sure that like all other operating systems, Mac OS lets you
> > change the resolution. Problem solved. I can't imagine what sort of hell
> > it would be working at 800x600 these days but each to their own.

Unless of course you are working under RISC OS which uses space efficiently
for most applications and doesn't need enormous resolution to see even small
text due to its anti-aliasing. It even beats windows at speed of
presentation and start up except for large graphics. (200MHz 64Mbt ARM v
1600Mhz 1Gbt PC).

John


--
_ _________________________________________
/ \._._ |_ _ _ /' Orpheus Internet Services
\_/| |_)| |(/_|_|_> / 'Internet for Everyone'
_______ | ___________./ http://www.orpheusinternet.co.uk
Rob Kendrick
2009-11-10 11:02:33 UTC
Permalink
On Tue, 10 Nov 2009 10:17:51 GMT
Mr John FO Evans <***@orpheusmail.co.uk> wrote:

> In article <***@rialto.bas.me.uk>, Ben Shimmin
> <***@llamaselector.com> wrote:
> > > I'm pretty sure that like all other operating systems, Mac OS
> > > lets you change the resolution. Problem solved. I can't imagine
> > > what sort of hell it would be working at 800x600 these days but
> > > each to their own.
>
> Unless of course you are working under RISC OS which uses space
> efficiently for most applications and doesn't need enormous
> resolution to see even small text due to its anti-aliasing. It even
> beats windows at speed of presentation and start up except for large
> graphics. (200MHz 64Mbt ARM v 1600Mhz 1Gbt PC).

I find RISC OS unusable at anything less than 1024x768. Window
management becomes a pain due to them all obscuring each other and
having to dig through them to find the right one.

Additionally, I find the RISC OS font manager's anti-aliasing makes
things worse due to its lack of hinting to pull lines onto pixel
boundaries, meaning you can fit less text into a small space with less
clarity, due to the blurring effect. (Although this approach is fine
for DTP where you want to keep the precise metrics, for web pages and
email it is less ideal.)

I find Linux, however, completely usable at 320x480. ie, my phone,
because the GUI is designed for it :)

B.
Martin Bazley
2009-11-10 19:03:27 UTC
Permalink
The following bytes were arranged on 10 Nov 2009 by Rob Kendrick :

> I find RISC OS unusable at anything less than 1024x768. Window
> management becomes a pain due to them all obscuring each other and
> having to dig through them to find the right one.
>
I use !Switch to solve that problem -

http://www.starfighter.acornarcade.com/mysite/utilities.htm#switch

> Additionally, I find the RISC OS font manager's anti-aliasing makes
> things worse due to its lack of hinting to pull lines onto pixel
> boundaries, meaning you can fit less text into a small space with less
> clarity, due to the blurring effect. (Although this approach is fine
> for DTP where you want to keep the precise metrics, for web pages and
> email it is less ideal.)
>
I've never had any such problem.

--
__<^>__ "Did you know that polar bears stay white all year round? ...The
/ _ _ \ white colour makes them less visible to the seals and penguins they
( ( |_| ) ) hunt." - Nelson Thornes AQA-endorsed GCSE science textbook
\_> <_/ ======================= Martin Bazley ==========================
Jeremy Nicoll - news posts
2009-11-10 19:20:00 UTC
Permalink
Martin Bazley <***@blueyonder.co.uk> wrote:

> The following bytes were arranged on 10 Nov 2009 by Rob Kendrick :
>
> > I find RISC OS unusable at anything less than 1024x768. Window
> > management becomes a pain due to them all obscuring each other and
> > having to dig through them to find the right one.
> >
> I use !Switch to solve that problem -
>
> http://www.starfighter.acornarcade.com/mysite/utilities.htm#switch

I use VRPC on a laptop, using a little less than the full screen (because if
it is using full-screen alerts from anti-virus and firewall software on XP
can go unnoticed, and it's often useful to have a tiny amount of another
windows app's window visible at the same time). RO therefore runs in a
1200x700 window.

I find !MoreDesk invaluable for this, not least because I've configured it
so that different combinations of windows are present in each virtual
desktop. Some will follow me around as I change from desktop to desktop (eg
the !Reporter log window which I always have open, though small, on each
screen). Others only ever open on a particular desktop. So eg for working
with !AntiSpam and its logs, I use one specific desktop only.

--
Jeremy C B Nicoll - my opinions are my own.

Email sent to my from-address will be deleted. Instead, please reply
to ***@wingsandbeaks.org.uk replacing "nnn" by "284".
Dave Higton
2009-11-11 21:29:12 UTC
Permalink
In message <***@trite.i.flarn.net.i.flarn.net>
Rob Kendrick <***@rjek.com> wrote:

> I find RISC OS unusable at anything less than 1024x768. Window
> management becomes a pain due to them all obscuring each other and
> having to dig through them to find the right one.

Curious. RISC OS in general wastes lass space on window furniture
and useless blank areas that most other OSs, so you can get the
same amount of information in less pixels.

> Additionally, I find the RISC OS font manager's anti-aliasing makes
> things worse due to its lack of hinting to pull lines onto pixel
> boundaries, meaning you can fit less text into a small space with less
> clarity, due to the blurring effect.

I recall we've disagreed on this topic before. I haven't used
Mac OS [1], but I have often found the effect that you describe,
on Windows, is irritating and sometimes slows me down - is that
a spurious space between those letters? Careful examination -
no, it's just Windows buggering the spacing up so it looks
unnatural.

Dave

[1] And, since Apple has taken over Microsoft's title as The
Evil Empire, I'm not likely to any time soon.
John M Ward
2009-11-11 23:41:36 UTC
Permalink
In article <***@dsl.pipex.com>,
Dave Higton <***@dsl.pipex.com> wrote:
> In message <***@trite.i.flarn.net.i.flarn.net>
> Rob Kendrick <***@rjek.com> wrote:

> > I find RISC OS unusable at anything less than 1024x768. Window
> > management becomes a pain due to them all obscuring each other and
> > having to dig through them to find the right one.

> Curious. RISC OS in general wastes lass space on window furniture
> and useless blank areas that most other OSs, so you can get the
> same amount of information in less pixels.

True enough, I find, largely because of all the toolbars, status bars and
the rest that one finds in mainstream OSes' apps -- which seems to be the
way their Style Guide or equivalent expects things.

Also, there are sets of slimmer (and chunkier, for that matter) window
furniture freely available, and I do recommend looking at those for
lower-res RISC OS displays.

> > Additionally, I find the RISC OS font manager's anti-aliasing makes
> > things worse due to its lack of hinting to pull lines onto pixel
> > boundaries, meaning you can fit less text into a small space with less
> > clarity, due to the blurring effect.

I don't find it blurred and find it better than the other systems I have
encountered; but it's simple enough to switch off sub-pixel anti-aliasing
in Configure. At least we have that choice...

> I recall we've disagreed on this topic before. I haven't used Mac OS,
> but I have often found the effect that you describe, on Windows, is
> irritating and sometimes slows me down - is that a spurious space
> between those letters?

It isn't very accurate, and we have become accustomed to doing it properly
during the past twenty years.

When I was trying to introduce DTP into my office in the then DTI, not
only was there no anti-aliasing at all, but smaller text was "greeked"
(represented by a grey block) -- this was in top-end DTP software.

As Druck has written previously, Microsoft do not implement "font
smoothing" (as they call it) at all well. Although they licensed the
technology from Acorn, it's as if (again) they don't really understand it
themselves -- like Bill "what's a network?" Gates when first visiting
Acorn some years ago. These people have the money and the sheer brute
force to dominate, but it doesn't mean they're any good!

--
John Ward in Medway, Kent - using RISC OS since 1987
Now using an Iyonix, an A9home, 2 RiscPCs and Virtual-RPC!
Acorn/RISC OS web page: www.john-ward.org.uk/personal/john/computers
Rob Kendrick
2009-11-12 01:14:12 UTC
Permalink
On Wed, 11 Nov 2009 23:41:36 GMT
John M Ward <***@acornusers.org> wrote:

> Also, there are sets of slimmer (and chunkier, for that matter) window
> furniture freely available, and I do recommend looking at those for
> lower-res RISC OS displays.

Yes; I always use my own set that are about 33% smaller :)

> > > Additionally, I find the RISC OS font manager's anti-aliasing
> > > makes things worse due to its lack of hinting to pull lines onto
> > > pixel boundaries, meaning you can fit less text into a small
> > > space with less clarity, due to the blurring effect.
>
> I don't find it blurred and find it better than the other systems I
> have encountered; but it's simple enough to switch off sub-pixel
> anti-aliasing in Configure. At least we have that choice...

Sure, but this doesn't effect the blurryness, only the accuracy of
character positioning. Also, note that RISC OS doesn't have sub-pixel
anti-aliasing, it has sub-pixel *positioning*. (And Windows, Linux and
Mac OS let you turn off sub-pixel anti-aliasing, and positioning is
enabled on a per-app basis; ie, it's not normally on for textual
display, but it is for art, word processing, etc)

> As Druck has written previously, Microsoft do not implement "font
> smoothing" (as they call it) at all well. Although they licensed the
> technology from Acorn, it's as if (again) they don't really
> understand it themselves

Woah, FUD. Windows has had excellent font rasterising that was not the
crap positioning of grey pixels on edges since XP. And they've never
licenced anything of Acorn as far as I can tell; I'd like to see a
citeation for that. Acorn's font render anti-alaises by rendering at
16 times the normal size, and then resampling the resulting image
down. Microsoft's anti-aliasing attempts have always done it at the
size of render. Doing so is essential for sub-pixel anti-aliasing.

Additionally, Microsoft and Apple own dozens of patents on the
widely-believed "correct" way of on-screen rasterising of text. Acorn
own none, and their system is the most simple and naive imaginable.

Still, it's what you're used to. Some people like all that blurry text.

B.
Martin Bazley
2009-11-12 16:13:11 UTC
Permalink
The following bytes were arranged on 12 Nov 2009 by Rob Kendrick :

> Still, it's what you're used to. Some people like all that blurry text.
>
Sorry, we may be talking on different wavelengths here. Somebody who
uses Windows regularly is referring to the Acorn font manager's
anti-aliasing as 'blurry'?

This statement does not compute. Windows is blurry. Try entering text
in Word in a suitable font (Franklin Gothic Book works well) and then
call RISC OS 'blurry'. With a straight face.

--
__<^>__
/ _ _ \ You always find something in the last place you look.
( ( |_| ) )
\_> <_/ ======================= Martin Bazley ==========================
Rob Kendrick
2009-11-12 16:20:06 UTC
Permalink
On Thu, 12 Nov 2009 16:13:11 GMT
Martin Bazley <***@blueyonder.co.uk> wrote:

> The following bytes were arranged on 12 Nov 2009 by Rob Kendrick :
>
> > Still, it's what you're used to. Some people like all that blurry
> > text.
> >
> Sorry, we may be talking on different wavelengths here. Somebody who
> uses Windows regularly is referring to the Acorn font manager's
> anti-aliasing as 'blurry'?

Absolutely. Perhaps you're stuck using an ancient version of Windows.
Modern versions of Windows have excellent font rasterising along with
good hinting.

Where RISC OS still uses the draw-it-big-and-scale-it-down hack.

B.
John M Ward
2009-11-12 16:51:06 UTC
Permalink
In article <***@blueyonder.co.uk>,
Martin Bazley <***@blueyonder.co.uk> wrote:
> The following bytes were arranged on 12 Nov 2009 by Rob Kendrick :

> > Still, it's what you're used to. Some people like all that blurry
> > text.

> Sorry, we may be talking on different wavelengths here. Somebody who
> uses Windows regularly is referring to the Acorn font manager's
> anti-aliasing as 'blurry'?

I agree, from twenty years of experience of the system at all resolutions
from 640 x 256 to 1600 x 1200 pixels.

> This statement does not compute. Windows is blurry. Try entering text
> in Word in a suitable font (Franklin Gothic Book works well) and then
> call RISC OS 'blurry'. With a straight face.

It all depends on point size, according to something I once read here from
Druck, and I have noticed that it is very variable on the Microsoft system
from one size to another. If you get one of the "right" point sizes it
seems to be okay, but not otherwise. It isn't really, it just looks OK.

It /could/ be to do with the poor scaffolding that I have mentioned
elsewhere in this thread today. They just don't seem to understand the
concept of anything of higher (finer) resolution than an actual pixel, as
far as I can determine from inspection and experience over many years.

There might be another explanation, of course; but whatever, it just isn't
a very good system, though much improved in XP and beyond from earlier
versions, which is something at least.

--
John Ward in Medway, Kent - using RISC OS since 1987
Now using an Iyonix, an A9home, 2 RiscPCs and Virtual-RPC!
Acorn/RISC OS web page: www.john-ward.org.uk/personal/john/computers
Rob Kendrick
2009-11-12 17:01:46 UTC
Permalink
On Thu, 12 Nov 2009 16:51:06 GMT
John M Ward <***@acornusers.org> wrote:

> It /could/ be to do with the poor scaffolding that I have mentioned
> elsewhere in this thread today. They just don't seem to understand
> the concept of anything of higher (finer) resolution than an actual
> pixel, as far as I can determine from inspection and experience over
> many years.

The way Windows, Mac OS and Linux do this is *because* they understand
the scaffolding and sub pixels. (Look at that screenshot more closely
to see how it's using knowledge of the workings of the TFT to light
fractional pixels, instead of applying the hack of using a lighter
colour to simulation fractional pixels.)

And the scaffold manipulation is called hinting, and it is done to
improve the readability of the text while sacrificing absolute accuracy
of the shapes of glyphs. And this manipulation is done to *minimise*
the amount of anti-aliasing that occurs.

B.
Ben Shimmin
2009-11-12 17:09:22 UTC
Permalink
Rob Kendrick <***@rjek.com>:
> On Thu, 12 Nov 2009 16:51:06 GMT
> John M Ward <***@acornusers.org> wrote:
>> It /could/ be to do with the poor scaffolding that I have mentioned
>> elsewhere in this thread today. They just don't seem to understand
>> the concept of anything of higher (finer) resolution than an actual
>> pixel, as far as I can determine from inspection and experience over
>> many years.
>
> The way Windows, Mac OS and Linux do this is *because* they understand
> the scaffolding and sub pixels. (Look at that screenshot more closely
> to see how it's using knowledge of the workings of the TFT to light
> fractional pixels, instead of applying the hack of using a lighter
> colour to simulation fractional pixels.)
>
> And the scaffold manipulation is called hinting, and it is done to
> improve the readability of the text while sacrificing absolute accuracy
> of the shapes of glyphs. And this manipulation is done to *minimise*
> the amount of anti-aliasing that occurs.

This is an interesting read:

<URL:http://www.xvsxp.com/system/fonts_view.php>

(for some values of `interesting', anyway).

b.

--
<***@bas.me.uk> <URL:http://bas.me.uk/>
`...the Plain English Campaign [...] says some officials only use Latin to
make themselves feel important. A Campaign spokesman said the ban might
stop people confusing the Latin abbreviation e.g. with the word "egg".'
Mr John FO Evans
2009-11-12 19:34:47 UTC
Permalink
In article <***@rialto.bas.me.uk>, Ben Shimmin
<***@llamaselector.com> wrote:
> This is an interesting read:
>
> <URL:http://www.xvsxp.com/system/fonts_view.php>
>
> (for some values of `interesting', anyway).
>
> b.

Looking at the examples in the above might explain why we have so much
trouble printing non-colour pages on our Xerox printer/collator/finisher.

This printer charges about 7 times more for a colour page than a B/W page and
even one colour pixel will cause it to charge for a colour page. If we
use a Windows DTP program we have to set the printer driver to B/W (and also
ensure that it is not set to 'process black'). If we want a single colour
page in the document, and also retain the full print/collate/finish
capability of the printer, we cannot do this since the driver has to be set
for colour. We then have to go to extreme lengths to 'doctor' the B/W pages
at print time. The only program we have found so far which does not have
this fault is Ovation Pro for Windows!!

I have made a mental note to ensure that our next printer is a single
transfer drum type in which the colour premium can be far less.

John


--
_ _________________________________________
/ \._._ |_ _ _ /' Orpheus Internet Services
\_/| |_)| |(/_|_|_> / 'Internet for Everyone'
_______ | ___________./ http://www.orpheusinternet.co.uk
Rob Kendrick
2009-11-12 21:06:38 UTC
Permalink
On Thu, 12 Nov 2009 19:34:47 GMT
Mr John FO Evans <***@orpheusmail.co.uk> wrote:

> Looking at the examples in the above might explain why we have so much
> trouble printing non-colour pages on our Xerox
> printer/collator/finisher.
>
> This printer charges about 7 times more for a colour page than a B/W
> page and even one colour pixel will cause it to charge for a colour
> page.

The sub-pixel anti-aliasing only makes sense on TFTs, and as such is
only enabled when rasterising text to the screen. On a printing, no
anti-aliasing is performed at all, because it is assumed your printer
is much higher DPI than your monitor.

B.
Mr John FO Evans
2009-11-13 09:38:04 UTC
Permalink
In article <***@trite.i.flarn.net.i.flarn.net>, Rob
Kendrick <***@rjek.com> wrote:
> The sub-pixel anti-aliasing only makes sense on TFTs, and as such is
> only enabled when rasterising text to the screen. On a printing, no
> anti-aliasing is performed at all, because it is assumed your printer
> is much higher DPI than your monitor.

OK then there must be some other reason why Windows DTP programs (other than
Ovation Pro) almost always put some colour pixels in B/W pages at print
time. Great for the printer suppliers but very bad when you are on a budget
- one print bill went up by some hundreds of pounds before we discovered
what was happening.

Worse it is often very difficult a) to discover that it has happened
b)to get rid of it!

I must emphasise that the problem only really bites when you actually want
colour on selected pages and you are printing the whole document collated.

John


--
_ _________________________________________
/ \._._ |_ _ _ /' Orpheus Internet Services
\_/| |_)| |(/_|_|_> / 'Internet for Everyone'
_______ | ___________./ http://www.orpheusinternet.co.uk
druck
2009-11-12 20:31:00 UTC
Permalink
John M Ward wrote:
> It all depends on point size, according to something I once read here from
> Druck, and I have noticed that it is very variable on the Microsoft system
> from one size to another. If you get one of the "right" point sizes it
> seems to be okay, but not otherwise. It isn't really, it just looks OK.

I'm afraid that comment was from many years ago, when Windows only had
naive "font smoothing", now it has Clear Type and we are all using LCD
monitors, it is a different story.

RISC OS fonts on a very high resolution CRT looked great at the
anti-aliasing was a good match the to slightly indistinct pixels on that
technology, but on an LCD with pin sharp pixels, it just looks fuzzy.

Clear Type on Windows, and similar technology on Linux, effectively
renders the font at 3x the horizontal pixel resolution making the fonts
look a lot smoother and pleasing to the eye than on RISC OS which is
limited to whole pixels.

RISC OS still wins on accuracy as it far better at portraying font
weights and sizes to 1/16th of a point, which is good for DTP. The other
systems optimise for screen rendering with scafolding which leads to
uneven increments, but the vast majority of the time you are looking at
screen orientated content, not printing accuracy.

---druck
Rob Kendrick
2009-11-12 21:08:14 UTC
Permalink
On Thu, 12 Nov 2009 20:31:00 +0000
druck <***@druck.org.uk> wrote:

> RISC OS still wins on accuracy as it far better at portraying font
> weights and sizes to 1/16th of a point, which is good for DTP. The
> other systems optimise for screen rendering with scafolding which
> leads to uneven increments, but the vast majority of the time you are
> looking at screen orientated content, not printing accuracy.

Although any DTP or graphics package worth a jot will use the RISC OS
approach. (Or, failing that, convert the font to shapes and use the
package's own renderer, which will be designed for print.)

Certainly, Xara (both under Windows and Linux), Scribus, etc etc all do
this.

B.
John M Ward
2009-11-13 00:22:46 UTC
Permalink
In article <***@druck.org.uk>,
druck <***@druck.org.uk> wrote:
> John M Ward wrote:
> > It all depends on point size, according to something I once read here
> > from Druck, and I have noticed that it is very variable on the
> > Microsoft system from one size to another. If you get one of the
> > "right" point sizes it seems to be okay, but not otherwise. It isn't
> > really, it just looks OK.

> I'm afraid that comment was from many years ago, when Windows only had
> naive "font smoothing", now it has Clear Type and we are all using LCD
> monitors, it is a different story.

Ah, I'm pleased you wrote this, as I have been thinking about the matter
of monitor types while I have been out at a Council meeting (just got back
a few minutes ago). It seems I have a long memory!

> RISC OS fonts on a very high resolution CRT looked great at the
> anti-aliasing was a good match the to slightly indistinct pixels on
> that technology, but on an LCD with pin sharp pixels, it just looks
> fuzzy.

I wondered about that, but it doesn't seem to be so here -- and I have
both types of monitor side by side here in what I call "Mission Control".
However, on the bedroom computer, which is at a lower res on a TFT
display, I do accept that it looks a little fuzzy when viewed close-up.

> Clear Type on Windows, and similar technology on Linux, effectively
> renders the font at 3x the horizontal pixel resolution making the fonts
> look a lot smoother and pleasing to the eye than on RISC OS which is
> limited to whole pixels.

Another thing that occurred to me while I was out was that it would be
useful to have a link to what one of you experts considered to be an
authoritative web page on this Clear Type (or the Linux equivalent)
technology. I can easily Google if you have no preferences, but first
wanted to offer this opportunity to suggest a suitable webpage. I find it
an interesting subject, actually. Perhaps I'm just sad...

> RISC OS still wins on accuracy as it far better at portraying font
> weights and sizes to 1/16th of a point, which is good for DTP. The
> other systems optimise for screen rendering with scafolding which leads
> to uneven increments, but the vast majority of the time you are looking
> at screen orientated content, not printing accuracy.

Well, I still do a fair amount of document design for printing purposes,
so it's okay for me, and I certainly don't notice any blurring on what are
mostly 1280 x 1024 displays or greater. It works for me, as they say :-)

--
John Ward in Medway, Kent - using RISC OS since 1987
Now using an Iyonix, an A9home, 2 RiscPCs and Virtual-RPC!
Acorn/RISC OS web page: www.john-ward.org.uk/personal/john/computers
Tim Hill
2009-11-13 12:45:07 UTC
Permalink
In article <***@druck.org.uk>, druck <***@druck.org.uk>
wrote:
> John M Ward wrote:
> > It all depends on point size, according to something I once read here
> > from Druck, and I have noticed that it is very variable on the
> > Microsoft system from one size to another. If you get one of the
> > "right" point sizes it seems to be okay, but not otherwise. It isn't
> > really, it just looks OK.

> I'm afraid that comment was from many years ago, when Windows only had
> naive "font smoothing", now it has Clear Type and we are all using LCD
> monitors, it is a different story.

In spite of what may have been implied further up this thread, ClearType
is present in windows XP. I have both 'doze XP and RISC OS 5.15 running
on 1280x1024 LCD monitors and my eyes tell me that the RISC OS font
rendering is better throughout the desktop and one reason I still use
Pluto. Perhaps this is because 'doze doesn't have a consistent approach
and apply ClearType all the time. The amount of pixellation of fonts in
the desktop is astonishing for a 'modern OS' and is little better than a
Spectrum in places.

> RISC OS fonts on a very high resolution CRT looked great at the
> anti-aliasing was a good match the to slightly indistinct pixels on
> that technology, but on an LCD with pin sharp pixels, it just looks
> fuzzy.

Fuzzy = nicer to read! YMMV if you like reading lumpy-bumpy fonts.

> Clear Type on Windows, and similar technology on Linux, effectively
> renders the font at 3x the horizontal pixel resolution making the fonts
> look a lot smoother and pleasing to the eye than on RISC OS which is
> limited to whole pixels.

It is only possible for any rendering on any display to render at the
pixel level so I'm not sure what you're getting at. I thought RISC OS
does have so-called sub-pixel anti-aliasing or did I dream the contents
of Configure > Fonts????

> RISC OS still wins on accuracy as it far better at portraying font
> weights and sizes to 1/16th of a point, which is good for DTP. The
> other systems optimise for screen rendering with scafolding which
> leads to uneven increments, but the vast majority of the time you are
> looking at screen orientated content, not printing accuracy.

Layout looks more pleasing when the text is rendered accurately rather
than, as mentioned elsewhere, 'doze sometimes introduces an apparent
spa ce where you least expect it.

Incidentally, Safari runs on WinXP and its own font rendering looks
/much/ better than 'doze XP and more "blurry" like RISC OS.

The 'doze XP ClearType Tuner is here:
http://www.microsoft.com/typography/cleartype/tuner/step1.aspx

Note that it describes the two types as 'Black and White' and 'ClearType'
so anyone with problematic coloured edge pixels may want to resort to the
former.

--
Have an ethical share in cheaper telecoms: http://tinyurl.com/phone-coop
Genuine spam-proof addresses for Usenet: http://www.invalid.org.uk/
Email address for replies: substitute postmaster@ for tim@

... "O, beauty, till now I never knew thee!" Henry VIII, Act i, Sc.4
druck
2009-11-13 22:04:46 UTC
Permalink
Tim Hill wrote:
> In article <***@druck.org.uk>, druck <***@druck.org.uk>
> wrote:
>> Clear Type on Windows, and similar technology on Linux, effectively
>> renders the font at 3x the horizontal pixel resolution making the
>> fonts look a lot smoother and pleasing to the eye than on RISC OS
>> which is limited to whole pixels.
>
> It is only possible for any rendering on any display to render at the
> pixel level so I'm not sure what you're getting at.

No it is. Where as CRTs had several different patterns of R, & B dots or
stripes, LCDs are always patterns of vertical RGB strips. Therefore the
font rendering technology can treat the colour components as
individually addressable sub pixels in order give a greater apparent
horizontal resolution. The side effect of this is colour fringes at the
edges of characters, but this can be compensated for by altering the
colours of surrounding pixels. Obviously this works best with black &
white text and backgrounds, and becomes less and less effective as
either text or background is closer to a saturated colour.

> I thought RISC OS does have so-called sub-pixel anti-aliasing or did
> I dream the contents of Configure > Fonts????

It's called sub pixel anti-aliasing, but what it actually is sub pixel
positioning, where it caches several copies of a character at quarter
pixel horizontal (and/or vertical) offsets. This can be seen if you do a
row of iiiiiiiiiiiiiii's where the character size isn't a whole pixel.
Without sub pixel a-a there will be clumps of identical i's separated by
a blank pixel to preserve positioning accuracy. With sub pixel a-a on,
the i's will appear equally spaced, as the different sub pixel variants
are used depending on the fractional pixel offset of the character
position. In all cases the actual anti-aliasing is always plotting whole
pixels.

---druck
Tim Hill
2009-11-14 17:18:47 UTC
Permalink
In article <***@druck.org.uk>, druck <***@druck.org.uk>
wrote:
> Tim Hill wrote:
> > In article <***@druck.org.uk>, druck <***@druck.org.uk>
> > wrote:
> >> Clear Type on Windows, and similar technology on Linux, effectively
> >> renders the font at 3x the horizontal pixel resolution making the
> >> fonts look a lot smoother and pleasing to the eye than on RISC OS
> >> which is limited to whole pixels.
> >
> > It is only possible for any rendering on any display to render at the
> > pixel level so I'm not sure what you're getting at.

> No it is. Where as CRTs had several different patterns of R, & B dots
> or stripes, LCDs are always patterns of vertical RGB strips. Therefore
> the font rendering technology can treat the colour components as
> individually addressable sub pixels in order give a greater apparent
> horizontal resolution. The side effect of this is colour fringes at the
> edges of characters, but this can be compensated for by altering the
> colours of surrounding pixels. Obviously this works best with black &
> white text and backgrounds, and becomes less and less effective as
> either text or background is closer to a saturated colour.

Yes, thanks for that D. I was getting my knickers in a twist. Of course,
ClearType does render at the sub (R-G-B) pixel level I was trying not to
confuse the sub- and non-sub types and it all ended in confusion by me
wrongly referring to the RGB pixelettes as pixels. It was the "fonts look
a lot smoother and pleasing to the eye than on RISC OS" I was really
disagreeing with because I dislike the colour fringing ClearType causes.

> > I thought RISC OS does have so-called sub-pixel anti-aliasing or did
> > I dream the contents of Configure > Fonts????

> It's called sub pixel anti-aliasing, but what it actually is sub pixel
> positioning, where it caches several copies of a character at quarter
> pixel horizontal (and/or vertical) offsets. This can be seen if you do
> a row of iiiiiiiiiiiiiii's where the character size isn't a whole
> pixel. Without sub pixel a-a there will be clumps of identical i's
> separated by a blank pixel to preserve positioning accuracy. With sub
> pixel a-a on, the i's will appear equally spaced, as the different sub
> pixel variants are used depending on the fractional pixel offset of
> the character position. In all cases the actual anti-aliasing is
> always plotting whole pixels.

Exactly. Another Acorn misnomer perhaps.

--
Have an ethical share in cheaper telecoms: http://tinyurl.com/phone-coop
Genuine spam-proof addresses for Usenet: http://www.invalid.org.uk/
Email address for replies: substitute postmaster@ for tim@

... "All that lives must die, passing through nature to eternity" Hamlet, Act i, Sc.2
Martin Bazley
2009-11-13 18:21:08 UTC
Permalink
The following bytes were arranged on 12 Nov 2009 by druck :

> Clear Type on Windows, and similar technology on Linux, effectively
> renders the font at 3x the horizontal pixel resolution making the fonts
> look a lot smoother and pleasing to the eye than on RISC OS which is
> limited to whole pixels.
>
I'm not quite clear here - how is that effectively different from
rendering the font at three times its normal size and then scaling it
down? How is this 'better' than the RISC OS approach? (Apart from
using a lot more processor power, of course.)

I once turned on RISC OS's sub-pixel anti-aliasing, but I turned it off
again as it incurred a performance hit and I honestly couldn't see the
difference. This was on a CRT, though - I don't have enough experience
of LCDs to say whether it's an improvement there.

--
__<^>__ "Did you know that polar bears stay white all year round? ...The
/ _ _ \ white colour makes them less visible to the seals and penguins they
( ( |_| ) ) hunt." - Nelson Thornes AQA-endorsed GCSE science textbook
\_> <_/ ======================= Martin Bazley ==========================
Rob Kendrick
2009-11-13 19:23:36 UTC
Permalink
On Fri, 13 Nov 2009 18:21:08 GMT
Martin Bazley <***@blueyonder.co.uk> wrote:

> The following bytes were arranged on 12 Nov 2009 by druck :
>
> > Clear Type on Windows, and similar technology on Linux, effectively
> > renders the font at 3x the horizontal pixel resolution making the
> > fonts look a lot smoother and pleasing to the eye than on RISC OS
> > which is limited to whole pixels.
> >
> I'm not quite clear here - how is that effectively different from
> rendering the font at three times its normal size and then scaling it
> down? How is this 'better' than the RISC OS approach? (Apart from
> using a lot more processor power, of course.)

Because you get better control over those pixels. And don't count on
ClearType-like rendering being slower; it can be faster. For a start,
you're not using as much memory bandwidth, and it means (in the case of
TrueType) that you don't need to perform a second pass when rendering
it in black and white removing orphaned pixels. (FreeType, for
example, the rasteriser used under Linux and many other systems, is
actually faster in anti-aliased mode than in black and white.)

> I once turned on RISC OS's sub-pixel anti-aliasing, but I turned it
> off again as it incurred a performance hit and I honestly couldn't
> see the difference. This was on a CRT, though - I don't have enough
> experience of LCDs to say whether it's an improvement there.

RISC OS doesn't have sub-pixel anti-aliasing. It has a mode where it
will do sub-pixel positioning while anti-aliasing. ie, when you ask
for text of 12.343pt, it's very unlikely that the gap between
characters will be an integer number of pixels wide. It's solving a
completely different problem.

B.
Martin Bazley
2009-11-13 22:05:59 UTC
Permalink
The following bytes were arranged on 13 Nov 2009 by Rob Kendrick :

> On Fri, 13 Nov 2009 18:21:08 GMT
> Martin Bazley <***@blueyonder.co.uk> wrote:
>
> > The following bytes were arranged on 12 Nov 2009 by druck :
> >
> > > Clear Type on Windows, and similar technology on Linux, effectively
> > > renders the font at 3x the horizontal pixel resolution making the
> > > fonts look a lot smoother and pleasing to the eye than on RISC OS
> > > which is limited to whole pixels.
> > >
> > I'm not quite clear here - how is that effectively different from
> > rendering the font at three times its normal size and then scaling it
> > down? How is this 'better' than the RISC OS approach? (Apart from
> > using a lot more processor power, of course.)
>
> Because you get better control over those pixels.

[snip rest of paragraph, which was all about processor power]

...what?

OK, to put it another way, how is rendering a vector graphic to a 180dpi
sprite and averaging out the values of each group of four pixels
different to rendering it twice as big to a 90dpi sprite, then doing an
anti-aliased scale down? I'm even having trouble thinking up a suitable
analogy, the two methods described above being so utterly identical.

> > I once turned on RISC OS's sub-pixel anti-aliasing, but I turned it
> > off again as it incurred a performance hit and I honestly couldn't
> > see the difference. This was on a CRT, though - I don't have enough
> > experience of LCDs to say whether it's an improvement there.
>
> RISC OS doesn't have sub-pixel anti-aliasing. It has a mode where it
> will do sub-pixel positioning while anti-aliasing. ie, when you ask
> for text of 12.343pt, it's very unlikely that the gap between
> characters will be an integer number of pixels wide. It's solving a
> completely different problem.
>
Perhaps somebody needs to open a ROOL ticket about the Configure section
advertising 'Sub-pixel anti-aliasing', then!

--
__<^>__
/ _ _ \ I don't have a problem with God; it's his fan club I can't stand.
( ( |_| ) )
\_> <_/ ======================= Martin Bazley ==========================
Rob Kendrick
2009-11-13 22:08:05 UTC
Permalink
On Fri, 13 Nov 2009 22:05:59 GMT
Martin Bazley <***@blueyonder.co.uk> wrote:

> OK, to put it another way, how is rendering a vector graphic to a
> 180dpi sprite and averaging out the values of each group of four
> pixels different to rendering it twice as big to a 90dpi sprite, then
> doing an anti-aliased scale down? I'm even having trouble thinking
> up a suitable analogy, the two methods described above being so
> utterly identical.

Because in the first case, you're dealing with twice as much data, and
thus it'll take twice as long?

B.
Martin Bazley
2009-11-13 22:40:23 UTC
Permalink
The following bytes were arranged on 13 Nov 2009 by Rob Kendrick :

> On Fri, 13 Nov 2009 22:05:59 GMT
> Martin Bazley <***@blueyonder.co.uk> wrote:
>
> > OK, to put it another way, how is rendering a vector graphic to a
> > 180dpi sprite and averaging out the values of each group of four
> > pixels different to rendering it twice as big to a 90dpi sprite, then
> > doing an anti-aliased scale down? I'm even having trouble thinking
> > up a suitable analogy, the two methods described above being so
> > utterly identical.
>
> Because in the first case, you're dealing with twice as much data, and
> thus it'll take twice as long?
>
No, you're not. Read it carefully - I said twice as big (which you may
interpret as the 'Magnify' option set to a scale factor of 2). Also,
wasn't the first case the one that was supposed to be quicker?

--
__<^>__
/ _ _ \ It is written that Geeks shall inherit the Earth.
( ( |_| ) )
\_> <_/ ======================= Martin Bazley ==========================
druck
2009-11-14 19:33:00 UTC
Permalink
Martin Bazley wrote:
> OK, to put it another way, how is rendering a vector graphic to a 180dpi
> sprite and averaging out the values of each group of four pixels
> different to rendering it twice as big to a 90dpi sprite, then doing an
> anti-aliased scale down? I'm even having trouble thinking up a suitable
> analogy, the two methods described above being so utterly identical.

In the first case you are averaging 4x4 pixels giving 16 levels of
anti-aliasing, in the second case you are averaging 2x2 which only gives
4 levels. i.e. not as good.

> Perhaps somebody needs to open a ROOL ticket about the Configure section
> advertising 'Sub-pixel anti-aliasing', then!

It would be a lot clearer if it was called "cache sub pixel character
variants" or similar.

---druck
Martin Bazley
2009-11-15 11:33:03 UTC
Permalink
The following bytes were arranged on 14 Nov 2009 by druck :

> Martin Bazley wrote:
> > OK, to put it another way, how is rendering a vector graphic to a 180dpi
> > sprite and averaging out the values of each group of four pixels
> > different to rendering it twice as big to a 90dpi sprite, then doing an
> > anti-aliased scale down? I'm even having trouble thinking up a suitable
> > analogy, the two methods described above being so utterly identical.
>
> In the first case you are averaging 4x4 pixels giving 16 levels of
> anti-aliasing, in the second case you are averaging 2x2 which only gives
> 4 levels. i.e. not as good.
>
By 'group of four pixels' I meant '2x2 = 4 pixels' - i.e. exactly the
same.

--
__<^>__
/ _ _ \ I don't have a problem with God; it's his fan club I can't stand.
( ( |_| ) )
\_> <_/ ======================= Martin Bazley ==========================
druck
2009-11-13 22:16:03 UTC
Permalink
Martin Bazley wrote:
> The following bytes were arranged on 12 Nov 2009 by druck :
>
>> Clear Type on Windows, and similar technology on Linux, effectively
>> renders the font at 3x the horizontal pixel resolution making the fonts
>> look a lot smoother and pleasing to the eye than on RISC OS which is
>> limited to whole pixels.
>>
> I'm not quite clear here - how is that effectively different from
> rendering the font at three times its normal size and then scaling it
> down? How is this 'better' than the RISC OS approach?

See the links in Ben's post.

> (Apart from using a lot more processor power, of course.)

But it doesn't, and even if did with an x86 box having 10x to 100x the
speed of a RISC OS machine, there wouldn't be an issue

> I once turned on RISC OS's sub-pixel anti-aliasing, but I turned it off
> again as it incurred a performance hit and I honestly couldn't see the
> difference. This was on a CRT, though - I don't have enough experience
> of LCDs to say whether it's an improvement there.

View this in Homerton 12pt with horizontal s-p a-a on and off :-

iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii

---druck
Rob Kendrick
2009-11-12 01:09:28 UTC
Permalink
On Wed, 11 Nov 2009 21:29:12 GMT
Dave Higton <***@dsl.pipex.com> wrote:

> Curious. RISC OS in general wastes lass space on window furniture
> and useless blank areas that most other OSs, so you can get the
> same amount of information in less pixels.

Fewer? :) The RISC OS window furniture is approximately the same size
as those on Windows and most themes for the desktop environments under
Linux. One big problem, alluded to earlier, are fonts. Sure, you can
make fonts in the desktop smaller, but because of the inflexible static
nature of how how WIMP windows work, it doesn't make most windows any
smaller.

> > Additionally, I find the RISC OS font manager's anti-aliasing makes
> > things worse due to its lack of hinting to pull lines onto pixel
> > boundaries, meaning you can fit less text into a small space with
> > less clarity, due to the blurring effect.
>
> I recall we've disagreed on this topic before. I haven't used
> Mac OS [1], but I have often found the effect that you describe,
> on Windows, is irritating and sometimes slows me down - is that
> a spurious space between those letters? Careful examination -
> no, it's just Windows buggering the spacing up so it looks
> unnatural.

On the plus side, it makes the characters crisper, and serifs more
obvious. RISC OS just looks blurry to me.

> [1] And, since Apple has taken over Microsoft's title as The
> Evil Empire, I'm not likely to any time soon.

Apple have for a long time been far more evil than Microsoft could ever
hope to be. We should count ourselves fortunate that they're not as
powerful.

B.
Vince M Hudd
2009-11-12 11:46:03 UTC
Permalink
Rob Kendrick <***@rjek.com> wrote:
> On Wed, 11 Nov 2009 21:29:12 GMT Dave Higton <***@dsl.pipex.com>
> wrote:

> > Curious. RISC OS in general wastes lass space on window furniture and
> > useless blank areas that most other OSs, so you can get the same amount
> > of information in less pixels.

> Fewer? :) The RISC OS window furniture is approximately the same size as
> those on Windows and most themes for the desktop environments under Linux.

I think the default XP furniture is larger, so I wonder if those who believe
the RISC OS furniture to be significantly smaller are using that on XP?
Things like that are amongst the first to go for me when setting up. They're
just pointless gloss.

[...]

> > I recall we've disagreed on this topic before. I haven't used Mac OS
> > [1], but I have often found the effect that you describe, on Windows, is
> > irritating and sometimes slows me down - is that a spurious space
> > between those letters? Careful examination - no, it's just Windows
> > buggering the spacing up so it looks unnatural.

> On the plus side, it makes the characters crisper, and serifs more
> obvious. RISC OS just looks blurry to me.

I've never been entirely happy with how fonts look on RISC OS since I
started using an LCD display - though it does depend on what I'm using. Some
things I find more bearable than others.

[...]
--
Vince M Hudd - Soft Rock Software
http://www.softrock.co.uk
http://misc.vinceh.com
Rob Kendrick
2009-11-12 12:02:26 UTC
Permalink
On Thu, 12 Nov 2009 11:46:03 +0000
Vince M Hudd <***@softrock.co.uk> wrote:

> I think the default XP furniture is larger, so I wonder if those who
> believe the RISC OS furniture to be significantly smaller are using
> that on XP? Things like that are amongst the first to go for me when
> setting up. They're just pointless gloss.

Oh, I always forget it defaults to Teletubby mode. I turn it back to
Windows 2000-style furniture as the first thing I do on an
installation :)

B.
John M Ward
2009-11-12 12:12:44 UTC
Permalink
In article <***@trite.i.flarn.net.i.flarn.net>,
Rob Kendrick <***@rjek.com> wrote:
> On Thu, 12 Nov 2009 11:46:03 +0000
> Vince M Hudd <***@softrock.co.uk> wrote:

> > I think the default XP furniture is larger, so I wonder if those who
> > believe the RISC OS furniture to be significantly smaller are using
> > that on XP? Things like that are amongst the first to go for me when
> > setting up. They're just pointless gloss.

> Oh, I always forget it defaults to Teletubby mode. I turn it back to
> Windows 2000-style furniture as the first thing I do on an
> installation :)

Here we see the huge difference between what a handful of techies do and
the entirety of the rest of the world -- much of which I have been able to
witness over the years, and never pass up the chance to evaluate (and
memorise) what just about everyone else has with which to work.

--
John Ward in Medway, Kent - using RISC OS since 1987
Now using an Iyonix, an A9home, 2 RiscPCs and Virtual-RPC!
Acorn/RISC OS web page: www.john-ward.org.uk/personal/john/computers
Roger Darlington
2010-02-04 16:33:00 UTC
Permalink
On 12 Nov 2009, Vince M Hudd wrote:

> I've never been entirely happy with how fonts look on RISC OS since I
> started using an LCD display - though it does depend on what I'm using. Some
> things I find more bearable than others.
>

I am driving a 1920 x 1200 LCD screen using the analogue output on the
Iyonix.

If I use the supplied cable, which has a greater analogue bandwidth,
then the characters look crisper, but annoyingly when a window with
text in is moved about, the text characters shimmer annoyingly as they
are traversed the screen pixels.

If I use instead a lower bandwidth cable, the text is slightly blurred
out, but does not exhibit the annoying shimmering described above.

I prefer the latter where the cable is doing some of the anti-aliasing
:-))



--

Cheers
Roger
"An invisibility cloak that works at optical wavelengths is nowhere in
sight"
druck
2010-02-04 22:12:45 UTC
Permalink
Roger Darlington wrote:
> On 12 Nov 2009, Vince M Hudd wrote:
>
>> I've never been entirely happy with how fonts look on RISC OS since I
>> started using an LCD display - though it does depend on what I'm using. Some
>> things I find more bearable than others.
>>
>
> I am driving a 1920 x 1200 LCD screen using the analogue output on the
> Iyonix.
>
> If I use the supplied cable, which has a greater analogue bandwidth,
> then the characters look crisper, but annoyingly when a window with
> text in is moved about, the text characters shimmer annoyingly as they
> are traversed the screen pixels.

The monitor isn't set up correctly, causing a misalignment between
computer and monitor pixels. Try the monitors auto-setup feature,
failing that get the zebra test card and the manual adjustment.

> If I use instead a lower bandwidth cable, the text is slightly blurred
> out, but does not exhibit the annoying shimmering described above.
>
> I prefer the latter where the cable is doing some of the anti-aliasing
> :-))

Erk!

---druck
Roger Darlington
2010-02-05 08:05:14 UTC
Permalink
On 4 Feb 2010, druck wrote:
> Roger Darlington wrote:
>> On 12 Nov 2009, Vince M Hudd wrote:
>>
>>> I've never been entirely happy with how fonts look on RISC OS since I
>>> started using an LCD display - though it does depend on what I'm using. Some
>>> things I find more bearable than others.
>>>
>>
>> I am driving a 1920 x 1200 LCD screen using the analogue output on the
>> Iyonix.
>>
>> If I use the supplied cable, which has a greater analogue bandwidth,
>> then the characters look crisper, but annoyingly when a window with
>> text in is moved about, the text characters shimmer annoyingly as they
>> are traversed the screen pixels.
>
> The monitor isn't set up correctly, causing a misalignment between
> computer and monitor pixels. Try the monitors auto-setup feature,
> failing that get the zebra test card and the manual adjustment.

I'm not at all convinced that there is not already a 1:1
correspondence between computer pixels and screen pixels on my
computer/monitor set-up.

It is auto-set by the monitor (HP w2408h 24 inch). It locks on very
well, the monitor telling me it is set to 1920x1200 (60Hz) pixels,
filling the screen exactly with the 1920 x 1200 pixels that the
computer tells me that it is outputting.

But just to test this, from where do I download this Zebra Test Card;
Google hasn't found it for me. (or is it something I buy?)



>
>> If I use instead a lower bandwidth cable, the text is slightly blurred
>> out, but does not exhibit the annoying shimmering described above.
>>
>> I prefer the latter where the cable is doing some of the anti-aliasing
>> :-))
>
> Erk!
>
> ---druck
>
>
>


--

Cheers
Roger

Atomic Software http://atomicsoftware.org.uk/
Isotope Info http://rogerarm.freeuk.com/rogerarm/
Wild Flowers http://wildflowerfinder.org.uk/
Bury Walkers http://burystrollers.org.uk/
Red sky at night: Turkish delight.
David Pitt
2010-02-05 08:47:16 UTC
Permalink
Roger Darlington <***@freeuk.com> wrote:

[snip]
> But just to test this, from where do I download this Zebra Test Card;
> Google hasn't found it for me. (or is it something I buy?)

zipped. !Zebra by Peter Gaunt, it goes up to 1280*1024.

--
David Pitt

MessengerPro on iMac, Snow Leopard
Roger Darlington
2010-02-05 09:32:40 UTC
Permalink
On 5 Feb 2010, David Pitt wrote:
> Roger Darlington <***@freeuk.com> wrote:
>
> [snip]
>> But just to test this, from where do I download this Zebra Test Card;
>> Google hasn't found it for me. (or is it something I buy?)
>
> zipped. !Zebra by Peter Gaunt, it goes up to 1280*1024.

displaying this test card shows me that there is absolutely no Moire
patterns present in the vertical lines across the screen, which must
surely be the case IF there is not 1:1 correspondence between screen
and computer pixels (as would be the case if, for instance, there was
a 1: 1.013 correspondence).

I conclude that the shimmering of the text I see when I move a window
of text about the screen in a horizontal direction is what would be
expected. I have sub-pixel aliasing switched on in both directions up
to 18 point. My text size is 12 point.

BTW, whilst trying to run Zebra (which isn't 32 bit) I ran Aemulor.
Only afterwards did I see that running Zebra is un-necessary, it is
only necessary to run the sprite file it contains.

But I notice Aemulor now gives an error message when it didn't before
I installed RO5.16. 'Filer quits' or somesuch. So now off to report
this on the Iyo list...

>


--

Cheers
Roger
If you collect them, then, one day - you'll have a lot.
druck
2010-02-05 18:59:59 UTC
Permalink
On 5 Feb 2010 Roger Darlington <***@freeuk.com> wrote:

> On 5 Feb 2010, David Pitt wrote:
>> Roger Darlington <***@freeuk.com> wrote:
>>
>> [snip]
>>> But just to test this, from where do I download this Zebra Test Card;
>>> Google hasn't found it for me. (or is it something I buy?)
>>
>> zipped. !Zebra by Peter Gaunt, it goes up to 1280*1024.

> displaying this test card shows me that there is absolutely no Moire
> patterns present in the vertical lines across the screen, which must
> surely be the case IF there is not 1:1 correspondence between screen
> and computer pixels (as would be the case if, for instance, there was
> a 1: 1.013 correspondence).

You can't be using it correctly if it's only 1280x1024 and you have a
1920x1200 monitor. It needs to be used at the native resolution of the
monitor. I'll send you one which goes up to 2048x1536, but you'll need
to crop it to the correct size and add a single pixel white border for
it to work.

> I conclude that the shimmering of the text I see when I move a window
> of text about the screen in a horizontal direction is what would be
> expected. I have sub-pixel aliasing switched on in both directions up
> to 18 point. My text size is 12 point.

The other thing you may be seeing is the reduced number of lumenance
levels when an object is moved on an LCD screen. It's easiest to see
with coloured objects which tend towards their primary colour. With
anti-aliased text it will look sharper as the grey levels are reduced.

---druck

--
The ARM Club Free Software - http://www.armclub.org.uk/free/
32 bit Conversions Page - http://www.armclub.org.uk/32bit/
druck
2010-02-05 19:25:34 UTC
Permalink
On 5 Feb 2010 druck <***@druck.freeuk.com> wrote:
> You can't be using [Zebra] correctly if it's only 1280x1024 and you have a
> 1920x1200 monitor. It needs to be used at the native resolution of the
> monitor. I'll send you one which goes up to 2048x1536, but you'll need
> to crop it to the correct size and add a single pixel white border for
> it to work.

I was feeling generous so made a 1920x1200 one, as I can use it on our
other monitor too.

---druck

--
The ARM Club Free Software - http://www.armclub.org.uk/free/
32 bit Conversions Page - http://www.armclub.org.uk/32bit/
Roger Darlington
2010-02-10 18:55:11 UTC
Permalink
On 5 Feb 2010, druck wrote:
> On 5 Feb 2010 druck <***@druck.freeuk.com> wrote:
>> You can't be using [Zebra] correctly if it's only 1280x1024 and you have a
>> 1920x1200 monitor. It needs to be used at the native resolution of the
>> monitor. I'll send you one which goes up to 2048x1536, but you'll need
>> to crop it to the correct size and add a single pixel white border for
>> it to work.
>
> I was feeling generous so made a 1920x1200 one, as I can use it on our
> other monitor too.

I've looked on both the URLs below, but cannot find it... Can you
please tell me where it is?


--

Cheers
Roger
Scene one - seen 'em all. Scene two - seen too many.
David Pitt
2010-02-05 09:44:19 UTC
Permalink
David Pitt <***@pittdj.co.uk> wrote:

> Roger Darlington <***@freeuk.com> wrote:
>
> [snip]
> > But just to test this, from where do I download this Zebra Test Card;
> > Google hasn't found it for me. (or is it something I buy?)
>
> zipped. !Zebra by Peter Gaunt, it goes up to 1280*1024.
>
I have found the Iiyama test.bmp now :-

http://www.iiyama.com/en_GB/Downloads

See 'Test software'.

--
David Pitt

MessengerPro on iMac, Snow Leopard
Martin Bazley
2009-11-12 16:18:04 UTC
Permalink
The following bytes were arranged on 12 Nov 2009 by Rob Kendrick :

> On Wed, 11 Nov 2009 21:29:12 GMT
> Dave Higton <***@dsl.pipex.com> wrote:
>
> > Curious. RISC OS in general wastes lass space on window furniture
> > and useless blank areas that most other OSs, so you can get the
> > same amount of information in less pixels.
>
> Fewer? :) The RISC OS window furniture is approximately the same size
> as those on Windows and most themes for the desktop environments under
> Linux.

Erm, no. Windows XP title bars and icons are far bigger.

Not that it makes any difference in the grand scheme of things, but we
might as well start off on the right tack.

> One big problem, alluded to earlier, are fonts. Sure, you can make
> fonts in the desktop smaller, but because of the inflexible static
> nature of how how WIMP windows work, it doesn't make most windows any
> smaller.
>
True enough - but then, most windows are smaller anyway.

> > I recall we've disagreed on this topic before. I haven't used
> > Mac OS [1], but I have often found the effect that you describe,
> > on Windows, is irritating and sometimes slows me down - is that
> > a spurious space between those letters? Careful examination -
> > no, it's just Windows buggering the spacing up so it looks
> > unnatural.
>
> On the plus side, it makes the characters crisper, and serifs more
> obvious. RISC OS just looks blurry to me.
>
Pardon? Did you just call Windows font anti-aliasing 'crisp'?

I won't even dignify that with a response.

--
__<^>__ "Start off every day with a smile and get it over with."
/ _ _ \ - W.C. Fields
( ( |_| ) )
\_> <_/ ======================= Martin Bazley ==========================
Rob Kendrick
2009-11-12 16:28:37 UTC
Permalink
On Thu, 12 Nov 2009 16:18:04 GMT
Martin Bazley <***@blueyonder.co.uk> wrote:

> Pardon? Did you just call Windows font anti-aliasing 'crisp'?
>
> I won't even dignify that with a response.

Straight face, and all: http://www.rjek.com/winxp-ro4.png
Windows font rasterising here may look a little odd if your monitor
does not share the same colour ordering as mine.

You'll also note that the size difference in window furniture is too
close to call without counting pixels. (And these are the default size
settings for Windows XP.)

Sorry, but I count all that grey smudge around the body of the
characters "blur".

B.
John M Ward
2009-11-12 16:41:06 UTC
Permalink
In article <***@trite.i.flarn.net.i.flarn.net>,
Rob Kendrick <***@rjek.com> wrote:
> On Thu, 12 Nov 2009 16:18:04 GMT
> Martin Bazley <***@blueyonder.co.uk> wrote:

> > Pardon? Did you just call Windows font anti-aliasing 'crisp'?
> >
> > I won't even dignify that with a response.

> Straight face, and all: http://www.rjek.com/winxp-ro4.png
> Windows font rasterising here may look a little odd if your monitor
> does not share the same colour ordering as mine.

No problems here, you'll be pleased to know.

> You'll also note that the size difference in window furniture is too
> close to call without counting pixels. (And these are the default size
> settings for Windows XP.)

Yes, this is why I made a point (earlier) of mentioning some of the other
"clutter" that encroaches upon what could otherwise have been what might
be termed "data space", which is more significant in practice.

> Sorry, but I count all that grey smudge around the body of the
> characters "blur".

Hmm. It does seem to be at a different point size, by the way. As Druck
reported a while back -- and this was in regard to XP -- at certain point
sizes, the Microsoft method seems generally okay, but not at what I might
call "inconvenient" point sizes.

A useful clue, though, is looking at the two examples under a magnifier,
even at just 2x magnification. The RISC OS font is correctly scaffolded
and rendered accordingly, whereas the Windows version displays a broken
skeleton, particularly in the capital S. I have saved the image as a
useful demonstration of this effect.

--
John Ward in Medway, Kent - using RISC OS since 1987
Now using an Iyonix, an A9home, 2 RiscPCs and Virtual-RPC!
Acorn/RISC OS web page: www.john-ward.org.uk/personal/john/computers
Rob Kendrick
2009-11-12 16:49:52 UTC
Permalink
On Thu, 12 Nov 2009 16:41:06 GMT
John M Ward <***@acornusers.org> wrote:

> > Sorry, but I count all that grey smudge around the body of the
> > characters "blur".
>
> Hmm. It does seem to be at a different point size, by the way. As
> Druck reported a while back -- and this was in regard to XP -- at
> certain point sizes, the Microsoft method seems generally okay, but
> not at what I might call "inconvenient" point sizes.

ClearType works just as well at any font size. It doesn't work by
just adding pixels; anti-aliasing happens at an earlier point than that.

Lots of people seem to confuse how Windows currently anti-aliases text
(which is designed for TFT use specifically) with how Windows 98 and
2000 anti-aliased text (which was to fill pixels with a shade of grey
that mapped to the proportion of the pixel that the glyph filled).
RISC OS pays no regard to TFTs, and the resampling approach it takes
does not work well at all for modern displays.

> A useful clue, though, is looking at the two examples under a
> magnifier, even at just 2x magnification. The RISC OS font is
> correctly scaffolded and rendered accordingly, whereas the Windows
> version displays a broken skeleton, particularly in the capital S. I
> have saved the image as a useful demonstration of this effect.

That's because the text has been hinted so the main scaffolds that are
essential for readability exist on whole pixel boundries, which is one
of the reasons it's crisper: it doesn't need to anti-alias most of it.

The larger the point size, the less this effect is obvious, obviously.
It suffers slightly at /tiny/ point sizes, but then the hinting takes a
different approach. (And by tiny, I mean text that's 4 pixels tall.)

Graphics and publishing applications under Windows and Linux will use
the same approach as RISC OS however, sacrificing clarity for accurate
representation, which is important for WYSIWYG, but serves no purpose
for a web browser, or email client, etc.

B.
Mr John FO Evans
2009-11-12 19:38:14 UTC
Permalink
In article <***@acornusers.org>, John M Ward
<***@acornusers.org> wrote:
> Hmm. It does seem to be at a different point size, by the way. As Druck
> reported a while back -- and this was in regard to XP -- at certain point
> sizes, the Microsoft method seems generally okay, but not at what I might
> call "inconvenient" point sizes.

Well with white text on a black ground in Word under Windows XP I had a
complete page in which only the bottom half of the characters was displayed
making it imposible to edit. After wasting a great deal of time I discovered
that the document printed correctly. I have never seen anthing as bad as
that under RISCOS.

In general the text in windows DTP is far less readable than under RISCOS.
We have had a number of ridiculous text errors in our local magazine as a
result of this since we changed from RISCOS to XP.

John


--
_ _________________________________________
/ \._._ |_ _ _ /' Orpheus Internet Services
\_/| |_)| |(/_|_|_> / 'Internet for Everyone'
_______ | ___________./ http://www.orpheusinternet.co.uk
Martin Bazley
2009-11-12 20:36:26 UTC
Permalink
The following bytes were arranged on 12 Nov 2009 by Rob Kendrick :

> On Thu, 12 Nov 2009 16:18:04 GMT
> Martin Bazley <***@blueyonder.co.uk> wrote:
>
> > Pardon? Did you just call Windows font anti-aliasing 'crisp'?
> >
> > I won't even dignify that with a response.
>
> Straight face, and all: http://www.rjek.com/winxp-ro4.png
> Windows font rasterising here may look a little odd if your monitor
> does not share the same colour ordering as mine.
>
> You'll also note that the size difference in window furniture is too
> close to call without counting pixels. (And these are the default size
> settings for Windows XP.)
>
Well, for certain values of 'default' - as pointed out elsewhere, the
defaults are much bigger.

> Sorry, but I count all that grey smudge around the body of the
> characters "blur".
>
What do you call that much bigger white smudge around the edge of the
words 'Document - WordPad', then?

Like I said, some fonts show it worse than others. Arial is practically
made for the Windows font manager (earlier versions didn't anti-alias at
all), but Franklin Gothic Book is a shocker. The lower-case 'l', at the
right size, actually turns into a thin grey stroke! (Or it did when I
last tried it, which was the last time I used Office for serious office
work, which was... ooh, two years ago.)

I think a lot of the difference here is due to Arial being a very thin
font, whereas Homerton is slightly fatter and therefore overflows the
pixel margins slightly more.

Having just run some tests, it confirms opinions I have read elsewhere
in that one of the RISC OS font manager's chief strengths is in clarity
at small point sizes - it doesn't really matter how it does it or how it
looks, it gets the point across. And that, surely, is the whole
rationale behind a font manager? (Whereas I clearly remember attempting
to enter very small text in MS Paint - it just looked like a random
collection of pixels.)

Which version of Windows was this?

<takes a good look at screenshot with CloseUp>

...I don't believe it.

Do you realise the Windows fonts - even those on white backgrounds - are
anti-aliased in orange and blue?

WTF?

--
__<^>__ "Did you know that polar bears stay white all year round? ...The
/ _ _ \ white colour makes them less visible to the seals and penguins they
( ( |_| ) ) hunt." - Nelson Thornes AQA-endorsed GCSE science textbook
\_> <_/ ======================= Martin Bazley ==========================
Ben Shimmin
2009-11-12 21:01:31 UTC
Permalink
Martin Bazley <***@blueyonder.co.uk>:

[...]

> Which version of Windows was this?
>
> <takes a good look at screenshot with CloseUp>
>
> ...I don't believe it.
>
> Do you realise the Windows fonts - even those on white backgrounds - are
> anti-aliased in orange and blue?
>
> WTF?

Um, do you know anything at all about subpixel anti-aliasing?

<URL:http://alienryderflex.com/sub_pixel/>
<URL:http://en.wikipedia.org/wiki/Subpixel_rendering>

b.

--
<***@bas.me.uk> <URL:http://bas.me.uk/>
`Property, marriage, the law; as the bed to the river, so rule
and convention to the instinct; and woe to him who tampers with
the banks while the flood is flowing.' -- Samuel Butler, _Erewhon_
Rob Kendrick
2009-11-12 21:11:06 UTC
Permalink
On Thu, 12 Nov 2009 20:36:26 GMT
Martin Bazley <***@blueyonder.co.uk> wrote:

> > You'll also note that the size difference in window furniture is too
> > close to call without counting pixels. (And these are the default
> > size settings for Windows XP.)
> >
> Well, for certain values of 'default' - as pointed out elsewhere, the
> defaults are much bigger.

Default for the "classic" theme. Teletubby theme is about 50% larger,
of course. But who uses that when they know there's an option not to?

> > Sorry, but I count all that grey smudge around the body of the
> > characters "blur".
> >
> What do you call that much bigger white smudge around the edge of the
> words 'Document - WordPad', then?

Bold text. Zoom in, and you'll see it's crisp on all the verticals and
horizontals, where the RISC OS text is just a fuzz.

> Like I said, some fonts show it worse than others. Arial is
> practically made for the Windows font manager (earlier versions
> didn't anti-alias at all)

Anti-aliasing has nothing to do with typeface. At all.

> Which version of Windows was this?

XP.

B.
Tim Hill
2009-11-13 13:09:26 UTC
Permalink
In article <***@trite.i.flarn.net.i.flarn.net>, Rob
Kendrick <***@rjek.com> wrote:
> On Thu, 12 Nov 2009 16:18:04 GMT Martin Bazley
> <***@blueyonder.co.uk> wrote:

> > Pardon? Did you just call Windows font anti-aliasing 'crisp'?
> >
> > I won't even dignify that with a response.

> Straight face, and all: http://www.rjek.com/winxp-ro4.png Windows font
> rasterising here may look a little odd if your monitor does not share
> the same colour ordering as mine.

Blowing the thing up reveals all sorts of blues and pinks being used.
It's ClearType then. You prefer lumpy-bumpy windows fonts and are tying
to tell us that blurring the edges with colour is better than RISC OS'
b&w effort at a different point size. :-/

> You'll also note that the size difference in window furniture is too
> close to call without counting pixels. (And these are the default size
> settings for Windows XP.)

You forget the inch of unnecessary clutter at the top of the wordpad
window, and have you ever seen how much crap some people add to their
browser windows? As it's always there, the menu strip at least should be
included in any measurement, not just the window furniture.

> Sorry, but I count all that grey smudge around the body of the
> characters "blur".

Which applies to both, of course, though with TrueType it pink and blue
and green rather than grey because it uses the RGB sub-pixels
independently rather than RISC OS method of using each combined R+G+B
pixel to produce greys.

Consequently, TrueType /should/ look better but I think it looks horrid.
Lumpy-bumpy and coloured, rather than smooth and monochrome.

There is also a difference in using a font specifically designed for
screen display (Arial) and one designed for print (Homerton).

--
Have an ethical share in cheaper telecoms: http://tinyurl.com/phone-coop
Genuine spam-proof addresses for Usenet: http://www.invalid.org.uk/
Email address for replies: substitute postmaster@ for tim@

... "Sacred and sweet was all I saw in her" T of the S, Act i, Sc.1
Rob Kendrick
2009-11-13 13:36:10 UTC
Permalink
On Fri, 13 Nov 2009 13:09:26 +0000 (GMT)
Tim Hill <***@invalid.org.uk> wrote:

> > Straight face, and all: http://www.rjek.com/winxp-ro4.png Windows
> > font rasterising here may look a little odd if your monitor does
> > not share the same colour ordering as mine.
>
> Blowing the thing up reveals all sorts of blues and pinks being used.
> It's ClearType then. You prefer lumpy-bumpy windows fonts and are
> tying to tell us that blurring the edges with colour is better than
> RISC OS' b&w effort at a different point size. :-/

The fact that you had to zoom into the image before you noticed the
colours is evidence enough that the system works. I have no idea what
"lumpy-bumpy" means. Perhaps you're refering to hinting? Personally,
making slight alterations to a glyph to make its shapes readable at
variable sizes (which RISC OS also does, if the font supports it, IIRC)
in conjuction with an understanding of how to make things look clear on
such a low-resolution device (such as a monitor) makes much more sense
than just smudging the edges. RISC OS fonts just look like I need to
have my glasses prescription updated.

> There is also a difference in using a font specifically designed for
> screen display (Arial) and one designed for print (Homerton).

Buh? Where do you get this from? Both fonts are metric clones of
Helvetica; spotting the difference is difficult without a trained eye.
And Helvetica was designed long before computer typesetting was
thought about. (1957.)

B.
Ben Shimmin
2009-11-13 14:40:57 UTC
Permalink
Rob Kendrick <***@rjek.com>:

[...]

> Buh? Where do you get this from? Both fonts are metric clones of
> Helvetica; spotting the difference is difficult without a trained eye.

Oh, I don't know about that -- look at the letter `R'!

b.

--
<***@bas.me.uk> <URL:http://bas.me.uk/>
`Property, marriage, the law; as the bed to the river, so rule
and convention to the instinct; and woe to him who tampers with
the banks while the flood is flowing.' -- Samuel Butler, _Erewhon_
Rob Kendrick
2009-11-13 14:47:04 UTC
Permalink
On Fri, 13 Nov 2009 14:40:57 +0000
Ben Shimmin <***@llamaselector.com> wrote:

> Rob Kendrick <***@rjek.com>:
>
> [...]
>
> > Buh? Where do you get this from? Both fonts are metric clones of
> > Helvetica; spotting the difference is difficult without a trained
> > eye.
>
> Oh, I don't know about that -- look at the letter `R'!

That's the profoundest difference, yes. And I bet most people miss it
unless they're looking for it.

This site is fun: http://www.iliveonyourvisits.com/helvetica/

B.
Ben Shimmin
2009-11-13 15:45:59 UTC
Permalink
Rob Kendrick <***@rjek.com>:
> On Fri, 13 Nov 2009 14:40:57 +0000
> Ben Shimmin <***@llamaselector.com> wrote:
>> Rob Kendrick <***@rjek.com>:
>> > Buh? Where do you get this from? Both fonts are metric clones of
>> > Helvetica; spotting the difference is difficult without a trained
>> > eye.
>>
>> Oh, I don't know about that -- look at the letter `R'!
>
> That's the profoundest difference, yes. And I bet most people miss it
> unless they're looking for it.
>
> This site is fun: http://www.iliveonyourvisits.com/helvetica/

Bah, easy -- 10/10.

This one is slightly harder (well, a couple of them are, anyway):

<URL:http://www.ironicsans.com/helvarialquiz/>

I only got 19!

b.

--
<***@bas.me.uk> <URL:http://bas.me.uk/>
`Property, marriage, the law; as the bed to the river, so rule
and convention to the instinct; and woe to him who tampers with
the banks while the flood is flowing.' -- Samuel Butler, _Erewhon_
Martin Bazley
2009-11-13 18:36:38 UTC
Permalink
The following bytes were arranged on 13 Nov 2009 by Ben Shimmin :

> Rob Kendrick <***@rjek.com>:
> > On Fri, 13 Nov 2009 14:40:57 +0000
> > Ben Shimmin <***@llamaselector.com> wrote:
> >> Rob Kendrick <***@rjek.com>:
> >> > Buh? Where do you get this from? Both fonts are metric clones of
> >> > Helvetica; spotting the difference is difficult without a trained
> >> > eye.
> >>
> >> Oh, I don't know about that -- look at the letter `R'!
> >
> > That's the profoundest difference, yes. And I bet most people miss it
> > unless they're looking for it.
> >
> > This site is fun: http://www.iliveonyourvisits.com/helvetica/
>
> Bah, easy -- 10/10.
>
> This one is slightly harder (well, a couple of them are, anyway):
>
> <URL:http://www.ironicsans.com/helvarialquiz/>
>
> I only got 19!
>
Me too - it was Mattel that did it.

It's easy once you get past the first few and get to know what to look
for - principally, Helvetica is slightly bolder (as I pointed out
earlier), and the tails of its letters are always square.

--
__<^>__
/ _ _ \ You always find something in the last place you look.
( ( |_| ) )
\_> <_/ ======================= Martin Bazley ==========================
druck
2009-11-13 22:23:28 UTC
Permalink
Ben Shimmin wrote:
> Rob Kendrick <***@rjek.com>:
>
>> Buh? Where do you get this from? Both fonts are metric clones of
>> Helvetica; spotting the difference is difficult without a trained eye.

> Oh, I don't know about that -- look at the letter `R'!

Both are Helvetica like fonts, but not identical to avoid having to pay
the Monotype Corporation licences. All the Acorn supplied RISC OS fonts
are look-a-likes named after Cambridge colleges starting with the same
letter as the Monotype font.

---druck
Ben Shimmin
2009-11-13 22:52:20 UTC
Permalink
druck <***@druck.org.uk>:
> Ben Shimmin wrote:
>> Rob Kendrick <***@rjek.com>:
>>> Buh? Where do you get this from? Both fonts are metric clones of
>>> Helvetica; spotting the difference is difficult without a trained eye.
>
> > Oh, I don't know about that -- look at the letter `R'!
>
> Both are Helvetica like fonts, but not identical to avoid having to pay
> the Monotype Corporation licences.

That's a confusing sentence. Arial is owned by Monotype. Helvetica
is owned by Linotype. Unless they're... yes, they're now the same
company. How droll.

> All the Acorn supplied RISC OS fonts are look-a-likes named after
> Cambridge colleges starting with the same letter as the Monotype font.

I always thought that was rather sweet, though I seem to recall the
first letter thing isn't always true -- isn't Clare the equivalent of
AvantGarde, and... I'm sure there's another, but I forget.

b.

--
<***@bas.me.uk> <URL:http://bas.me.uk/>
`Property, marriage, the law; as the bed to the river, so rule
and convention to the instinct; and woe to him who tampers with
the banks while the flood is flowing.' -- Samuel Butler, _Erewhon_
John M Ward
2009-11-14 07:07:16 UTC
Permalink
In article <***@rialto.bas.me.uk>,
Ben Shimmin <***@llamaselector.com> wrote:
> druck <***@druck.org.uk>:

[some snipped]

> > All the Acorn supplied RISC OS fonts are look-a-likes named after
> > Cambridge colleges starting with the same letter as the Monotype font.

> I always thought that was rather sweet, though I seem to recall the
> first letter thing isn't always true -- isn't Clare the equivalent of
> AvantGarde, and... I'm sure there's another, but I forget.

The idea was to suggest the "original" name via the initial letter (or
more) of the college name; but such a scheme depends on there being a
Cambridge college with the right initial, so inevitably there would come a
time when a font was simulated without having a suitable college name.

Just for the record, here are the ones I know (excluding Porterhouse,
which was a bitmap font rather than a scalable outline font):


* Clare - AvantGarde

* Corpus - Courier

* Homerton - Helvetica

* New Hall - New Century Schoolbook

* Selwyn - Dingbats

* Sidney - Symbol

* Trinity - Times


I'm sure there are at least one or two more.

--
John Ward in Medway, Kent - using RISC OS since 1987
Now using an Iyonix, an A9home, 2 RiscPCs and Virtual-RPC!
Acorn/RISC OS web page: www.john-ward.org.uk/personal/john/computers
Tim Hill
2009-11-14 17:25:51 UTC
Permalink
In article <***@acornusers.org>, John M Ward
<***@acornusers.org> wrote:
> such a scheme depends on there being a Cambridge college with the right
> initial, so inevitably there would come a time when a font was
> simulated without having a suitable college name.

When dealing with such problems the answer would be to remove head from
arse and use a name not connected to any Cambridge college. It isn't
important.

"All Souls" would have done for "A" ;-))
charles
2009-11-14 18:21:30 UTC
Permalink
In article <***@invalid.org.uk>,
Tim Hill <***@invalid.org.uk> wrote:
> In article <***@acornusers.org>, John M Ward
> <***@acornusers.org> wrote:
> > such a scheme depends on there being a Cambridge college with the right
> > initial, so inevitably there would come a time when a font was
> > simulated without having a suitable college name.

> When dealing with such problems the answer would be to remove head from
> arse and use a name not connected to any Cambridge college. It isn't
> important.

> "All Souls" would have done for "A" ;-))

wrong colour blue.

--
From KT24

Using a RISC OS computer running v5.11
Tim Hill
2009-11-13 14:45:07 UTC
Permalink
In article <***@trite.i.flarn.net.i.flarn.net>, Rob
Kendrick <***@rjek.com> wrote:
> On Fri, 13 Nov 2009 13:09:26 +0000 (GMT) Tim Hill <***@invalid.org.uk>
> wrote:

> > > Straight face, and all: http://www.rjek.com/winxp-ro4.png Windows
> > > font rasterising here may look a little odd if your monitor does
> > > not share the same colour ordering as mine.
> >
> > Blowing the thing up reveals all sorts of blues and pinks being used.
> > It's ClearType then. You prefer lumpy-bumpy windows fonts and are
> > tying to tell us that blurring the edges with colour is better than
> > RISC OS' b&w effort at a different point size. :-/

> The fact that you had to zoom into the image before you noticed the
> colours is evidence enough that the system works.

I had to zoom in only to check; I could see a blue and pink fringes
already and dislike ClearType for this reason. I said 'blowing the thing
up' to cater for those who can't see such things with their eyes.

> I have no idea what "lumpy-bumpy" means.

Can you be that thick? Think what it might mean when applied to
font-rendering which resembles a staircase in profile (top curve of the
window's 'S') or has an extra blobby bit on the top end of the upper-case
'S', as in your example. They are also referred to as 'jaggies'.

> Perhaps you're refering to hinting?

Perhaps: bad hinting.

Hinting is about maintaining the appearance at different sizes. Windows
TrueType doesn't. It all falls to bits at small sizes, becomes pixellated
and adds colour where none is wanted. Your example demonstrates this. If
you can't see it, aren't you lucky!

> Personally,
> making slight alterations to a glyph to make its shapes readable at
> variable sizes (which RISC OS also does, if the font supports it, IIRC)
> in conjuction with an understanding of how to make things look clear on
> such a low-resolution device (such as a monitor) makes much more sense
> than just smudging the edges. RISC OS fonts just look like I need to
> have my glasses prescription updated.

If you can't see the coloured fringing in TrueType without magnification,
perhaps you do. If you can see RISC OS 'blur' you should be able to see
the color (sic.) too. And the hard edges in the wrong places; but you
seem to like those as they make it 'crisp'. Each to his own.

> > There is also a difference in using a font specifically designed for
> > screen display (Arial) and one designed for print (Homerton).

> Buh? Where do you get this from?

<crosses fingers and toes />
http://en.wikipedia.org/wiki/Arial

"Though nearly identical to Linotype Helvetica in both proportion and
weight (see figure), the design of Arial is in fact a variation of
Monotype Grotesque,[3] and was designed for IBM's laserxerographic
printer.[2] Subtle changes and variations were made to both the
letterforms and the spacing between characters, in order to make it more
readable on screen and at various resolutions."

> Both fonts are metric clones of
> Helvetica; spotting the difference is difficult without a trained eye.

Arial is not a clone of Helvetica (see above) and it /does not/ need a
trained eye to see the differences, just an observant one. Look at an 'R'
as shown on Wikip.

> And Helvetica was designed long before computer typesetting was thought
> about. (1957.)

Computer typesetting was thought about long before 1957. In 1949, the
book 1984 included a 'speak-write' machine, didn't it?
--
Tim Hill
tjrh.eu
Rob Kendrick
2009-11-13 15:12:34 UTC
Permalink
On Fri, 13 Nov 2009 14:45:07 +0000 (GMT)
Tim Hill <***@invalid.org.uk> wrote:

> > The fact that you had to zoom into the image before you noticed the
> > colours is evidence enough that the system works.
>
> I had to zoom in only to check; I could see a blue and pink fringes
> already and dislike ClearType for this reason. I said 'blowing the
> thing up' to cater for those who can't see such things with their
> eyes.

As I said, if your TFT has a different pixel ordering to mine, it'll
look odd. The text looks entirely black on mine.

> > I have no idea what "lumpy-bumpy" means.
>
> Can you be that thick?

I'm giving up now. You've already demonstrated a wide and complete
ignorance of typography, and now we get insults.

B.
Tim Hill
2009-11-13 15:58:02 UTC
Permalink
In article <***@trite.i.flarn.net.i.flarn.net>, Rob
Kendrick <***@rjek.com> wrote:
> On Fri, 13 Nov 2009 14:45:07 +0000 (GMT) Tim Hill <***@invalid.org.uk>
> wrote:

> > > The fact that you had to zoom into the image before you noticed the
> > > colours is evidence enough that the system works.
> >
> > I had to zoom in only to check; I could see a blue and pink fringes
> > already and dislike ClearType for this reason. I said 'blowing the
> > thing up' to cater for those who can't see such things with their
> > eyes.

> As I said, if your TFT has a different pixel ordering to mine, it'll
> look odd. The text looks entirely black on mine.

Well if you're using ClearType it /will/ use the R G B sub-pixels to
create the illusion so the colours /are/ there you just can't or choose
not to see them. The layout of your TFT is irrelevant, just that
different colours may show through.

Try using a magnifier to see. (A real one).

> > > I have no idea what "lumpy-bumpy" means.
> >
> > Can you be that thick?

> I'm giving up now. You've already demonstrated a wide and complete
> ignorance of typography,

Really? Have I? Where?

> and now we get insults.

If the cap fits. Or do you still want to say that you have no idea what I
meant when describing a font rendering as 'lumpy-bumpy'?

Go ahead, feign offence, write nonsense and run away again.
John M Ward
2009-11-13 16:47:59 UTC
Permalink
In article <***@invalid.org.uk>,
Tim Hill <***@invalid.org.uk> wrote:
> In article <***@trite.i.flarn.net.i.flarn.net>, Rob
> Kendrick <***@rjek.com> wrote:
> > On Fri, 13 Nov 2009 14:45:07 +0000 (GMT) Tim Hill
> > <***@invalid.org.uk> wrote:

> > > > The fact that you had to zoom into the image before you noticed
> > > > the colours is evidence enough that the system works.
> > >
> > > I had to zoom in only to check; I could see a blue and pink fringes
> > > already and dislike ClearType for this reason. I said 'blowing the
> > > thing up' to cater for those who can't see such things with their
> > > eyes.

> > As I said, if your TFT has a different pixel ordering to mine, it'll
> > look odd. The text looks entirely black on mine.

Works OK here, though not as well as I'd been led to believe. At normal
size (i.e. non-magnified) the t and l verticals aren't realy 'all there'
in the Windows version, and the RISC OS l is a tiny bit lacking as well,
though nowhere near as much.

The use of the colours as physical sub-pixels is a cleverly-devised
optical illusion, and one has to give credit to those who tried to devise
a workable system -- indeed, it seems to work for some, though not all
that well for me.

I did wonder whether it was just because of what was familiar, but I've
done enough tests and "tricks" to take that variable out of the equation
(too long to detail here) and it is genuine. Neither is perfect (of
course) but I find the RISC OS version better overall, including in web
pages on TFT displays, which I wasn't necessarily expecting any more.

> Well if you're using ClearType it /will/ use the R G B sub-pixels to
> create the illusion so the colours /are/ there you just can't or choose
> not to see them. The layout of your TFT is irrelevant, just that
> different colours may show through.

It might be to do with a known optical effect with which I have to admit I
(as non-expert) am not familiar. Again: clever in its way, and it's good
that someone is making the effort to try to improve this area of
technology, but definitely not there yet...

> Try using a magnifier to see. (A real one).

It is instructive, to see how it's being done. The technique does seem to
result in displayed lettering that seems "wrong" somehow and I find it
impairs readability in terms of actually reading a chunk of text. This
applies to serif fonts as well as the example provided in the PNG image.

Despite an undoubtedly clever algorithm, the method just doesn't get text
right for readability, for which purpose body fonts in particular were
carefully designed. It is perhaps unsurprising that the changed form that
ClearType renders is therefore less easily read.

[small snip to aid the flow!]

> ...Or do you still want to say that you have no idea what I meant when
> describing a font rendering as 'lumpy-bumpy'?

It's actually quite a good description! I might use it myself one day,
somewhere, if it isn't copyrighted :-)

--
John Ward in Medway, Kent - using RISC OS since 1987
Now using an Iyonix, an A9home, 2 RiscPCs and Virtual-RPC!
Acorn/RISC OS web page: www.john-ward.org.uk/personal/john/computers
Martin Bazley
2009-11-13 18:42:13 UTC
Permalink
The following bytes were arranged on 13 Nov 2009 by Tim Hill :

> > > > I have no idea what "lumpy-bumpy" means.
> > >
> > > Can you be that thick?
[...]
> > and now we get insults.
>
> If the cap fits. Or do you still want to say that you have no idea what I
> meant when describing a font rendering as 'lumpy-bumpy'?
>
> Go ahead, feign offence, write nonsense and run away again.
>

Actually, I think he was objecting to being called 'thick'.

Still, it wouldn't be a csa.advocacy thread without a good old
mudslinging match...

--
__<^>__
/ _ _ \ I don't have a problem with God; it's his fan club I can't stand.
( ( |_| ) )
\_> <_/ ======================= Martin Bazley ==========================
druck
2009-11-13 22:36:49 UTC
Permalink
Tim Hill wrote:
>> As I said, if your TFT has a different pixel ordering to mine, it'll
>> look odd. The text looks entirely black on mine.
>
> Well if you're using ClearType it /will/ use the R G B sub-pixels to
> create the illusion so the colours /are/ there you just can't or choose
> not to see them. The layout of your TFT is irrelevant, just that
> different colours may show through.

No, you aren't even thinking about what you are writing now.

Sub pixel rendering has to be taylored to the ordering of the component
stripes. Rendering which expects RGBRGBRGB will not work on BGRBGRBGR

> Try using a magnifier to see. (A real one).

You don't view sub pixel a-a through a magnifier any more than whole
pixel a-a. It's what it looks like in use, and your own personal
preferences to which one is preferable on the displays you use.

> Go ahead, feign offence, write nonsense and run away again.

Don't forget that advice applies equally to yourself.

RISC OS is an operating system not a religion. If a new technology comes
along which isn't on RISC OS, it is not automatically worse/wrong/work
of the devil.

---druck
Tim Hill
2009-11-14 17:38:51 UTC
Permalink
In article <***@druck.org.uk>, druck <***@druck.org.uk> wrote:
> Tim Hill wrote:
> >> As I said, if your TFT has a different pixel ordering to mine, it'll
> >> look odd. The text looks entirely black on mine.
> >
> > Well if you're using ClearType it /will/ use the R G B sub-pixels to
> > create the illusion so the colours /are/ there you just can't or
> > choose not to see them. The layout of your TFT is irrelevant, just
> > that different colours may show through.

> No, you aren't even thinking about what you are writing now.

> Sub pixel rendering has to be taylored to the ordering of the component
> stripes. Rendering which expects RGBRGBRGB will not work on BGRBGRBGR

No indeed. The colours will look 'wrong' but surely TrueType and the Gfx
system will take care of all this so it shouldn't happen?

Given that there are so few LCD panel manufacturers, are there that many
that differ in this way?

[Snip]

> RISC OS is an operating system not a religion. If a new technology
> comes along which isn't on RISC OS, it is not automatically
> worse/wrong/work of the devil.

No indeed. RISC OS is no effing use for most browsing these days.

D. All I know is that I just got home from a Powerpoint presentation fed
to a small projector from a Vista laptop. All I can honestly say is that
the fonts looked terrible. I used to get better results with my A4
because - you've guessed it - the font rendering is better by default.

In fact all those years ago people were amazed at the on-screen quality
of the display (though back then it was one of those LCD panels which fit
on an OHP).

--
Tim Hill
--------
tjrh.eu

... "Praising what is lost, makes the remembrance dear" All's Well, Act v, Sc.3
Steve Fryatt
2009-11-14 18:20:39 UTC
Permalink
On 14 Nov, Tim Hill wrote in message
<***@invalid.org.uk>:

> D. All I know is that I just got home from a Powerpoint presentation fed
> to a small projector from a Vista laptop. All I can honestly say is that
> the fonts looked terrible.

How much of that was down to the fact that the system was rendering the
fonts for a TFT screen and not for whatever the projector used? And was the
resolution set to the native one for the projector, or was the projector
'downsampling' (for want of a better word -- dropping pixels seems to be
what usually happens) to make the display fit?

It's very easy to make RISC OS fonts look terrible on a projector if you
don't know what you're doing.

--
Steve Fryatt - Leeds, England

http://www.stevefryatt.org.uk/
Tim Hill
2009-11-21 11:32:53 UTC
Permalink
In article <***@stevefryatt.org.uk>, Steve Fryatt
<***@stevefryatt.org.uk> wrote:
> On 14 Nov, Tim Hill wrote in message <***@invalid.org.uk>:

> > D. All I know is that I just got home from a Powerpoint presentation
> > fed to a small projector from a Vista laptop. All I can honestly say
> > is that the fonts looked terrible.

> How much of that was down to the fact that the system was rendering the
> fonts for a TFT screen and not for whatever the projector used? And
> was the resolution set to the native one for the projector, or was the
> projector 'downsampling' (for want of a better word -- dropping pixels
> seems to be what usually happens) to make the display fit?

No it was the best they could get out of the set-up. Apparently.

> It's very easy to make RISC OS fonts look terrible on a projector if
> you don't know what you're doing.

And easy if you know how to operate the equipment properly. It's not
rocket science or brain surgery.

--
Have an ethical share in cheaper telecoms: http://tinyurl.com/phone-coop
Genuine spam-proof addresses for Usenet: http://www.invalid.org.uk/
Email address for replies: substitute postmaster@ for tim@

... "The elements be kind to thee, and make thy spirits all of comfort: fare thee well !" Ant & Cleo, Act iii, Sc.2
Ben Shimmin
2009-11-13 15:54:08 UTC
Permalink
Tim Hill <***@invalid.org.uk>:
> In article <***@trite.i.flarn.net.i.flarn.net>, Rob
> Kendrick <***@rjek.com> wrote:
>> Buh? Where do you get this from?
>
> <crosses fingers and toes />
> http://en.wikipedia.org/wiki/Arial
>
> "Though nearly identical to Linotype Helvetica in both proportion and
> weight (see figure), the design of Arial is in fact a variation of
> Monotype Grotesque,[3] and was designed for IBM's laserxerographic
> printer.[2] Subtle changes and variations were made to both the
> letterforms and the spacing between characters, in order to make it more
> readable on screen and at various resolutions."
>
>> Both fonts are metric clones of
>> Helvetica; spotting the difference is difficult without a trained eye.
>
> Arial is not a clone of Helvetica (see above)

Note that Rob wrote `metric clone'. In the Wikipedia page you cite,
note also that it says `Arial's glyph widths are nearly identical to
those of Helvetica' (not of Grotesque).

Arial is Microsoft's swap-in replacement for Helvetica, and it
conveniently side-steps any need for them to pay royalties or licence
fees. You might lastly note that the default system font on Mac OS X
is, yes, you guessed it, Helvetica.

b.

--
<***@bas.me.uk> <URL:http://bas.me.uk/>
`Property, marriage, the law; as the bed to the river, so rule
and convention to the instinct; and woe to him who tampers with
the banks while the flood is flowing.' -- Samuel Butler, _Erewhon_
Tim Hill
2009-11-13 16:52:23 UTC
Permalink
In article <***@rialto.bas.me.uk>, Ben Shimmin
<***@llamaselector.com> wrote:
> Tim Hill <***@invalid.org.uk>:
> > In article <***@trite.i.flarn.net.i.flarn.net>,
> > Rob Kendrick <***@rjek.com> wrote:
> >> Buh? Where do you get this from?
> >
> > <crosses fingers and toes /> http://en.wikipedia.org/wiki/Arial
> >
> > "Though nearly identical to Linotype Helvetica in both proportion and
> > weight (see figure), the design of Arial is in fact a variation of
> > Monotype Grotesque,[3] and was designed for IBM's laserxerographic
> > printer.[2] Subtle changes and variations were made to both the
> > letterforms and the spacing between characters, in order to make it
> > more readable on screen and at various resolutions."
> >
> >> Both fonts are metric clones of Helvetica; spotting the difference
> >> is difficult without a trained eye.
> >
> > Arial is not a clone of Helvetica (see above)

> Note that Rob wrote `metric clone'.

Noted. He did and it isn't.

> In the Wikipedia page you cite,
> note also that it says `Arial's glyph widths are nearly identical

A clone is not /nearly/ identical. A clone /is/ identical.

> to
> those of Helvetica' (not of Grotesque).

> Arial is Microsoft's swap-in replacement for Helvetica,

I wouldn't disagree.

> and it
> conveniently side-steps any need for them to pay royalties or licence
> fees.

No surprises there.

> You might lastly note that the default system font on Mac OS X
> is, yes, you guessed it, Helvetica.

So why would they bother to have a version of Arial from Mac OS X onwards
if it isn't different to Helvetica? Because it is. Even windoze has
'proper' Helvetica clones available as well as Arial.

This is a hoary old chestnut but let's not get hung up on what you and
Rob define as a 'clone'. I think you'll have to stick to your definition
(nearly the same) and I'll stick to mine which is that clones are
identical (as Homerton/Helvetica are meant to be). Arial and Helvetica
are not identical, metrically or otherwise. Though some (most) of the
appearance is similar, they are different in a similar but perhaps subtle
ways that RISC OS 5 and 6 are. Separated back along the family tree and
no longer exactly the same.

Just to mention that small differences to others are important to me
probably because I work with fonts at very large sizes at the final
printed stage where differences may be more marked than in the average
person's desktop where I agree that they look the same and could be
assumed to be clones. The conversation is a little moot, though, as I
would usually avoid using any common windows desktop or web fonts -
clones or otherwise - such as those discussed here. ;-)
Ben Shimmin
2009-11-13 21:00:41 UTC
Permalink
Tim Hill <***@invalid.org.uk>:
> In article <***@rialto.bas.me.uk>, Ben Shimmin
> <***@llamaselector.com> wrote:

[...]

>> In the Wikipedia page you cite,
>> note also that it says `Arial's glyph widths are nearly identical
>
> A clone is not /nearly/ identical. A clone /is/ identical.
>
>> to those of Helvetica' (not of Grotesque).
>
>> Arial is Microsoft's swap-in replacement for Helvetica,
>
> I wouldn't disagree.

I rather like the Jargon File's second and third definitions of the
word `clone', which certainly describe what I would intend the word to
mean in the context of the relationship between Helvetica and Arial:

2. A shoddy, spurious copy: "Their product is a clone of our
product."

3. A blatant ripoff, most likely violating copyright, patent, or
trade secret protections: "Your product is a clone of my product."
This use implies legal action is pending.

(Mostly only the first three words of the third definition apply here,
admittedly.)

b.

--
<***@bas.me.uk> <URL:http://bas.me.uk/>
`Property, marriage, the law; as the bed to the river, so rule
and convention to the instinct; and woe to him who tampers with
the banks while the flood is flowing.' -- Samuel Butler, _Erewhon_
Steve Fryatt
2009-11-14 18:29:13 UTC
Permalink
On 13 Nov, Tim Hill wrote in message
<***@invalid.org.uk>:

> In article <***@rialto.bas.me.uk>, Ben Shimmin
> <***@llamaselector.com> wrote:
> > Tim Hill <***@invalid.org.uk>:
> >
> > > Arial is not a clone of Helvetica (see above)
>
> > Note that Rob wrote `metric clone'.
>
> Noted. He did and it isn't.

I may be misunderstanding this, but AFAIK the metrics or Arial, Homerton and
Helvetica are near enough identical that you can take a document flowed in
one and print it in another and it will still look correct. Indeed, this is
what would happen if you printed a RISC OS document via PS without including
fonts, and I have done exactly this on *many* occasions so that I know that
it works.

> > In the Wikipedia page you cite, note also that it says `Arial's glyph
> > widths are nearly identical
>
> A clone is not /nearly/ identical. A clone /is/ identical.

Having used Google, other sites seem to back up my understanding above.

--
Steve Fryatt - Leeds, England

http://www.stevefryatt.org.uk/
Tim Hill
2009-11-21 11:34:02 UTC
Permalink
In article <***@stevefryatt.org.uk>, Steve Fryatt
<***@stevefryatt.org.uk> wrote:

[snip]

> Having used Google, other sites seem to back up my understanding above.


Good for you.
Mr John FO Evans
2009-11-13 15:50:17 UTC
Permalink
In article <***@invalid.org.uk>, Tim Hill <***@invalid.org.uk>
wrote:
> If you can't see the coloured fringing in TrueType without magnification,
> perhaps you do. If you can see RISC OS 'blur' you should be able to see
> the color (sic.) too. And the hard edges in the wrong places; but you
> seem to like those as they make it 'crisp'. Each to his own.

We sometimes forget that we are often conditioned by experience. If you are
used to non-aliased fonts as in earlier windows systems you may well feel
that the windows version of anti-alias is right. If on the other hand you
are used to RISCOS antialiasing (very few people will have seen the earlier
non-aliased Acorn Fonts) you are likely to prefer that.

All I know is that one visually impaired friend who uses magnified windows
was highly impressed by the visibility of the RISCOS desktop.

John


--
_ _________________________________________
/ \._._ |_ _ _ /' Orpheus Internet Services
\_/| |_)| |(/_|_|_> / 'Internet for Everyone'
_______ | ___________./ http://www.orpheusinternet.co.uk
John M Ward
2009-11-13 16:17:58 UTC
Permalink
In article <***@orpheusmail.co.uk>,
Mr John FO Evans <***@orpheusmail.co.uk> wrote:

[some snipped]

> All I know is that one visually impaired friend who uses magnified
> windows was highly impressed by the visibility of the RISCOS desktop.

Visitors accustomed to other systems have been consistently impressed by
the clarity on my systems during the past twenty years. I even lent my
old A540 to my office for a couple of months, some years ago, just to show
my colleagues what I meant.

Although the situation has changed to some extent in recent years, I am
still finding that first-time visitors' spontaneous reaction is
consistently positive; and I take that as a better and more representative
sample of "real people" than the occasional individual posting negative
stuff here. Am I wrong?

--
John Ward in Medway, Kent - using RISC OS since 1987
Now using an Iyonix, an A9home, 2 RiscPCs and Virtual-RPC!
Acorn/RISC OS web page: www.john-ward.org.uk/personal/john/computers
Ollie Clark
2009-11-10 12:25:28 UTC
Permalink
Mr John FO Evans wrote:
> In article <***@rialto.bas.me.uk>, Ben Shimmin
><***@llamaselector.com> wrote:
>> > I'm pretty sure that like all other operating systems, Mac OS lets you
>> > change the resolution. Problem solved. I can't imagine what sort of hell
>> > it would be working at 800x600 these days but each to their own.
>
> Unless of course you are working under RISC OS which uses space efficiently
> for most applications and doesn't need enormous resolution to see even small
> text due to its anti-aliasing.

No, even if I'm using RISC OS I rarely drop down to 800x600. It just
doesn't fit enough on the screen. When I do go down to 800x600 to get
16 million colours it feels very cramped indeed and I go back up to
1280x1024 as soon as I can.

Cheers,

Ollie
druck
2009-11-10 21:47:25 UTC
Permalink
Mr John FO Evans wrote:
> Unless of course you are working under RISC OS which uses space
> efficiently for most applications and doesn't need enormous
> resolution to see even small text due to its anti-aliasing.

I find I have to use RISC Os at least 25% higher resolution than others
systems because the desktop font is effectively fixed size and quite large.

---druck
Mr John FO Evans
2009-11-11 23:06:55 UTC
Permalink
In article <hdcn1f$rq6$***@news.eternal-september.org>, druck
<***@druck.org.uk> wrote:
> Mr John FO Evans wrote:
> > Unless of course you are working under RISC OS which uses space
> > efficiently for most applications and doesn't need enormous
> > resolution to see even small text due to its anti-aliasing.
>
> I find I have to use RISC Os at least 25% higher resolution than others
> systems because the desktop font is effectively fixed size and quite
> large.

Yes I agree that changing the font size without changing the layout is a bit
pointless. However the standard font is very good for those with impaired
vision and I personally find I need more screen workspace in windows because
the (standard) directory presentation uses so much of it.

I suspect that all our preferences depend largely on the type of work we do.

I recently spent two or three hours trying to stop MS Word displaying only
the bottom half of lines of text on the screen (white on dark ground) - then
finding that the document printed correctly even though it was almost
impossible to edit. So far that has not happened to me under RISC OS.

John


--
_ _________________________________________
/ \._._ |_ _ _ /' Orpheus Internet Services
\_/| |_)| |(/_|_|_> / 'Internet for Everyone'
_______ | ___________./ http://www.orpheusinternet.co.uk
David Boddie
2009-10-31 18:03:35 UTC
Permalink
On Saturday 31 October 2009 08:22, Alexander Ausserstorfer wrote:

> 3.) Linux: With KDE you will get almost the same desktop as with
> Windows. Programs often crash down (also Internet). Trying Open Office
> I need four times to get the work done because of crashs but the
> fourth resulted in a corrupt document file and all the work and time
> were lost.

Sorry to hear that. I found OpenOffice to be pretty flaky in many respects
before version 2.3.

Which version of KDE did you try?

> What is so good on these systems? I will keep both Acorn computers and
> invest further in them so long as possible.

The good thing about KDE (and other Linux-based desktops) is the openness
of the development process. You can file bugs against the applications and
system components. OK, so people may or may not fix them, but at least
there's a certain level of transparency in the process.

I've been happily using KDE for about seven years now. At some point I'll
probably start using KDE 4 instead of KDE 3.

David
Alexander Ausserstorfer
2009-11-07 04:58:57 UTC
Permalink
In der Nachricht
<hchu5n$9v6$***@get-news01.get.basefarm.net>
David Boddie <***@boddie.org.uk>
hat geschrieben:

> On Saturday 31 October 2009 08:22, Alexander Ausserstorfer wrote:
>
> > 3.) Linux: With KDE you will get almost the same desktop as with
> > Windows. Programs often crash down (also Internet). Trying Open Office
> > I need four times to get the work done because of crashs but the
> > fourth resulted in a corrupt document file and all the work and time
> > were lost.
>
> Sorry to hear that. I found OpenOffice to be pretty flaky in many respects
> before version 2.3.
>
> Which version of KDE did you try?

2.38, I think. I'm not sure at the moment (the machines are far away at
the moment). Is it outdated?

> > What is so good on these systems? I will keep both Acorn computers and
> > invest further in them so long as possible.
>
> The good thing about KDE (and other Linux-based desktops) is the openness
> of the development process. You can file bugs against the applications and
> system components. OK, so people may or may not fix them, but at least
> there's a certain level of transparency in the process.

I can understand your argument, though I think it is irrelevant to the
user who knows nothing in programming.

> I've been happily using KDE for about seven years now. At some point I'll
> probably start using KDE 4 instead of KDE 3.

Glad to hear that.

Alex'

--
E-Mail: bavariasound [ at ] chiemgau-net [ point] de
PGP (key): http://home.chiemgau-net.de/ausserstorfer/kontakt/key
Web: http://home.chiemgau-net.de/ausserstorfer/
David Boddie
2009-11-07 20:02:02 UTC
Permalink
On Saturday 07 November 2009 05:58, Alexander Ausserstorfer wrote:

> In der Nachricht <hchu5n$9v6$***@get-news01.get.basefarm.net> David Boddie
> hat geschrieben:

>> Sorry to hear that. I found OpenOffice to be pretty flaky in many
>> respects before version 2.3.
>>
>> Which version of KDE did you try?
>
> 2.38, I think. I'm not sure at the moment (the machines are far away at
> the moment). Is it outdated?

If you mean KDE 2.38, then yes. If you mean OpenOffice 2.38, then it's not
that old.

>> The good thing about KDE (and other Linux-based desktops) is the openness
>> of the development process. You can file bugs against the applications
>> and system components. OK, so people may or may not fix them, but at
>> least there's a certain level of transparency in the process.
>
> I can understand your argument, though I think it is irrelevant to the
> user who knows nothing in programming.

I think it depends on the project and its developers. Some projects could
really use input from users, either in terms of help with usability or with
documentation and resources. Filing bug reports is just one way to interact
with the developers.

David
Alexander Ausserstorfer
2009-11-09 06:12:56 UTC
Permalink
On 7 Nov., 21:02, David Boddie <***@boddie.org.uk> wrote:
> On Saturday 07 November 2009 05:58, Alexander Ausserstorfer wrote:

[snip]

> >> Which version of KDE did you try?
>
> > 2.38, I think. I'm not sure at the moment (the machines are far away at
> > the moment). Is it outdated?
>
> If you mean KDE 2.38, then yes. If you mean OpenOffice 2.38, then it's not
> that old.

It's OpenOffice 2.3.

I'm currently checking (for six weeks now) different Windows, Linux
and Apple systems / computers for daily work but don't find them very
usable. I cannot explain this and won't buy one of these systems.

Alex'
Mr John FO Evans
2009-10-31 22:38:13 UTC
Permalink
In article <***@invalid.org.uk>, Tim Hill <***@invalid.org.uk>
wrote:
> In my eyes, all manufacturers
> of gadgets and computers are all completely irresponsible; using our
> scarce resources to make stuff which chews more energy. We're
> irresponsible for fuelling that industrybbc!!

How about the new Beagle Board which can be powered from a USB connection
and provides a 600MHz (possibly 1.2GHz in the future).

Couple that with a micro projector using a laser diode (when it gets to SVGA
it is VGA at the moment) and you have a pretty low powered system.

John

--
_ _________________________________________
/ \._._ |_ _ _ /' Orpheus Internet Services
\_/| |_)| |(/_|_|_> / 'Internet for Everyone'
_______ | ___________./ http://www.orpheusinternet.co.uk
Tim Hill
2009-11-01 12:41:11 UTC
Permalink
In article <***@orpheusmail.co.uk>, Mr John FO
Evans <***@orpheusmail.co.uk> wrote:
> In article <***@invalid.org.uk>, Tim Hill
> <***@invalid.org.uk> wrote:
> > In my eyes, all manufacturers of gadgets and computers are all
> > completely irresponsible; using our scarce resources to make stuff
> > which chews more energy. We're irresponsible for fuelling that
> > industrybbc!!

> How about the new Beagle Board which can be powered from a USB
> connection and provides a 600MHz (possibly 1.2GHz in the future).

> Couple that with a micro projector using a laser diode (when it gets to
> SVGA it is VGA at the moment) and you have a pretty low powered system.

It's not just the consumption, it's the resources used in development and
manufacture to get to where we are now. Starting with the Spinning Jenny.
;-)

--
Have an ethical share in cheaper telecoms: http://tinyurl.com/phone-coop
Genuine spam-proof addresses for Usenet: http://www.invalid.org.uk/
Email address for replies: substitute postmaster@ for tim@

... "Woman will love her, that she is a woman more worth than any man; men, that she is the rarest of all women" Win Tale, Act v, Sc.1
Alexander Ausserstorfer
2009-11-07 04:54:13 UTC
Permalink
In der Nachricht
<***@orpheusmail.co.uk>
Mr John FO Evans <***@orpheusmail.co.uk>
hat geschrieben:

> In article <***@invalid.org.uk>, Tim Hill <***@invalid.org.uk>
> wrote:
> > In my eyes, all manufacturers
> > of gadgets and computers are all completely irresponsible; using our
> > scarce resources to make stuff which chews more energy. We're
> > irresponsible for fuelling that industrybbc!!
>
> How about the new Beagle Board which can be powered from a USB connection
> and provides a 600MHz (possibly 1.2GHz in the future).

I want to buy a complette computer which does the job - not just a
"board".

[snip]

Alex'

--
E-Mail: bavariasound [ at ] chiemgau-net [ point] de
PGP (key): http://home.chiemgau-net.de/ausserstorfer/kontakt/key
Web: http://home.chiemgau-net.de/ausserstorfer/
Stuart
2009-11-07 07:56:14 UTC
Permalink
In article <***@chiemgau-net.de>,
Alexander Ausserstorfer <***@usenet.cnntp.org> wrote:
> > How about the new Beagle Board which can be powered from a USB
> > connection and provides a 600MHz (possibly 1.2GHz in the future).

> I want to buy a complette computer which does the job - not just a
> "board".

But the mere fact that people are working on such a project gives hope for
the future. They are doing an absolutly brilliant job and I have nothing
but praise for them.

If the time comes when I can by a board, load RO and some applications and
run them, I will be happy.

If I have to fabricate my own enclosure, I have no problems with that.
Alexander Ausserstorfer
2009-11-13 06:18:50 UTC
Permalink
On 31 Okt., 23:38, Mr John FO Evans <***@orpheusmail.co.uk> wrote:
> In article <***@invalid.org.uk>, Tim Hill <***@invalid.org.uk>
> wrote:

[snip]

> How about the new Beagle Board which can be powered from a USB connection
> and provides a 600MHz (possibly 1.2GHz in the future).

Distributation / shops for Germany / Europa? From where can I buy it?

Alex'
Rob Kendrick
2009-11-13 09:24:06 UTC
Permalink
On Thu, 12 Nov 2009 22:18:50 -0800 (PST)
Alexander Ausserstorfer <***@chiemgau-net.de> wrote:

> On 31 Okt., 23:38, Mr John FO Evans <***@orpheusmail.co.uk> wrote:
> > In article <***@invalid.org.uk>, Tim Hill
> > <***@invalid.org.uk> wrote:
>
> [snip]
>
> > How about the new Beagle Board which can be powered from a USB
> > connection and provides a 600MHz (possibly 1.2GHz in the future).
>
> Distributation / shops for Germany / Europa? From where can I buy it?

Digikey. Consult Google.

B.
Loading...