[tex-live] ICC profiles for PDF/A compliance

[Mojca Miklavec]

Hi,

I would like to ask whether it would be acceptable to include some ICC
color profiles from ghoscript (GS is GPL-licenced, so I assume those
files fall under the same licence) into TeX Live which are needed for
PDF/A compliance of PDF documents.

The files from GhostScript which are needed are:
- default_cmyk.icc
- default_gray.icc
- default_rgb.icc
plus one more file from http://www.eci.org/en/downloads

A while ago there was a discussion on this mailing list as well:
https://www.tug.org/pipermail/tex-live/2013-December/034625.html

What would have to be done to make these profiles a viable inclusion
for TeX Live?

Thank you very much,
Mojca

[Reinhard Kotucha]

On 2018-09-05 at 19:11:54 +0200, Mojca Miklavec wrote:

> Hi,
>
> I would like to ask whether it would be acceptable to include some ICC
> color profiles from ghoscript (GS is GPL-licenced, so I assume those
> files fall under the same licence) into TeX Live which are needed for
> PDF/A compliance of PDF documents.
>
> The files from GhostScript which are needed are:
> - default_cmyk.icc
> - default_gray.icc
> - default_rgb.icc
> plus one more file from http://www.eci.org/en/downloads
>
> A while ago there was a discussion on this mailing list as well:
> https://www.tug.org/pipermail/tex-live/2013-December/034625.html
>
> What would have to be done to make these profiles a viable inclusion
> for TeX Live?

Hi Mojca,
the pdfx LaTeX package already provides an ICC profile

sRGB_IEC61966-2-1_black_scaled.icc

from http://color.org . AFAICS this is sufficient to create valid
PDF/A.

Regards,
Reinhard

--
------------------------------------------------------------------
Reinhard Kotucha Phone: [phone]
Marschnerstr. 25
D-30167 Hannover mailto:***@web.de
------------------------------------------------------------------

[Mojca Miklavec]

Dear Reinhard,

On Wed, 5 Sep 2018 at 20:53, Reinhard Kotucha wrote:
> On 2018-09-05 at 19:11:54 +0200, Mojca Miklavec wrote:
>
> > Hi,
> >
> > I would like to ask whether it would be acceptable to include some ICC
> > color profiles from ghoscript (GS is GPL-licenced, so I assume those
> > files fall under the same licence) into TeX Live which are needed for
> > PDF/A compliance of PDF documents.
> >
> > The files from GhostScript which are needed are:
> > - default_cmyk.icc
> > - default_gray.icc
> > - default_rgb.icc
> > plus one more file from http://www.eci.org/en/downloads
> >
> > A while ago there was a discussion on this mailing list as well:
> > https://www.tug.org/pipermail/tex-live/2013-December/034625.html
> >
> > What would have to be done to make these profiles a viable inclusion
> > for TeX Live?
>
> Hi Mojca,
> the pdfx LaTeX package already provides an ICC profile
>
> sRGB_IEC61966-2-1_black_scaled.icc
>
> from http://color.org . AFAICS this is sufficient to create valid
> PDF/A.

Thanks a lot for pointing this out.

It doesn't help though that this file is well hidden behind the "latex" prefix.

Mojca

[David Carlisle]

> It doesn't help though that this file is well hidden behind the "latex" prefix.
>
> Mojca

If there was a directory set up somewhere under tex/generic that had
this and possibly other icc profile files I would be surprised if the
pdfx package maintainers would object to stopping including a copy
with the package to avoid having duplicate files in the input tree.


David
(not that I have any right to speak for the pdfx maintainers:-)

[Mojca Miklavec]

On Wed, 5 Sep 2018 at 21:36, David Carlisle wrote:
> > It doesn't help though that this file is well hidden behind the "latex" prefix.
>
> If there was a directory set up somewhere under tex/generic that had
> this and possibly other icc profile files I would be surprised if the
> pdfx package maintainers would object to stopping including a copy
> with the package to avoid having duplicate files in the input tree.
>
> David
> (not that I have any right to speak for the pdfx maintainers:-)

There was also an idea to create a separate directory
<texmf>/colors/<packagename>, in line with "fonts" as a special type
of files.

Mojca

[Reinhard Kotucha]

On 2018-09-05 at 20:36:00 +0100, David Carlisle wrote:

> > It doesn't help though that this file is well hidden behind the
> > "latex" prefix.
> >
> > Mojca
>
> If there was a directory set up somewhere under tex/generic that had
> this and possibly other icc profile files I would be surprised if the
> pdfx package maintainers would object to stopping including a copy
> with the package to avoid having duplicate files in the input tree.
>
>
> David
> (not that I have any right to speak for the pdfx maintainers:-)

The current maintainer of pdfx is Ross Moore. He already asked before
he added the profile to his package. Since there was no place for ICC
profiles in the TDS specification and Ross was the only one who needed
them, maintaining the profiles wihtin his package was the most obvious
and easiest solution.

The file isn't hidden behind the "latex" prefix at all. See the
TEXINPUTS entries in texmf.cnf. For plain TeX we have, for example:

TEXINPUTS.tex = .;$TEXMF/tex/{plain,generic,}//

You see that the third element between the braces is empty. This
means that when nothing is found in $TEXMF/tex/plain and
$TEXMF/tex/generic, kpathsea scans $TEXMF/tex, which undoubtedly
contains $TEXMF/tex/latex.


On 2018-09-05 at 22:33:31 +0200, Mojca Miklavec wrote:

> There was also an idea to create a separate directory
> <texmf>/colors/<packagename>, in line with "fonts" as a special
> type of files.

This only makes sense if TeX Live can provide many PDF/X color
profiles as well.

A PDF/X color profile specifies how a particular document is printed,
which type of paper has to be used, etc. It's sole purpose is to
allow the printer to blame the customer if something goes wrong.

There are not many PDF/X profiles available with appropriate licenses
for TeX Live.

Adobe offers a bunch of PDF/X color profiles which can be used freely
but I doubt that their licenses are appropriate for TeX Live. I also
doubt that printers don't accept anything else but the files provided
by Adobe because they are familiar with them. Please note that ICC
profiles are binary files and you can't easily find out what they
actually do.

TeX users who need PDF/X color profiles should download the
appropriate files from Adobe and maintain them in their current
working directory, TEXMFHOME, or TEXMFLOCAL.

At the moment I don't see a good reason to create a new directory
which only contains two files. Don't know what happens in the future.
Adobe recently released a lot of formerly proprietary files and even a
new font family with acceptable license conditions.

If Adobe decides to put their PDF/X color profiles under a reasonable
license, we can evaluate whether a distinct directory for such files
is sensible. But I don't see any reason to change anything now.

Regards,
Reinhard


--
------------------------------------------------------------------
Reinhard Kotucha Phone: [phone]
Marschnerstr. 25
D-30167 Hannover mailto:***@web.de
------------------------------------------------------------------

[Karl Berry]

<texmf>/colors/<packagename>

I don't want to make a new top-level directory unless there is some
reason why it is absolutely necessary (clearly not, since we've gotten
along with it until now, context notwithstanding).

Is it so hard to type "locate .icc" or "find . -type f -name \*icc" or
"fgrep .icc /whatever/texlive.tlpdb" if one needs to do a brute force
search? I fail to see the problem we are trying to solve.

[Karl Berry]

GS is GPL-licenced, so I assume those files fall under the same licence

I hope so, but it's not a safe assumption. It would be nice to have more
.icc files.

- default_cmyk.icc
- default_gray.icc
- default_rgb.icc
plus one more file from http://www.eci.org/en/downloads

Is there any information anywhere on the licensing of those specific
files, rather than just being in the same tarball?

What would have to be done to make these profiles a viable inclusion
for TeX Live?

Lacking specific license information, ask the GS/ECI people to make a
statement about it.

Also, as you know, upload a package to CTAN with them. But that's the
easy part. -k

[Norbert Preining]

HI Karl, hi Mojca, hi all,

a few comments from the Debian side:

sRGB_IEC61966-2-1_black_scaled.icc is non-free and should be removed
from TeX Live.

See the long thread of discussion about that here
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=786946
and the final clarification of the license in 2017.

I know I should have written about it earlier, but it completely dropped
out of my memory.

We should inform Ross that he removes/replaces the files in his package,
and the file should be removed from TeX Live.

> GS is GPL-licenced, so I assume those files fall under the same licence
>
> I hope so, but it's not a safe assumption. It would be nice to have more
> .icc files.

I agree with Karl here. Too much black voodoo and huge assumptions,
including wrong statements on websites.

Unless there is a clear statement of ECI I would not include any of
them.

Best

Norbert

--
PREINING Norbert http://www.preining.info
Accelia Inc. + JAIST + TeX Live + Debian Developer
GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13

[Mojca Miklavec]

On Wed, 5 Sep 2018 at 23:10, Karl Berry wrote:
>
> GS is GPL-licenced, so I assume those files fall under the same licence
>
> I hope so, but it's not a safe assumption.

I'll try to dig further, but given that Debian already packages those
files, I assume there is a reasonable chance that they would be free
enough

> It would be nice to have more
> .icc files.

Unfortunately a lot of those files might be non-free.

> - default_cmyk.icc
> - default_gray.icc
> - default_rgb.icc
> plus one more file from http://www.eci.org/en/downloads
>
> Is there any information anywhere on the licensing of those specific
> files, rather than just being in the same tarball?

I'll try to dig further.
Ironically a GS update just broke asymptote for me due to apparent
inability to find default_gray.icc.

> What would have to be done to make these profiles a viable inclusion
> for TeX Live?
>
> Lacking specific license information, ask the GS/ECI people to make a
> statement about it.

I just did.

Thank you very much,
Mojca

[Karl Berry]

sRGB_IEC61966-2-1_black_scaled.icc is non-free and should be removed
from TeX Live.

Ok, done. (I wonder if http://argyllcms.com/ has any other free profiles.)

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=786946
...
and the final clarification of the license in 2017.

I guess at http://www.color.org/srgbprofiles.xalter,
which says the file may not be changed.

We should inform Ross that he removes/replaces the files in [pdfx]

.. and update the code and doc, which now use the sRGB profile.
Cc-ing Ross.

I hope the gs profiles are free. -k

[Ross Moore]

Hi Karl, Norbert, Peter,

On 7 Sep 2018, at 7:19 am, Karl Berry <***@freefriends.org<mailto:***@freefriends.org>> wrote:

sRGB_IEC61966-2-1_black_scaled.icc is non-free and should be removed
from TeX Live.

Ok, done. (I wonder if http://argyllcms.com/<https://protect-au.mimecast.com/s/-HlJClx1OYUo9g8quqbkbq?domain=argyllcms.com> has any other free profiles.)

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=786946<https://protect-au.mimecast.com/s/GQANCmO5wZs50E3LcQH9Vl?domain=bugs.debian.org>



This page says that the ICC profiles *are* free.

Debian Bug report logs - #786946
lintian: false positive: icc-profiles *are* free

But that was back in 2015; then you were arguing that they are free.


and the final clarification of the license in 2017.

OK. There is a single-line statement saying: "not free"
without any explanation of why Debian has determined this.


I guess at http://www.color.org/srgbprofiles.xalter<https://protect-au.mimecast.com/s/l0kuCnx1Z5UGyM2psEdNnH?domain=color.org>,
which says the file may not be changed.

It is not changed, so far as I’m aware;
pdfTeX just includes the binary file verbatim.
Not sure about XeTeX and LuaTeX, but I cannot imagine why
they would change anything about it.



We should inform Ross that he removes/replaces the files in [pdfx]

.. and update the code and doc, which now use the sRGB profile.
Cc-ing Ross.

If this is going to stop Debian from distributing TeXLive, then I suppose
we should change the profile.

Can we switch to: sRGB2014.icc from
http://www.color.org/srgbprofiles.xalter ?
Or will that be subject to the same objection?

What about the updated v4 profiles, on that same page?

sRGB_v4_ICC_preference.icc
sRGB_v4_ICC_preference_displayclass.icc

Their licensing seems to be a bit more liberal than before;
viz. from http://www.color.org/profiles2.xalter#license

Licensing

The copyright owner and terms of use of an ICC profile are normally identified in the Creator field in the profile header and in the Copyright tag. Where ICC is the copyright owner, the following license terms apply:

"This profile is made available by the International Color Consortium, and may be copied, distributed, embedded, made, used, and sold without restriction. Altered versions of this profile shall have the original identification and copyright information removed and shall not be misrepresented as the original profile."

ICC recommends that other profile creators and copyright owners adopt a similar wording for profiles that are intended to be freely distributed. See the Profile Registration page for more details.


On the other hand, sRGB_IEC61966-2-1_black_scaled.icc
is still available at http://www.color.org/black_scaled_2009_srgb.xalter
where it says:

Legacy v2 sRGB profile

This sRGB profile has been superseded by the sRGB2014.icc profile, following a 2014 corrigendum to IEC 61966. This previous version of the profile is made available for those who need it for compatibility with existing workflows.
Note that this version of the profile has different licensing terms from the standard ICC profile license

License

To anyone who acknowledges that the file "sRGB_IEC61966-2-1_no_black_scaling.icc" is provided "AS IS" WITH NO EXPRESS OR IMPLIED WARRANTY, permission to use, copy and distribute these file for any purpose is hereby granted without fee, provided that the file is not changed including the ICC copyright notice tag, and that the name of ICC shall not be used in advertising or publicity pertaining to distribution of the software without specific, written prior permission. ICC makes no representations about the suitability of this software for any purpose.


So it would seem:
sRGB2014.icc is free
sRGB_IEC61966-2-1_black_scaled.icc is not free.

Agreed?


And what about the CMYK profile? coated_FOGRA39L_argl.icc
Has an objection to this ever been raised?

That file is no longer available at: http://www.color.org/registry/index.xalter
which is where we got it from, I think.
The licensing on the files available there now says things like:

Profile Name: PSOuncoated_v3_FOGRA52.icc

Profile Provider: European Color Initiative (ECI)

Contact: Florian Suessl

Copyright: Heidelberger Druckmaschinen AG

License: This profile is made available by ECI European Color Initiative, with permission of Heidelberger Druckmaschinen AG, and may be used, embedded and exchanged without restriction. It may not be distributed, sold or altered without written permission of ECI European Color Initiative. Color Toolbox 17.0.0 - (c) Copyright 2015 Heidelberger Druckmaschinen AG. All Rights Reserved.

So these would be classed as “not free” right?
Is it reasonable for TeXLive to request permission to distribute
from some of these places: ECI, IDEAlliance, VIGC, 
 ?
But if granted, would Debian be able to re-distribute?



I hope the gs profiles are free. -k

Please clarify, and give a clear reason why you think there is a violation
by including these files.


Alternatively, there is this site:

https://sourceforge.net/projects/openicc/files/OpenICC-Profiles/

Worth investigating?



Cheers,

Ross

[Norbert Preining]

Hi Ross, Peter, Karl,

first of all, I have to agree it is all very convoluted, and I am not an
expert, but here is what I understood and what I think was the timeline:

* at some point in the past all the icc profiles somehow shared the same
license statement (or it wasn't made clear on profiles2.xalter
This is probably when Peter included the _black_sealed one.

* Debian hat a lintian error about this being non-free, and I was stymed
because the license statement was sufficiently free for me (including
a renaming clause but that is fine).
The reason for this is that in the license statement, as of now, there
is no *explicite* clause stating that changes are allowed (a
requirement for OSS).
But it is stated that *altered* versions are to be renamed etc.

* I asked clarification from Debian ftp-masters and they agreed that the
intention is that changes are allowed with the renaming restrictions,
and thus the lintian error should be changed.

* Then it turned out that on different parts of the color.org page were
different license statements, and after clarifications it turned out
that the _black_sealed is non-free.

That brings me to Ross' analysis, with which I agree:
> So it would seem:
> sRGB2014.icc is free
> sRGB_IEC61966-2-1_black_scaled.icc is not free.

This is what I would say, too.

Checking with Debian, the sRGB2015.icc is in the package icc-profiles,
which is in non-free, and the license statement attributed to the file
is recorded as
License: ICC
This profile is made available by the International Color Consortium,
and may be copied, distributed, embedded, made, used, and sold without
restriction. Altered versions of this profile shall have the original
identification and copyright information removed and shall not be
misrepresented as the original profile.
Comment:
Apparently applies generally to color profiles with ICC listed as
copyright holder in the embedded Creator field, according to
<http://www.color.org/profiles2.xalter#license>.
So with the ftp-master decision I think it should actually moved into
the icc-profile-free package. I will report there and see if I can move
somthing forward, but this is in fact unrelated to TeX Live!!!

Now a few more comments:

On Fri, 07 Sep 2018, Ross Moore wrote:
> OK. There is a single-line statement saying: "not free"
> without any explanation of why Debian has determined this.

Because as you found, the _black_sealed has different license terms.

> It is not changed, so far as I’m aware;

Irrelevant. What is necessary is the permission to change ;-)

> If this is going to stop Debian from distributing TeXLive, then I suppose
> we should change the profile.

No no no. I have for now removed the file in the Debian package and
suggested the icc-profiles package to get it back, so no worry on the
Debian side.

Anyway, in all these discussion I normally *completely* ignore my Debian
hat and decide for TeX Live, the rest must be managed from Debian in
some way (I guess I'm a bit schizophrenic ;-)

> And what about the CMYK profile? coated_FOGRA39L_argl.icc
> Has an objection to this ever been raised?

Not that I know.

> License: This profile is made available by ECI European Color Initiative, with permission of Heidelberger Druckmaschinen AG, and may be used, embedded and exchanged without restriction. It may not be distributed, sold or altered without written permission of ECI European Color Initiative. Color Toolbox 17.0.0 - (c) Copyright 2015 Heidelberger Druckmaschinen AG. All Rights Reserved.

> So these would be classed as “not free” right?

Right, no way to have it in TeX Live.

> Is it reasonable for TeXLive to request permission to distribute

No, because we need software/files that follow the FSF guidelines, and
that means that any user of the files can use the obtained files freely
(not going into details here now), which wouldn't be achieved.

> from some of these places: ECI, IDEAlliance, VIGC, … ?
> But if granted, would Debian be able to re-distribute?

No, out of the same reasons. And TeX Live also cannot use them.
The TeX Live DVDs are sold by Lehmann which would be in breach
of the copyright agreement.

> I hope the gs profiles are free. -k
>
> Please clarify, and give a clear reason why you think there is a violation
> by including these files.

The question is always whether there is a clear license statement
allowing for free and unrestricted redistribution within the FSF
guilelines.

All the best

Norbert

--
PREINING Norbert http://www.preining.info
Accelia Inc. + JAIST + TeX Live + Debian Developer
GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13

[Peter Selinger]

Hi Ross,

when I added the files sRGB_IEC61966-2-1_black_scaled.icc and
coated_FOGRA39L_argl.icc to the pdfx package in 2015, they came with
the following respective licenses, which are included in the package
(in the file ICC_LICENSE.txt):

----------------------------------------------------------------------
For the file sRGB_IEC61966-2-1_black_scaled.icc:

Copyright International Color Consortium, 2009

It is hereby acknowledged that the file "sRGB_IEC61966-2-1_black
scaled.icc" is provided "AS IS" WITH NO EXPRESS OR IMPLIED WARRANTY.

Licensing

This profile is made available by the International Color Consortium,
and may be copied, distributed, embedded, made, used, and sold without
restriction. Altered versions of this profile shall have the original
identification and copyright information removed and shall not be
misrepresented as the original profile.

Terms of use

To anyone who acknowledges that the file "sRGB_IEC61966-2-1_black
scaled.icc" is provided "AS IS" WITH NO EXPRESS OR IMPLIED WARRANTY,
permission to use, copy and distribute these file for any purpose is
hereby granted without fee, provided that the file is not changed
including the ICC copyright notice tag, and that the name of ICC shall
not be used in advertising or publicity pertaining to distribution of
the software without specific, written prior permission. ICC makes no
representations about the suitability of this software for any
purpose.

For the file coated_FOGRA39L_argl.icc:

The zlib/libpng License

Copyright (c) 2008 Kai-Uwe Behrmann

This software is provided 'as-is', without any express or implied
warranty. In no event will the authors be held liable for any damages
arising from the use of this software.

Permission is granted to anyone to use this software for any purpose,
including commercial applications, and to alter it and redistribute
it freely, subject to the following restrictions:

1. The origin of this software must not be misrepresented; you
must not claim that you wrote the original software. If you use
this software in a product, an acknowledgment in the product
documentation would be appreciated but is not required.

2. Altered source versions must be plainly marked as such, and
must not be misrepresented as being the original software.

3. This notice may not be removed or altered from any source
distribution.
----------------------------------------------------------------------

I don't know on what basis someone now claims to have determined that
sRGB_IEC61966-2-1_black_scaled.icc is "not free". The license above is
the one that came with the file when I downloaded it from the
International Color Consortium's website, and it is explicitly
included in the package so the Debian developer should have seen
it. Such licenses are not revocable, so the file is free.

-- Peter

Ross Moore wrote:
>
>
> Hi Karl, Norbert, Peter,
>
> On 7 Sep 2018, at 7:19 am, Karl Berry <***@freefriends.org<mailto:***@freefriends.org>> wrote:
>
> sRGB_IEC61966-2-1_black_scaled.icc is non-free and should be removed
> from TeX Live.
>
> Ok, done. (I wonder if http://argyllcms.com/<https://protect-au.mimecast.com/s/-HlJClx1OYUo9g8quqbkbq?domain=argyllcms.com> has any other free profiles.)
>
> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=786946<https://protect-au.mimecast.com/s/GQANCmO5wZs50E3LcQH9Vl?domain=bugs.debian.org>
> …
>
> This page says that the ICC profiles *are* free.
>
> Debian Bug report logs - #786946
> lintian: false positive: icc-profiles *are* free
>
> But that was back in 2015; then you were arguing that they are free.
>
>
> and the final clarification of the license in 2017.
>
> OK. There is a single-line statement saying: "not free"
> without any explanation of why Debian has determined this.
>
>
> I guess at http://www.color.org/srgbprofiles.xalter<https://protect-au.mimecast.com/s/l0kuCnx1Z5UGyM2psEdNnH?domain=color.org>,
> which says the file may not be changed.
>
> It is not changed, so far as I’m aware;
> pdfTeX just includes the binary file verbatim.
> Not sure about XeTeX and LuaTeX, but I cannot imagine why
> they would change anything about it.
>
>
>
> We should inform Ross that he removes/replaces the files in [pdfx]
>
> .. and update the code and doc, which now use the sRGB profile.
> Cc-ing Ross.
>
> If this is going to stop Debian from distributing TeXLive, then I suppose
> we should change the profile.
>
> Can we switch to: sRGB2014.icc from
> http://www.color.org/srgbprofiles.xalter ?
> Or will that be subject to the same objection?
>
> What about the updated v4 profiles, on that same page?
>
> sRGB_v4_ICC_preference.icc
> sRGB_v4_ICC_preference_displayclass.icc
>
> Their licensing seems to be a bit more liberal than before;
> viz. from http://www.color.org/profiles2.xalter#license
>
> Licensing
>
> The copyright owner and terms of use of an ICC profile are normally identified in the Creator field in the profile header and in the Copyright tag. Where ICC is the copyright owner, the following license terms apply:
>
> "This profile is made available by the International Color Consortium, and may be copied, distributed, embedded, made, used, and sold without restriction. Altered versions of this profile shall have the original identification and copyright information removed and shall not be misrepresented as the original profile."
>
> ICC recommends that other profile creators and copyright owners adopt a similar wording for profiles that are intended to be freely distributed. See the Profile Registration page for more details.
>
>
> On the other hand, sRGB_IEC61966-2-1_black_scaled.icc
> is still available at http://www.color.org/black_scaled_2009_srgb.xalter
> where it says:
>
> Legacy v2 sRGB profile
>
> This sRGB profile has been superseded by the sRGB2014.icc profile, following a 2014 corrigendum to IEC 61966. This previous version of the profile is made available for those who need it for compatibility with existing workflows.
> Note that this version of the profile has different licensing terms from the standard ICC profile license
>
> License
>
> To anyone who acknowledges that the file "sRGB_IEC61966-2-1_no_black_scaling.icc" is provided "AS IS" WITH NO EXPRESS OR IMPLIED WARRANTY, permission to use, copy and distribute these file for any purpose is hereby granted without fee, provided that the file is not changed including the ICC copyright notice tag, and that the name of ICC shall not be used in advertising or publicity pertaining to distribution of the software without specific, written prior permission. ICC makes no representations about the suitability of this software for any purpose.
>
>
> So it would seem:
> sRGB2014.icc is free
> sRGB_IEC61966-2-1_black_scaled.icc is not free.
>
> Agreed?
>
>
> And what about the CMYK profile? coated_FOGRA39L_argl.icc
> Has an objection to this ever been raised?
>
> That file is no longer available at: http://www.color.org/registry/index.xalter
> which is where we got it from, I think.
> The licensing on the files available there now says things like:
>
> Profile Name: PSOuncoated_v3_FOGRA52.icc
>
> Profile Provider: European Color Initiative (ECI)
>
> Contact: Florian Suessl
>
> Copyright: Heidelberger Druckmaschinen AG
>
> License: This profile is made available by ECI European Color Initiative, with permission of Heidelberger Druckmaschinen AG, and may be used, embedded and exchanged without restriction. It may not be distributed, sold or altered without written permission of ECI European Color Initiative. Color Toolbox 17.0.0 - (c) Copyright 2015 Heidelberger Druckmaschinen AG. All Rights Reserved.
>
> So these would be classed as “not free” right?
> Is it reasonable for TeXLive to request permission to distribute
> from some of these places: ECI, IDEAlliance, VIGC, … ?
> But if granted, would Debian be able to re-distribute?
>
>
>
> I hope the gs profiles are free. -k
>
> Please clarify, and give a clear reason why you think there is a violation
> by including these files.
>
>
> Alternatively, there is this site:
>
> https://sourceforge.net/projects/openicc/files/OpenICC-Profiles/
>
> Worth investigating?
>
>
>
> Cheers,
>
> Ross

[Peter Selinger]

> * at some point in the past all the icc profiles somehow shared the same
> license statement (or it wasn't made clear on profiles2.xalter
> This is probably when Peter included the _black_sealed one.

No, this is not true. Unfortunately I don't have screen shots from
2015 to prove it, but I am always very careful about licenses. When I
downloaded the file sRGB_IEC61966-2-1_black_scaled.icc from ICC in
2015, I copied and pasted the exact license that came with it into the
file ICC_LICENSE.txt, and I did not alter the text of the license. The
beginning of the license explicitly states:

It is hereby acknowledged that the file "sRGB_IEC61966-2-1_black
scaled.icc" is provided "AS IS" WITH NO EXPRESS OR IMPLIED WARRANTY.

These words were not written by me. They were copied and pasted from
the license from color.org. And it explicitly refers to the
"sRGB_IEC61966-2-1_black scaled.icc" profile, so it was not some
blanket license somehow shared by all the icc profiles or
misattributed by me.

It is possible that ICC has decided some time after 2015 to distribute
that file under a different license. However, since I downloaded it in
2015, the terms from 2015 still apply, because a license, once granted,
cannot be retroactively changed.

-- Peter


Norbert Preining wrote:
>
> Hi Ross, Peter, Karl,
>
> first of all, I have to agree it is all very convoluted, and I am not an
> expert, but here is what I understood and what I think was the timeline:
>
> * at some point in the past all the icc profiles somehow shared the same
> license statement (or it wasn't made clear on profiles2.xalter
> This is probably when Peter included the _black_sealed one.
>
> * Debian hat a lintian error about this being non-free, and I was stymed
> because the license statement was sufficiently free for me (including
> a renaming clause but that is fine).
> The reason for this is that in the license statement, as of now, there
> is no *explicite* clause stating that changes are allowed (a
> requirement for OSS).
> But it is stated that *altered* versions are to be renamed etc.
>
> * I asked clarification from Debian ftp-masters and they agreed that the
> intention is that changes are allowed with the renaming restrictions,
> and thus the lintian error should be changed.
>
> * Then it turned out that on different parts of the color.org page were
> different license statements, and after clarifications it turned out
> that the _black_sealed is non-free.
>
> That brings me to Ross' analysis, with which I agree:
> > So it would seem:
> > sRGB2014.icc is free
> > sRGB_IEC61966-2-1_black_scaled.icc is not free.
>
> This is what I would say, too.
>
> Checking with Debian, the sRGB2015.icc is in the package icc-profiles,
> which is in non-free, and the license statement attributed to the file
> is recorded as
> License: ICC
> This profile is made available by the International Color Consortium,
> and may be copied, distributed, embedded, made, used, and sold without
> restriction. Altered versions of this profile shall have the original
> identification and copyright information removed and shall not be
> misrepresented as the original profile.
> Comment:
> Apparently applies generally to color profiles with ICC listed as
> copyright holder in the embedded Creator field, according to
> <http://www.color.org/profiles2.xalter#license>.
> So with the ftp-master decision I think it should actually moved into
> the icc-profile-free package. I will report there and see if I can move
> somthing forward, but this is in fact unrelated to TeX Live!!!
>
> Now a few more comments:
>
> On Fri, 07 Sep 2018, Ross Moore wrote:
> > OK. There is a single-line statement saying: "not free"
> > without any explanation of why Debian has determined this.
>
> Because as you found, the _black_sealed has different license terms.
>
> > It is not changed, so far as I’m aware;
>
> Irrelevant. What is necessary is the permission to change ;-)
>
> > If this is going to stop Debian from distributing TeXLive, then I suppose
> > we should change the profile.
>
> No no no. I have for now removed the file in the Debian package and
> suggested the icc-profiles package to get it back, so no worry on the
> Debian side.
>
> Anyway, in all these discussion I normally *completely* ignore my Debian
> hat and decide for TeX Live, the rest must be managed from Debian in
> some way (I guess I'm a bit schizophrenic ;-)
>
> > And what about the CMYK profile? coated_FOGRA39L_argl.icc
> > Has an objection to this ever been raised?
>
> Not that I know.
>
> > License: This profile is made available by ECI European Color Initiative, with permission of Heidelberger Druckmaschinen AG, and may be used, embedded and exchanged without restriction. It may not be distributed, sold or altered without written permission of ECI European Color Initiative. Color Toolbox 17.0.0 - (c) Copyright 2015 Heidelberger Druckmaschinen AG. All Rights Reserved.
>
> > So these would be classed as “not free” right?
>
> Right, no way to have it in TeX Live.
>
> > Is it reasonable for TeXLive to request permission to distribute
>
> No, because we need software/files that follow the FSF guidelines, and
> that means that any user of the files can use the obtained files freely
> (not going into details here now), which wouldn't be achieved.
>
> > from some of these places: ECI, IDEAlliance, VIGC, … ?
> > But if granted, would Debian be able to re-distribute?
>
> No, out of the same reasons. And TeX Live also cannot use them.
> The TeX Live DVDs are sold by Lehmann which would be in breach
> of the copyright agreement.
>
> > I hope the gs profiles are free. -k
> >
> > Please clarify, and give a clear reason why you think there is a violation
> > by including these files.
>
> The question is always whether there is a clear license statement
> allowing for free and unrestricted redistribution within the FSF
> guilelines.
>
> All the best
>
> Norbert
>
> --
> PREINING Norbert http://www.preining.info
> Accelia Inc. + JAIST + TeX Live + Debian Developer
> GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13
>

[Norbert Preining]

Hi Peter,

thanks for your answer.

> downloaded the file sRGB_IEC61966-2-1_black_scaled.icc from ICC in
> 2015, I copied and pasted the exact license that came with it into the
> file ICC_LICENSE.txt, and I did not alter the text of the license. The

Interesting. They seem to have a big mess up there ;-)

> It is hereby acknowledged that the file "sRGB_IEC61966-2-1_black
> scaled.icc" is provided "AS IS" WITH NO EXPRESS OR IMPLIED WARRANTY.
>
> These words were not written by me. They were copied and pasted from

Well, these words are anyway irrelevant, what counts is the next
paragraph you included
Licensing

This profile is made available by the International Color Consortium,
and may be copied, distributed, embedded, made, used, and sold without
restriction. Altered versions of this profile shall have the original
identification and copyright information removed and shall not be
misrepresented as the original profile.

The current license statement as also written by Ross and found in the
Debian package is quite different:
To anyone who acknowledges that the files
"sRGB_IEC61966-2-1_no_black_scaling.icc" and
"sRGB_IEC61966-2-1_black scaled.icc" are provided "AS IS" WITH NO
EXPRESS OR IMPLIED WARRANTY, permission to use, copy and distribute
these file for any purpose is hereby granted without fee, provided
that the files are not changed including the ICC copyright notice
tag, and that the name of ICC shall not be used in advertising or
publicity pertaining to distribution of the software without
specific, written prior permission. ICC makes no representations
about the suitability of this software for any purpose.

In particular does not allow for changes.

> It is possible that ICC has decided some time after 2015 to distribute
> that file under a different license. However, since I downloaded it in
> 2015, the terms from 2015 still apply, because a license, once granted,
> cannot be retroactively changed.

Indeed, you are right in this respect, but we are probably having a hard
time arguing that.

I am not a laywer, and although I think it would be worth it, I as
private person will not stand up in this fight. If the FSF or whoever
wants to fight for a clarification, that would be great, but well...

Best

Norbert

--
PREINING Norbert http://www.preining.info
Accelia Inc. + JAIST + TeX Live + Debian Developer
GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13

[Reinhard Kotucha]

On 2018-09-07 at 10:34:23 +0900, Norbert Preining wrote:

> Interesting. They seem to have a big mess up there ;-)

Yes, but maybe I can provide more information.

First of all, the pdfx package was formerly maintained by Thành and
Radhakrishnan at Sarovar. There was a web page with HTML
documentation and a file "sRGBIEC1966-2.1.icm". See attachment.

[Zdenek Wagner]

ne 9. 9. 2018 v 1:02 odesílatel Reinhard Kotucha <***@web.de>
napsal:

...

>
> Presumably HP shipped this file on CDs accompanied by their printers
> without further licensing information.
>

I have several ICC profiles coming with printers, scanners etc without
explicit license.

>
> The data in the file below is freely available, see
>
> https://en.wikipedia.org/wiki/SRGB
>
> So if we don't find an ICC sRGB profile with an appropriate license we
> can create one ourselves with a license which makes everybody happy.
> All we need is a specification of the .icc binary format.
>

Little CMS (LCMS) is a free OpenSource tool, information on the .icc format
can probably be found there and the specification is hopefully freely
available.

Zdeněk Wagner
http://ttsm.icpf.cas.cz/team/wagner.shtml
http://icebearsoft.euweb.cz




>
> For the time being, Norbert, could you add the file
>
> sRGB_IEC61966-2-1_black_scaled.icc
>
> to tlcontrib if you haven't done this already?
>
> Regards,
> Reinhard
>
> =========================================================================
> % Copyright (C) 2001-2018 Artifex Software, Inc.
> % All Rights Reserved.
> %
> % This software is provided AS-IS with no warranty, either express or
> % implied.
> %
> % This software is distributed under license and may not be copied,
> % modified or distributed except as expressly authorized under the terms
> % of the license contained in the file LICENSE in this distribution.
> %
> % Refer to licensing information at http://www.artifex.com or contact
> % Artifex Software, Inc., 1305 Grant Avenue - Suite 200, Novato,
> % CA 94945, U.S.A., [phone], for further information.
> %
>
> % sRGB color space derived from the public domain HP sample.
>
> currentglobal true setglobal
>
> /sRGB
> [ /CIEBasedABC
> 3 dict begin
> /WhitePoint [ 0.9505 1 1.0890 ] def % D65
>
> /DecodeLMN [ % sRGB gamma transform
> { dup 0.03928 le { 12.92321 div } { 0.055 add 1.055 div 2.4 exp }
> ifelse }
> bind dup dup ] def
>
> /MatrixLMN [ % sRGB to XYZ (D65) matrix (ITU-R BT.709-2 Primaries)
> 0.412457 0.212673 0.019334
> 0.357576 0.715152 0.119192
> 0.180437 0.072175 0.950301 ] def
> currentdict end
> ]
> /ColorSpace defineresource pop
>
> setglobal
> =========================================================================
>
>
> --
> ------------------------------------------------------------------
> Reinhard Kotucha Phone: [phone]
> Marschnerstr. 25
> D-30167 Hannover mailto:***@web.de
> ------------------------------------------------------------------
>

[Reinhard Kotucha]

On 2018-09-09 at 11:02:51 +0200, Zdenek Wagner wrote:

> ne 9. 9. 2018 v 1:02 odesílatel Reinhard Kotucha <***@web.de>
> napsal:
>
> ...
>
> So if we don't find an ICC sRGB profile with an appropriate
> license we can create one ourselves with a license which makes
> everybody happy. All we need is a specification of the .icc
> binary format.
>
> Little CMS (LCMS) is a free OpenSource tool, information on the
> .icc format can probably be found there and the specification is
> hopefully freely available.

I found the specification at color.org. But it's quite complex and
requires a lot of knowlege about colorimetry.

I also found some very useful programs at color.org. Two of them are
very interesting: iccToXml and iccFromXml. Since ICC profiles are
binary files, the most reasonable way to create them is write a
program in C, as assumed by the specification. Everything is easier
with an XML file as a starting point.

The C sources of these tools are based on autoconf and compiling the
programs was quite easy on Linux.

It's very nice to be able to convert binary files to XML and vice versa.

On my Raspberry I just installed the Debian icc-profiles-free package
and now there is a file

/usr/share/color/icc/ghostscript/srgb.icc

If Debian considers ghostscript/srgb.icc as free, isn't it free enough
for TeX Live as well? The file contains

"Copyright Artifex Software 2011"

Is this license appropriate for TeX Live? I think so because
otherwise neither Debian nor TeX Live can provide Ghostscript.

Debian also provides a file

/usr/share/color/icc/sRGB.icc

which obviously comes from LCMS and contains

"no copyright, use freely"

Nicely said. Sounds like "public domain" at a first glance. But who
is the author and where does the stuff come from?

After all, I think that if all GNU/Linux distributions already provide
Ghostscript, there is no obvious reason not to ship a file which is
part of it with the pdfx package.

Regards,
Reinhard

--
------------------------------------------------------------------
Reinhard Kotucha Phone: [phone]
Marschnerstr. 25
D-30167 Hannover mailto:***@web.de
------------------------------------------------------------------

[Ross Moore]

Hi Reinhard, Karl and others.

On 11 Sep 2018, at 8:38 am, Reinhard Kotucha <***@web.de<mailto:***@web.de>> wrote:

On my Raspberry I just installed the Debian icc-profiles-free package
and now there is a file

/usr/share/color/icc/ghostscript/srgb.icc

If Debian considers ghostscript/srgb.icc as free, isn't it free enough
for TeX Live as well? The file contains

"Copyright Artifex Software 2011"

Is this license appropriate for TeX Live? I think so because
otherwise neither Debian nor TeX Live can provide Ghostscript.

Debian also provides a file

/usr/share/color/icc/sRGB.icc

which obviously comes from LCMS and contains

"no copyright, use freely"

Nicely said. Sounds like "public domain" at a first glance. But who
is the author and where does the stuff come from?

After all, I think that if all GNU/Linux distributions already provide
Ghostscript, there is no obvious reason not to ship a file which is
part of it with the pdfx package.

Probably best is to ship a copy of this (i.e., one of these – which one?)
in the 
/tex/latex/pdfx/ directory.

Another possibility is to specify the full path to it, where needed in
the source of pdfx.sty .
The problem with this is that it needs to work on all platforms, so the
syntax of the path (or the path itself) could be different.
Maybe Ghostscript isn’t even installed in some cases, since GS normally
isn’t needed when pdfTeX is the engine ?


But there is one aspect of this that I don’t have the expertise in color usage to decide.
How different are the various profiles?
What is the difference between them? Does it matter?
If we make a choice, then there should be a reason for it, concerning the actual
colour output, either on-screen or when printed with different devices.

It would be good to have some explanation included within the pdfx documentation,
addressing this issue. That is, saying when it is best to stick with the default,
and when it is advisable to apply a different Colour profile, perhaps with some
examples of when alternatives are either useful or needed.

Who can help with this?


Regards,
Reinhard

--
------------------------------------------------------------------
Reinhard Kotucha Phone: [phone]
Marschnerstr. 25
D-30167 Hannover mailto:***@web.de
—————————————————————————————————


Cheers.

Ross



Dr Ross Moore

Mathematics Dept | 12 Wally’s Walk, 734
Macquarie University, NSW 2109, Australia

T: [phone] | F: [phone]<tel:%2B61%202%209850%209695>
M:[phone]<tel:%2B61%20409%20125%20670> | E: ***@mq.edu.au<mailto:***@mq.edu.au>

http://www.maths.mq.edu.au<http://mq.edu.au/>


[cid:***@01D030BE.D37A46F0]<http://mq.edu.au/>




CRICOS Provider Number 00002J. Think before you print.
Please consider the environment before printing this email.<http://mq.edu.au/>

This message is intended for the addressee named and may
contain confidential information. If you are not the intended
recipient, please delete it and notify the sender. Views expressed
in this message are those of the individual sender, and are not
necessarily the views of Macquarie University.<http://mq.edu.au/>

[Peter Selinger]

Hi Ross and Norbert,

yes. Although my recollection is as I stated, I do not have any
evidence to corroborate it. So I agree with you that the best thing to
do is to replace it with another, free profile (and hopefully to save
some screen shots proving what the license actually is).

-- Peter

Ross Moore wrote:
>
>
> Hi Peter,
>
> I’d meant to send this earlier, but got distracted with something else
> (also PDF related). Some of it is redundant after Norbert’s messages,
> but here it is anyway. :-)
>
> On 7 Sep 2018, at 10:27 am, Peter Selinger <***@mathstat.dal.ca<mailto:***@mathstat.dal.ca>> wrote:
>
> Hi Ross,
>
> when I added the files sRGB_IEC61966-2-1_black_scaled.icc and
> coated_FOGRA39L_argl.icc to the pdfx package in 2015, they came with
> the following respective licenses, which are included in the package
> (in the file ICC_LICENSE.txt):
>
> Yes, I was going to reply the same way …
>
> … until I read the email discussion thread on the page that Norbert indicated.
> viz.
> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=786946<https://protect-au.mimecast.com/s/Mg4ICD1vRkCB6ZBLt85J5p?domain=bugs.debian.org>
>
> (I hope my email system doesn’t munge this for you, by redirecting
> via a spam-management system, called mimecast .
> Clicking the link may not work for you, so you have to copy/paste
> the link into the browser’s location bar.)
>
>
> In 2015, a bug was filed, saying these profiles are non-free.
> This was countered, and the bug retracted.
> But in 2017, after a clarification was requested (2 years earlier!)
> it was finally decided that, for Debian, the licence was non-free!!!
>
> This is the first I’ve heard of this decision, which I find very hard to understand.
>
> I don't know on what basis someone now claims to have determined that
> sRGB_IEC61966-2-1_black_scaled.icc is "not free". The license above is
> the one that came with the file when I downloaded it from the
> International Color Consortium's website, and it is explicitly
> included in the package so the Debian developer should have seen
> it. Such licenses are not revocable, so the file is free.
>
> Just guessing, but maybe one may quibble about how the file copy is
> done, either onto disk or as it gets included into a PDF file.
>
> In any case, while Debian regards it as non-free, it poisons TeXLive for them.
> That’s certainly a bad thing, whether we agree with the reason or not.
> Fixing that should not be too hard.
>
> There are a few bugs with the current pdfx.sty anyway, which have emerged
> recently. So I’ll have to produce a new version.
>
> Besides, I have some great new features, not yet released:
>
> 1. support for dvips/GhostScript
> 2. XMP template for PDF/A & PDF/UA simultaneously
> 3. adjust how accents are placed, so that Copy/Paste gets
> the order correct for Unicode combining accents.
>
> plus minor bug fixes.
>
>
>
> -- Peter
>
> > Alternatively, there is this site:
> >
> > https://sourceforge.net/projects/openicc/files/OpenICC-Profiles/<https://protect-au.mimecast.com/s/5iQkCL7Eg9fPo2PDCGmeyK?domain=sourceforge.net>
> >
> > Worth investigating?
> >
> >
> >
> > Cheers,
> >
> > Ross
>
>

[Karl Berry]

If Debian considers ghostscript/srgb.icc as free, isn't it free enough
for TeX Live as well?

Yes. At least assuming it's not just an oversight of checking, which is
unlikely in this case. -k

[Norbert Preining]

Dear all,

On Sun, 09 Sep 2018, Reinhard Kotucha wrote:
> So if we don't find an ICC sRGB profile with an appropriate license we
> can create one ourselves with a license which makes everybody happy.

As far as *I* remember this is not a trivial task.

> For the time being, Norbert, could you add the file
> sRGB_IEC61966-2-1_black_scaled.icc
> to tlcontrib if you haven't done this already?

And where should it be installed? Into the pdfx directory or
texmf-dist/tex/generic/icc
maybe?

On Mon, 10 Sep 2018, Ross Moore wrote:
> But there is one aspect of this that I don’t have the expertise in color usage to decide.
> How different are the various profiles?
> What is the difference between them? Does it matter?
> If we make a choice, then there should be a reason for it, concerning the actual
> colour output, either on-screen or when printed with different devices.

I think this is an important question and one that *I* cannot answer,
but I guess it makes lot of difference. I used icc profiles paired with
calibrated targets to color-correct slides I have scanned. Without it
the scan wasn't really nice at times.

All the best

Norbert

--
PREINING Norbert http://www.preining.info
Accelia Inc. + JAIST + TeX Live + Debian Developer
GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13

[Mojca Miklavec]

On Tue, 11 Sep 2018 at 02:49, Norbert Preining wrote:
>
> > For the time being, Norbert, could you add the file
> > sRGB_IEC61966-2-1_black_scaled.icc
> > to tlcontrib if you haven't done this already?
>
> And where should it be installed? Into the pdfx directory or
> texmf-dist/tex/generic/icc
> maybe?

Please put it somewhere under "generic" rather than "latex", if
possible. (I would still vote for creating a separate "color" top
folder at some point, but that's not all that important.)

I'm still waiting for the answer from ECI (http://www.eci.org/en/downloads).

Just to better understand the current situation:
- sRGB_IEC61966-2-1_black_scaled.icc is not free enough for TL, but OK
for TLContrib?
- For ECI we need some clarification before it becomes clear whether
they are OK even for TLContrib?
- Ghostscript profiles should probably be OK for TL as well (might
need some additional checking first)?

I would also like to include some of these profiles to the ConTeXt
distribution, but it's not clear to me what is allowed (since ConTeXt
distribution is not being sold on the DVD like TL, we are probably on
a somewhat safer side, but I would still not like to run into
troubles).

> On Mon, 10 Sep 2018, Ross Moore wrote:
> > But there is one aspect of this that I don’t have the expertise in color usage to decide.
> > How different are the various profiles?
> > What is the difference between them? Does it matter?

Yes, it matters a lot as long as one wants faithful colour
reproduction. For "generic red titles" it doesn't matter all that much
and it's more important to be able to create valid PDF/X.

Thank you,
Mojca

[Peter Selinger]

I think the file (whichever one) should be included with pdfx,
typically in the same directory as the other pdfx files. Each
candidate RGB profile is less than 10 kilobytes (the Ghostscript one
is 2576 bytes in size, and the /usr/share/color one has 6922 bytes),
so there is very little to be gained by not duplicating the file.

As to which profile to include: I don't think it matters much. Any
user who cares about precise print color specifications can provide
and include their own color profile, using e.g. the
\setRGBcolorprofile or \setCMYKcolorprofile commands provided by the
pdfx package.

The point of providing a default profile is to make life easy to the
rest of the users (presumably the majority) who don't care that much
about professional quality color. Most TeX documents are
black-and-white, with perhaps a few HTML colors like \color{red}
thrown in. They will look okay no matter which profile is used.

Moreover, as I understand it, RGB profiles are not really meant for
professional use (most professional printers will prefer CMYK
colors). Rather, the point of RGB profiles is to allow users to print
something that "looks reasonably similar to what they see on the
screen". The differences between various commonly-used RGB profiles
are probably quite negligible and I would not worry about it too much.

-- Peter

Norbert Preining wrote:
>
> Dear all,
>
> On Sun, 09 Sep 2018, Reinhard Kotucha wrote:
> > So if we don't find an ICC sRGB profile with an appropriate license we
> > can create one ourselves with a license which makes everybody happy.
>
> As far as *I* remember this is not a trivial task.
>
> > For the time being, Norbert, could you add the file
> > sRGB_IEC61966-2-1_black_scaled.icc
> > to tlcontrib if you haven't done this already?
>
> And where should it be installed? Into the pdfx directory or
> texmf-dist/tex/generic/icc
> maybe?
>
> On Mon, 10 Sep 2018, Ross Moore wrote:
> > But there is one aspect of this that I don’t have the expertise in color usage to decide.
> > How different are the various profiles?
> > What is the difference between them? Does it matter?
> > If we make a choice, then there should be a reason for it, concerning the actual
> > colour output, either on-screen or when printed with different devices.
>
> I think this is an important question and one that *I* cannot answer,
> but I guess it makes lot of difference. I used icc profiles paired with
> calibrated targets to color-correct slides I have scanned. Without it
> the scan wasn't really nice at times.
>
> All the best
>
> Norbert
>
> --
> PREINING Norbert http://www.preining.info
> Accelia Inc. + JAIST + TeX Live + Debian Developer
> GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13
>

[Norbert Preining]

Hi Peter,

> I think the file (whichever one) should be included with pdfx,
> typically in the same directory as the other pdfx files. Each

I agree with that, the size is irrelevant.

> As to which profile to include: I don't think it matters much. Any
> user who cares about precise print color specifications can provide
> and include their own color profile, using e.g. the
> \setRGBcolorprofile or \setCMYKcolorprofile commands provided by the
> pdfx package.

Indeed, that is a good point.

> The point of providing a default profile is to make life easy to the
> rest of the users (presumably the majority) who don't care that much
> about professional quality color. Most TeX documents are
> black-and-white, with perhaps a few HTML colors like \color{red}
> thrown in. They will look okay no matter which profile is used.

Then a simple sRGB profile would suffice for sure. Like the one Reinhard
sent with a straight forward color space transformation.

> Moreover, as I understand it, RGB profiles are not really meant for
> professional use (most professional printers will prefer CMYK

It was like this for long time (I often did print posters for
conferences, books etc) and CMYK separation was necessary for most
printers. But in recent years I have seen a move - at least in the
normal print business, not the high class super value stuff though I
guess - to be completely fine with RGB.

All the best

Norbert

--
PREINING Norbert http://www.preining.info
Accelia Inc. + JAIST + TeX Live + Debian Developer
GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13

[Zdenek Wagner]

út 11. 9. 2018 v 3:36 odesílatel Norbert Preining <***@logic.at>
napsal:

> Hi Peter,
> ...
>
> > Moreover, as I understand it, RGB profiles are not really meant for
> > professional use (most professional printers will prefer CMYK
>
> It was like this for long time (I often did print posters for
> conferences, books etc) and CMYK separation was necessary for most
> printers. But in recent years I have seen a move - at least in the
> normal print business, not the high class super value stuff though I
> guess - to be completely fine with RGB.
>

Yes, CMYK is prefered because colours on print are mixed subtractively and
thus CMY is required. Black is added for technical reasons (I do not want
to go into details). RGB implies additive mixing used on monitors.
Luminants on the monitors as well as pigments for printing do not work
according simple equations, their spectra are complex. In addition, the
resulting colour of the printed output depend not only on the colours of
the pigment but also on the colour of the paper which is never exactly
white. Thus the selection of the CMYK profile depends on the type of the
process colours as well as on the paper. The RGB profiles differ mostly on
the gamut, i.e. the range of colours that can be represented. AdobeRGB and
Wide Gamut RGB are wider than sRGB. Each RGB is wider than CMYK profiles,
so the conversion process must know what to do with colours that are out of
the gamut of the target CMYK and how to optionally modify other colours in
order not to lose details in the out-of-gamut colours. This means that in
addition to a proper ICC selection it is necessary to set also "intent". If
you have plots with a few colours or a text with a few HTML colours, then
these colours are usually highly saturated RGB. they are always out of the
CMYK gamut and since there is no risk of losing details, the intent should
be set to saturation and any profiles can be used. Even the simplest
equations can be applied because nothing better can be done. This is not
the case of colour photographs where the correct intent is perception. I
have Q60 target (less precise than IT8) and my scanners are calibrated. I
use the software that came with the scanner to calibrate my colour inkjet
printer and using a colour table available as a file and printed on the
calibrated printer I visually set the monitor. Thus I use the RGB profiles
for my scanners. I choose the CMYK profile by making tests on the printer
which wil be used and using the same paper, if the exact colours are
important. It is of course expensive but if high quality is needed you have
to pay. Generic ICC profiles are not good enough for high quality printing.

Good printers have their equipment and papers for high quality colour
printing calibrated and they distribute their ICC profiles for free but
they are good for their equipment and their papers.

>
> All the best
>
> Norbert
>

Zdeněk Wagner
http://ttsm.icpf.cas.cz/team/wagner.shtml
http://icebearsoft.euweb.cz




>
> --
> PREINING Norbert http://www.preining.info
> Accelia Inc. + JAIST + TeX Live + Debian Developer
> GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13
>

[Norbert Preining]

Hi Zdenek,

well said, I didn't want to go into that much details. But don't you
have the experience that for normal printing (that is not art books etc)
printers nowadays accept RGB?

For the last book project (collection of scientific papers) I prepared
everything with CMYK etc only to be told they need it in RGB :-(

Anyway, back on topic - you seem to be technically well versed here, can
you suggest a profile that should be used by default unless a user
defines one in pdfx package?

Best

Norbert

--
PREINING Norbert http://www.preining.info
Accelia Inc. + JAIST + TeX Live + Debian Developer
GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13

[Lars Madsen]

The last time we had a book with colors in it the printing house had some trouble in using RGB and converting it correctly to the CMYK setup they use.

They had to dig up some older programmes to do the conversion, and then of course our colors were off (luckily we had parameterized all colors in our Tikz images, so we managed), but having better handling for CMYK would be nice.

I think at least earlier there were certain parts of tikz that forced RGB.


/Lars Madsen
Institut for Matematik / Department of Mathematics
Aarhus Universitet / Aarhus University
Mere info: http://au.dk/***@math / More information: http://au.dk/en/***@math


________________________________________
From: tex-live <tex-live-bounces+daleif=***@tug.org> on behalf of Norbert Preining <***@logic.at>
Sent: 11 September 2018 12:20
To: Zdenek Wagner
Cc: Ross Moore; TeX Live; Peter Selinger
Subject: Re: [tex-live] ICC profiles for PDF/A compliance

Hi Zdenek,

well said, I didn't want to go into that much details. But don't you
have the experience that for normal printing (that is not art books etc)
printers nowadays accept RGB?

For the last book project (collection of scientific papers) I prepared
everything with CMYK etc only to be told they need it in RGB :-(

Anyway, back on topic - you seem to be technically well versed here, can
you suggest a profile that should be used by default unless a user
defines one in pdfx package?

Best

Norbert

--
PREINING Norbert http://www.preining.info
Accelia Inc. + JAIST + TeX Live + Debian Developer
GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13

[Zdenek Wagner]

út 11. 9. 2018 v 12:20 odesílatel Norbert Preining <***@logic.at>
napsal:

> Hi Zdenek,
>
> well said, I didn't want to go into that much details. But don't you
> have the experience that for normal printing (that is not art books etc)
> printers nowadays accept RGB?
>
> For the last book project (collection of scientific papers) I prepared
> everything with CMYK etc only to be told they need it in RGB :-(
>

It is strange. When I prepare posters for conferences, I do it in RGB
because these small shops accept files from normal users who know nothing
about colours abd send RGB. When I prepare a real book, they always require
CMYK converted by me to the correct ICC profile and refuse RGB. Once they
told me that thay intentionally switched the automatic colour conversion
off. Once I was just the man in the middle, the author refused to convert
the image from RGB and did not require exect colours. Then the personnel
enabled colour conversion upon my special written request. They can even
automatically convert RGB to grayscale but again, there are several
strategies for such conversin and the automatic conversion need not be the
right one.

>
> Anyway, back on topic - you seem to be technically well versed here, can
> you suggest a profile that should be used by default unless a user
> defines one in pdfx package?
>

If a tool is aware of ICC profiles, then majority of them uses sRGB as
default so it seems logical to have it as a default profile in the pdfx
package.


> Best
>


> Zdeněk Wagner
> http://ttsm.icpf.cas.cz/team/wagner.shtml
> http://icebearsoft.euweb.cz
>
>
>
>
> Norbert
>
> --
> PREINING Norbert http://www.preining.info
> Accelia Inc. + JAIST + TeX Live + Debian Developer
> GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13
>

[Reinhard Kotucha]

On 2018-09-10 at 22:14:59 -0300, Peter Selinger wrote:

> As to which profile to include: [...]

I've converted the profiles provided by Artifex to XML:

http://ms25.ddns.net/pdfx/profiles/gs

I'm not sure whether all are useful. Some do not provide a gamma LUT
but there is a function

{ dup 0.03928 le { 12.92321 div } { 0.055 add 1.055 div 2.4 exp } ifelse }

in Resource/ColorSpace/sRGB, for instance. Maybe a profile without a
LUT is used as a template by gs in order to create profiles on-the-fly.

Regards,
Reinhard

--
------------------------------------------------------------------
Reinhard Kotucha Phone: [phone]
Marschnerstr. 25
D-30167 Hannover mailto:***@web.de
------------------------------------------------------------------

[Karl Berry]

And where should it be installed? Into the pdfx directory or
texmf-dist/tex/generic/icc

It should go in texmf-dist/tex/generic/<pkgname>, where <pkgname> is
*what gets uploaded to CTAN*. Please do not create it on the fly in TL
just because you can :).

As for what <pkgname> should be, how about "colorprofile"? Because, are
all profiles really "ICC"? In any case, I don't feel strongly about the
name.

As for a new top-level directory instead, Mojca, do you have an argument
to advance for it? As far as I can see, it would cause churn in many
search paths with no offsetting advantage. tex/generic seems simpler to
me in every way. --best, karl.

[Zdenek Wagner]

út 11. 9. 2018 v 23:10 odesílatel Karl Berry <***@freefriends.org> napsal:

> And where should it be installed? Into the pdfx directory or
> texmf-dist/tex/generic/icc
>
> It should go in texmf-dist/tex/generic/<pkgname>, where <pkgname> is
> *what gets uploaded to CTAN*. Please do not create it on the fly in TL
> just because you can :).
>
> As for what <pkgname> should be, how about "colorprofile"? Because, are
> all profiles really "ICC"? In any case, I don't feel strongly about the
> name.
>

ICC stands for International Color Consortium. The extension of the files
with the profile is either .icc or .icm (as I have seen).

>
> As for a new top-level directory instead, Mojca, do you have an argument
> to advance for it? As far as I can see, it would cause churn in many
> search paths with no offsetting advantage. tex/generic seems simpler to
> me in every way. --best, karl.
>


Zdeněk Wagner
http://ttsm.icpf.cas.cz/team/wagner.shtml
http://icebearsoft.euweb.cz

[Ross Moore]

Hi Karl, and others

On 12 Sep 2018, at 7:09 am, Karl Berry <***@freefriends.org<mailto:***@freefriends.org>> wrote:

And where should it be installed? Into the pdfx directory or
texmf-dist/tex/generic/icc

It should go in texmf-dist/tex/generic/<pkgname>, where <pkgname> is
*what gets uploaded to CTAN*. Please do not create it on the fly in TL
just because you can :).

As for what <pkgname> should be, how about "colorprofile"? Because, are
all profiles really "ICC"? In any case, I don't feel strongly about the
name.

I fully support this.
The colour profiles should be independent of the pdfx package.
Yes, the package needs to reference them, and include them in documents,
but the package in no sense “owns” them, nor has it created them.

There are many other resources that pdfx makes use of, as does any other
package, which are loaded from elsewhere within the texmf-* hierarchies.
I don’t see this as being any different.

Besides, I don’t see pdfx as (ultimately) being the only package that produces
standards-conforming PDFs requiring a Color Profile, nor that such documents
necessarily be tied to LaTeX production.

So having them under ‘generic’ seems most appropriate, and we shouldn’t
need to have to distribute them with pdfx .


Besides, I always have trouble preparing a complicated package for CTAN,
where the structure or the archive submitted tends to be different from the
TDS structure that I use myself under TeXLive for MacOS.
I’m thinking of the doc/ , font/, source/ subtrees as well as tex/latex/<pkgname> ;
but all are flattened – for use with MikTeX ???

How easy will it be for all TeX systems to implement a new branch like
/tex/generic/colorprofile/
or /tex/generic/color/public/
or /color/profiles/public/
or whatever we decide is the best naming ?


As for a new top-level directory instead, Mojca, do you have an argument
to advance for it? As far as I can see, it would cause churn in many
search paths with no offsetting advantage. tex/generic seems simpler to
me in every way. --best, karl.



Cheers,

Ross


Dr Ross Moore

Mathematics Dept | 12 Wally’s Walk, 734
Macquarie University, NSW 2109, Australia

T: [phone] | F: [phone]<tel:%2B61%202%209850%209695>
M:[phone]<tel:%2B61%20409%20125%20670> | E: ***@mq.edu.au<mailto:***@mq.edu.au>

http://www.maths.mq.edu.au<http://mq.edu.au/>


[cid:***@01D030BE.D37A46F0]<http://mq.edu.au/>




CRICOS Provider Number 00002J. Think before you print.
Please consider the environment before printing this email.<http://mq.edu.au/>

This message is intended for the addressee named and may
contain confidential information. If you are not the intended
recipient, please delete it and notify the sender. Views expressed
in this message are those of the individual sender, and are not
necessarily the views of Macquarie University.<http://mq.edu.au/>

[Norbert Preining]

Hi all,

> It should go in texmf-dist/tex/generic/<pkgname>, where <pkgname> is
> *what gets uploaded to CTAN*. Please do not create it on the fly in TL
> just because you can :).

Of course not. What about
colorprofiles
and
colorprofiles-nonfree
Both could be uploaded to CTAN, while the former one would be included
in TL, the later one in tlcontrib.

I can set up a github repo that collects all the profiles with license
information and creates both packages for upload to CTAN.

Target dir would be as discussed
TEXMF/tex/generic/$pkg

It is still unclear to me which one should be put into the free package,
so please everyone with interest here, please send me those that you
consider free with the respective license statement and where you got
it, so that I can include them in the package.

Norbert

--
PREINING Norbert http://www.preining.info
Accelia Inc. + JAIST + TeX Live + Debian Developer
GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13

[Mojca Miklavec]

On Thu, 13 Sep 2018 at 04:11, Norbert Preining wrote:
>
> Hi all,
>
> > It should go in texmf-dist/tex/generic/<pkgname>, where <pkgname> is
> > *what gets uploaded to CTAN*. Please do not create it on the fly in TL
> > just because you can :).
>
> Of course not. What about
> colorprofiles
> and
> colorprofiles-nonfree
> Both could be uploaded to CTAN, while the former one would be included
> in TL, the later one in tlcontrib.
>
> I can set up a github repo that collects all the profiles with license
> information and creates both packages for upload to CTAN.
>
> Target dir would be as discussed
> TEXMF/tex/generic/$pkg

This sounds reasonable to me.

> It is still unclear to me which one should be put into the free package,

Same here :)

On Thu, 13 Sep 2018 at 04:32, Peter Selinger wrote:
>
> As I understand it, there isn't really any reason to include non-free
> profiles. It should be possible to just replace the ones we currently
> have with free ones. -- Peter

That's approximately like saying that it should be fully sufficient to
ship just a single font with TeX Live, or to use just a single font in
every single document.
Sure, we need a default font, and if all you want to do is to transfer
some text to another person, you don't really care if it's all done in
a free version of Comic Sans.

I'm not an expert here, but if one single profile was all that was
needed, it wouldn't be required to specify it at all.

We are talking about having at least one free profile which could be
used by default, but some additional ones, even if distributed only
via tlcontrib, would certainly help.

Mojca

[Zdenek Wagner]

...

> On Thu, 13 Sep 2018 at 04:32, Peter Selinger wrote:
> >
> > As I understand it, there isn't really any reason to include non-free
> > profiles. It should be possible to just replace the ones we currently
> > have with free ones. -- Peter
>
> That's approximately like saying that it should be fully sufficient to
> ship just a single font with TeX Live, or to use just a single font in
> every single document.
> Sure, we need a default font, and if all you want to do is to transfer
> some text to another person, you don't really care if it's all done in
> a free version of Comic Sans.
>
> I'm not an expert here, but if one single profile was all that was
> needed, it wouldn't be required to specify it at all.
>
> We are talking about having at least one free profile which could be
> used by default, but some additional ones, even if distributed only
> via tlcontrib, would certainly help.


The specification requires an ICC profile to be included but no further
requirement is implied. For users who have just simple colour plots
any profile is acceptable and the default sRGB will do the job For
other applications a special (non-free) profile can be needed. Thus
having one or a few free profiles and other profiles in tlcontrib is
IMHO the best solutions.

>
>
> Mojca



Zdeněk Wagner
http://ttsm.icpf.cas.cz/team/wagner.shtml
http://icebearsoft.euweb.cz

[Karl Berry]

Please put it somewhere under "generic" rather than "latex",

I think we all agree on that.

- sRGB_IEC61966-2-1_black_scaled.icc is not free enough for TL, but OK
for TLContrib?
Yes.

- For ECI we need some clarification before it becomes clear whether
they are OK even for TLContrib?
Yes.

- Ghostscript profiles should probably be OK for TL as well (might
need some additional checking first)?

I would say "Debian" profiles rather than "Ghostscript", though it may
come to the same thing in this case. It's much more likely that Debian
has checked licenses with approximately the same criteria as TL.

since ConTeXt distribution is not being sold on the DVD like TL,

It's irrelevant whether TL is, in fact, being sold on DVD (that is,
nothing would change if Lehmann stopped selling the DVD). What matters
is that people (anyone) have the legal ability to do so. That's part of
being free-as-in-freedom software, in the definition that we use in our
little world.

If the ConTeXt distribution decides to admit nonfree files, that's a
major policy change (as far as I know; maybe it's already happened?).
For one thing, we have to be sure that any such files are excluded when
ConTeXt is imported into TL. --thanks, karl.

[Mojca Miklavec]

Hi,

On Wed, 12 Sep 2018 at 00:19, Karl Berry wrote:
>
> since ConTeXt distribution is not being sold on the DVD like TL,
>
> It's irrelevant whether TL is, in fact, being sold on DVD (that is,
> nothing would change if Lehmann stopped selling the DVD). What matters
> is that people (anyone) have the legal ability to do so. That's part of
> being free-as-in-freedom software, in the definition that we use in our
> little world.

(OK, that was a bit of a wrong wording on my side. I'm sorry.)

> If the ConTeXt distribution decides to admit nonfree files, that's a

I was talking about the ConTeXt distribution ("equivalent" of TeX Live
distribution) rather than about ConTeXt package itself.

> major policy change (as far as I know; maybe it's already happened?).

You did remove cow fonts from ConTeXt in the past due to "please do
not change these glyphs / work of art" policy violating the freedom to
modify the files.

> For one thing, we have to be sure that any such files are excluded when
> ConTeXt is imported into TL. --thanks, karl.

We are not talking about ConTeXt package at the moment, so nothing to
worry about.

Mojca

[Peter Selinger]

My only concern would be if people install the pdfx package
individually from CTAN, they have to go through the extra (and not
obvious) step of installing the new icc package as well. I don't see
any significant downside to simply making this 2k file part of the
pdfx package, or any significant upside to creating a new global
location for this one file.

There are other things that the pdfx package contains, such as
glyph-to-unicode mappings, which can in principle also be re-used by
other packages. But ultimately one cannot make a separate package for
each file, so I think it is better to keep related files together.

This is just my 2c, but you guys are the experts on this so ultimately
I agree with whatever you decide. -- Peter

Ross Moore wrote:
>
> Hi Karl, and others
>
> On 12 Sep 2018, at 7:09 am, Karl Berry <***@freefriends.org<mailto:***@freefriends.org>> wrote:
>
> And where should it be installed? Into the pdfx directory or
> texmf-dist/tex/generic/icc
>
> It should go in texmf-dist/tex/generic/<pkgname>, where <pkgname> is
> *what gets uploaded to CTAN*. Please do not create it on the fly in TL
> just because you can :).
>
> As for what <pkgname> should be, how about "colorprofile"? Because, are
> all profiles really "ICC"? In any case, I don't feel strongly about the
> name.
>
> I fully support this.
> The colour profiles should be independent of the pdfx package.
> Yes, the package needs to reference them, and include them in documents,
> but the package in no sense “owns” them, nor has it created them.
>
> There are many other resources that pdfx makes use of, as does any other
> package, which are loaded from elsewhere within the texmf-* hierarchies.
> I don’t see this as being any different.
>
> Besides, I don’t see pdfx as (ultimately) being the only package that produces
> standards-conforming PDFs requiring a Color Profile, nor that such documents
> necessarily be tied to LaTeX production.
>
> So having them under ‘generic’ seems most appropriate, and we shouldn’t
> need to have to distribute them with pdfx .
>
>
> Besides, I always have trouble preparing a complicated package for CTAN,
> where the structure or the archive submitted tends to be different from the
> TDS structure that I use myself under TeXLive for MacOS.
> I’m thinking of the doc/ , font/, source/ subtrees as well as tex/latex/<pkgname> ;
> but all are flattened – for use with MikTeX ???
>
> How easy will it be for all TeX systems to implement a new branch like
> /tex/generic/colorprofile/
> or /tex/generic/color/public/
> or /color/profiles/public/
> or whatever we decide is the best naming ?
>
>
> As for a new top-level directory instead, Mojca, do you have an argument
> to advance for it? As far as I can see, it would cause churn in many
> search paths with no offsetting advantage. tex/generic seems simpler to
> me in every way. --best, karl.
>
>
>
> Cheers,
>
> Ross
>
>
> Dr Ross Moore
>
> Mathematics Dept | 12 Wally’s Walk, 734
> Macquarie University, NSW 2109, Australia
>
> T: [phone] | F: [phone]<tel:%2B61%202%209850%209695>
> M:[phone]<tel:%2B61%20409%20125%20670> | E: ***@mq.edu.au<mailto:***@mq.edu.au>
>
> http://www.maths.mq.edu.au<http://mq.edu.au/>
>
>
> [cid:***@01D030BE.D37A46F0]<http://mq.edu.au/>
>
>
>
>
> CRICOS Provider Number 00002J. Think before you print.
> Please consider the environment before printing this email.<http://mq.edu.au/>
>
> This message is intended for the addressee named and may
> contain confidential information. If you are not the intended
> recipient, please delete it and notify the sender. Views expressed
> in this message are those of the individual sender, and are not
> necessarily the views of Macquarie University.<http://mq.edu.au/>
>

[Mojca Miklavec]

On Wed, 12 Sep 2018 at 03:40, Peter Selinger wrote:
>
> My only concern would be if people install the pdfx package
> individually from CTAN,

Who would want to do that? I'm not claiming there are no such
individuals around, but installing it is tedious.

Like already mentioned, the TDS structure is flattened on CTAN, so the
user would need to place every single file individually into the
proper location in the TEXMF tree.

> they have to go through the extra (and not
> obvious) step of installing the new icc package as well.

Also, if the user wants to install everything from scratch from CTAN,
what about all the fonts and all the other packages that are needed to
compile a simple document? Doing that without a TeX distribution is
literally a nightmare.

And inside TL the dependency can easily be implemented by adding a
line to tlpsrc if such a package exists.

> I don't see
> any significant downside to simply making this 2k file part of the
> pdfx package,

Other than the licence which apparently prohibits inclusion into TeX
Live in the first place?

Mojca

[Zdenek Wagner]

st 12. 9. 2018 v 11:08 odesílatel Mojca Miklavec <
***@gmail.com> napsal:

> On Wed, 12 Sep 2018 at 03:40, Peter Selinger wrote:
> >
> > My only concern would be if people install the pdfx package
> > individually from CTAN,
>
> Who would want to do that? I'm not claiming there are no such
> individuals around, but installing it is tedious.
>
> Like already mentioned, the TDS structure is flattened on CTAN, so the
> user would need to place every single file individually into the
> proper location in the TEXMF tree.
>

CTAN allows to upload the TDS version in addition to the flattened version
but only some packages are uploaded in both forms.

>
> > they have to go through the extra (and not
> > obvious) step of installing the new icc package as well.
>
> Also, if the user wants to install everything from scratch from CTAN,
> what about all the fonts and all the other packages that are needed to
> compile a simple document? Doing that without a TeX distribution is
> literally a nightmare.
>

There are too many packages which simply do not work if another package is
missing. Taking just a single package from CTAN is useful if it consists of
one file plu documentation, otherwise it is a nightmare...

>
> And inside TL the dependency can easily be implemented by adding a
> line to tlpsrc if such a package exists.
>
> > I don't see
> > any significant downside to simply making this 2k file part of the
> > pdfx package,
>
> Other than the licence which apparently prohibits inclusion into TeX
> Live in the first place?
>
> Mojca
>


Zdeněk Wagner
http://ttsm.icpf.cas.cz/team/wagner.shtml
http://icebearsoft.euweb.cz

[Reinhard Kotucha]

On 2018-09-12 at 11:07:53 +0200, Mojca Miklavec wrote:

> Like already mentioned, the TDS structure is flattened on CTAN, so
> the user would need to place every single file individually into
> the proper location in the TEXMF tree.

and all this has to be explained in the pdfx package documentation in
great detail. It blows up the manual unnecessarily.

I'm aware that these files are tiny. But I fear that when we allow
one and the same file to be installed in two different directories, we
potentially open Pandora's box.

Prior to TL 2008, spripts were maintained in bin/<platform>/ and we
noticed that scripts for different platforms had different version
numbers. It's almost impossible to maintain one and the same file in
two different directories and make sure that they are identical.

--------------------------------------------^

BTW, if your mail client is configured to use a monospaced font and to
obey the original linebreaks, you can see here what typographers call
a 'river'.

Regards,
Reinhard

--
------------------------------------------------------------------
Reinhard Kotucha Phone: [phone]
Marschnerstr. 25
D-30167 Hannover mailto:***@web.de
------------------------------------------------------------------

[Peter Selinger]

As I understand it, there isn't really any reason to include non-free
profiles. It should be possible to just replace the ones we currently
have with free ones. -- Peter

Norbert Preining wrote:
>
> Hi all,
>
> > It should go in texmf-dist/tex/generic/<pkgname>, where <pkgname> is
> > *what gets uploaded to CTAN*. Please do not create it on the fly in TL
> > just because you can :).
>
> Of course not. What about
> colorprofiles
> and
> colorprofiles-nonfree
> Both could be uploaded to CTAN, while the former one would be included
> in TL, the later one in tlcontrib.
>
> I can set up a github repo that collects all the profiles with license
> information and creates both packages for upload to CTAN.
>
> Target dir would be as discussed
> TEXMF/tex/generic/$pkg
>
> It is still unclear to me which one should be put into the free package,
> so please everyone with interest here, please send me those that you
> consider free with the respective license statement and where you got
> it, so that I can include them in the package.
>
> Norbert
>
> --
> PREINING Norbert http://www.preining.info
> Accelia Inc. + JAIST + TeX Live + Debian Developer
> GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13
>

[Peter Selinger]

Mojca Miklavec wrote:
>
> On Thu, 13 Sep 2018 at 04:32, Peter Selinger wrote:
> >
> > As I understand it, there isn't really any reason to include non-free
> > profiles. It should be possible to just replace the ones we currently
> > have with free ones. -- Peter
>
> That's approximately like saying that it should be fully sufficient to
> ship just a single font with TeX Live, or to use just a single font in
> every single document.
> Sure, we need a default font, and if all you want to do is to transfer
> some text to another person, you don't really care if it's all done in
> a free version of Comic Sans.
>
> I'm not an expert here, but if one single profile was all that was
> needed, it wouldn't be required to specify it at all.
>
> We are talking about having at least one free profile which could be
> used by default, but some additional ones, even if distributed only
> via tlcontrib, would certainly help.

No, it's not analogous. Color profiles are associated to (high-end)
devices like printers, scanners, digital cameras, and monitors. They
are generated by calibrating these devices, and are usually provided
by the device's manufacturer, or in some cases, can be generated by
users using (expensive) calibration hardware and software. If a user
needs to supply their own color profile for inclusion in a PDF file,
they will usually use a (non-free) profile that comes with their
hardware or that they otherwise associate with their image data. For
example, if I take a picture with a high-end digital camera, the
picture file includes the camera's (non-free) color profile as part of
its meta-data, and if I later use the image in a PDF file, the color
profile makes its way into the PDF. This ensures that the printer
interprets the colors in the way intended by the original device (the
camera).

There is no reason, in my opinion, for TeXLive to become a repository
for a large number of color profiles from different sources,
especially non-free ones.

The only reason a TeX distribution needs to include any profile at all
is that certain PDF standards require the inclusion of a profile,
*even if the document does not contain any color* or *even if the
document only includes vanilla RGB color*. Therefore, the reasonable
thing to do, in my opinion, is to include a single, free, sRGB
profile. It may also be reasonable to additionally include a single,
free CMYK profile (such as FOGRA39, which seems to be an ISO standard
for "typical print made by commercial offset litho") in case there are
any users who use a CMYK image from an unsophisticated source that
does not already have a color profile associated with it.

-- Peter

[Norbert Preining]

Dear all,

> There is no reason, in my opinion, for TeXLive to become a repository
> for a large number of color profiles from different sources,
> especially non-free ones.

That will never happen.

> document only includes vanilla RGB color*. Therefore, the reasonable
> thing to do, in my opinion, is to include a single, free, sRGB
> profile. It may also be reasonable to additionally include a single,
> free CMYK profile (such as FOGRA39, which seems to be an ISO standard

Ok, please, can *someone* please send either the sRGB or both, and I
will prepare a package. That would close the discussion ...

Thanks

Norbert

--
PREINING Norbert http://www.preining.info
Accelia Inc. + JAIST + TeX Live + Debian Developer
GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13

[Boris Veytsman]

PS> Date: Thu, 13 Sep 2018 08:53:57 -0300 (ADT)
PS> From: ***@mathstat.dal.ca (Peter Selinger)


PS> The only reason a TeX distribution needs to include any profile at all
PS> is that certain PDF standards require the inclusion of a profile,
PS> *even if the document does not contain any color* or *even if the
PS> document only includes vanilla RGB color*. Therefore, the reasonable
PS> thing to do, in my opinion, is to include a single, free, sRGB
PS> profile. It may also be reasonable to additionally include a single,
PS> free CMYK profile (such as FOGRA39, which seems to be an ISO standard
PS> for "typical print made by commercial offset litho") in case there are
PS> any users who use a CMYK image from an unsophisticated source that
PS> does not already have a color profile associated with it.

This is exactly what I saw with many of my customers: they need
*some* profile.



--
Good luck

-Boris

Cleanse area thoroughly before applying.

[Peter Selinger]

Good point. Attached are 4 files:

sRGB.iccprofile.xml
sRGB.icc
FOGRA39L_coated.iccprofile.xml
FOGRA39L_coated.icc

The XML files are taken from the Debian colord-1.4.3 source package.
Each XML files contains the text: "This profile is free of known
copyright restrictions".

The ICC files were generated from the XML files by me, today, using
the open-source program /usr/bin/cd-create-profile, which is also part
of the Debian colord-1.4.3 package.

Since the ICC files just contain the information from the XML files in
a different data format, the ICC files are not in themselves original
works and therefore not copyrightable in their own right. However, for
greater clarity, to the extent that I have any rights to these files
as their creator, I hereby waive all of my rights to the work
worldwide under copyright law, including all related and neighboring
rights, to the extent allowed by law.

Each ICC files also contains a UTF-16 encoding of the statement "This
profile is free of known copyright restrictions".

It will also be necessary to update the respective filenames in
pdfx.sty.

Thanks, -- Peter

Norbert Preining wrote:
>
> Dear all,
>
> > There is no reason, in my opinion, for TeXLive to become a repository
> > for a large number of color profiles from different sources,
> > especially non-free ones.
>
> That will never happen.
>
> > document only includes vanilla RGB color*. Therefore, the reasonable
> > thing to do, in my opinion, is to include a single, free, sRGB
> > profile. It may also be reasonable to additionally include a single,
> > free CMYK profile (such as FOGRA39, which seems to be an ISO standard
>
> Ok, please, can *someone* please send either the sRGB or both, and I
> will prepare a package. That would close the discussion ...
>
> Thanks
>
> Norbert
>
> --
> PREINING Norbert http://www.preining.info
> Accelia Inc. + JAIST + TeX Live + Debian Developer
> GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13
>

[Norbert Preining]

Dear Peter, dear all,

> sRGB.iccprofile.xml
> sRGB.icc
> FOGRA39L_coated.iccprofile.xml
> FOGRA39L_coated.icc

Thanks a lot.

BUT ...

I was pointed in private by a pdfx user to
https://ninedegreesbelow.com/photography/srgb-profile-comparison.html
which gives a *very* detailed and interesting comparison, though
slightly dated (2012/2015) of different sRGB color profiles, and argues
strongly in favor of the Argyllcms sRGB profile as being the one closest
to the specification with the best properties.
https://www.argyllcms.com/

The original mail to me also contained the following statement
> Regarding other sRGB profiles, there are some that are also close to the
> specification, but there are others that have got a different white
> point, a different tone response curve or a different black point.
> Please try to avoid the latter.

The file contained in the argyll-ref package in Debian is
/usr/share/color/argyll/ref/sRGB.icm

The discussion at the above mentioned web site - although I couldn't
check all the details - makes me somehow tempted to use the sRGB from
ArgyllCMS.

Peter, what is your opinion on this?

Thanks

Norbert

--
PREINING Norbert http://www.preining.info
Accelia Inc. + JAIST + TeX Live + Debian Developer
GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13

[Peter Selinger]

Yes, why not. The /usr/share/color/argyll/ref/sRGB.icm profile is also
free, it states "Created by Graeme W. Gill. Released into the public
domain. No Warranty, Use at your own risk."

I don't have personally prefer one sRGB profile over another. Since
that website makes a careful investigation and reasoned argument, it
is better than no data at all, so why not go with that one then.

-- Peter

Norbert Preining wrote:
>
> Dear Peter, dear all,
>
> > sRGB.iccprofile.xml
> > sRGB.icc
> > FOGRA39L_coated.iccprofile.xml
> > FOGRA39L_coated.icc
>
> Thanks a lot.
>
> BUT ...
>
> I was pointed in private by a pdfx user to
> https://ninedegreesbelow.com/photography/srgb-profile-comparison.html
> which gives a *very* detailed and interesting comparison, though
> slightly dated (2012/2015) of different sRGB color profiles, and argues
> strongly in favor of the Argyllcms sRGB profile as being the one closest
> to the specification with the best properties.
> https://www.argyllcms.com/
>
> The original mail to me also contained the following statement
> > Regarding other sRGB profiles, there are some that are also close to the
> > specification, but there are others that have got a different white
> > point, a different tone response curve or a different black point.
> > Please try to avoid the latter.
>
> The file contained in the argyll-ref package in Debian is
> /usr/share/color/argyll/ref/sRGB.icm
>
> The discussion at the above mentioned web site - although I couldn't
> check all the details - makes me somehow tempted to use the sRGB from
> ArgyllCMS.
>
> Peter, what is your opinion on this?
>
> Thanks
>
> Norbert
>
> --
> PREINING Norbert http://www.preining.info
> Accelia Inc. + JAIST + TeX Live + Debian Developer
> GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13
>

[Ross Moore]

Hi all.

We got a report today of the missing sRGB Profile, when processing
a document using pdfx.sty .

So we need to finalise this.
I’ve started to work on a revised package, but it means:
* including an sRGB profile, and
* adjusting pdfx.sty for the new default name.
* revising some documentation.

On 14 Sep 2018, at 1:30 pm, Peter Selinger <***@mathstat.dal.ca<mailto:***@mathstat.dal.ca>> wrote:

Yes, why not. The /usr/share/color/argyll/ref/sRGB.icm profile is also
free, it states "Created by Graeme W. Gill. Released into the public
domain. No Warranty, Use at your own risk.”

Nice and simple, but how do I get the file?
It’s from a Linux distribution. I’m on MacOS.

Aaah, here it is: http://www.argyllcms.com/sRGB.icm
but where is a simple License statement to be found?



An alternative is from ICC, with Licence:

"This profile is made available by the International Color Consortium, and may be copied, distributed, embedded, made, used, and sold without restriction. Altered versions of this profile shall have the original identification and copyright information removed and shall not be misrepresented as the original profile."

http://www.color.org/srgbprofiles.xalter#v2
It downloads as sRGB2014.icc when you click: “accept licensing terms”
The actual URL is http://www.color.org/profiles/sRGB2014.icc .

So I can go with this straight away, if that’s acceptable to everyone.



I don't have personally prefer one sRGB profile over another. Since
that website makes a careful investigation and reasoned argument, it
is better than no data at all, so why not go with that one then.

-- Peter

Norbert Preining wrote:
>
> Dear Peter, dear all,
>
> > sRGB.iccprofile.xml
> > sRGB.icc
> > FOGRA39L_coated.iccprofile.xml
> > FOGRA39L_coated.icc
>
> Thanks a lot.
>
> BUT ...
>
> I was pointed in private by a pdfx user to
> https://ninedegreesbelow.com/photography/srgb-profile-comparison.html<https://protect-au.mimecast.com/s/R0R0CBNqgBCVgYvgTjQA7t?domain=ninedegreesbelow.com>
> which gives a *very* detailed and interesting comparison, though
> slightly dated (2012/2015) of different sRGB color profiles, and argues
> strongly in favor of the Argyllcms sRGB profile as being the one closest
> to the specification with the best properties.
> https://www.argyllcms.com/<https://protect-au.mimecast.com/s/IT1FCD1vRkCBR4XRcl4M4p?domain=argyllcms.com>

Yeah, but I couldn’t find a licence that clearly applied just to the color Profile.
The whole Argyll CMS system is under
GNU AFFERO GENERAL PUBLIC LICENSE
which probably does allow us to take just a single file.
But I don’t really want to include all of that, when a couple of sentences suffices.


> The original mail to me also contained the following statement
> > Regarding other sRGB profiles, there are some that are also close to the
> > specification, but there are others that have got a different white
> > point, a different tone response curve or a different black point.
> > Please try to avoid the latter.
>
> The file contained in the argyll-ref package in Debian is
> /usr/share/color/argyll/ref/sRGB.icm
>
> The discussion at the above mentioned web site - although I couldn't
> check all the details - makes me somehow tempted to use the sRGB from
> ArgyllCMS.
>
> Peter, what is your opinion on this?
>
> Thanks
>
> Norbert


Cheers,

Ross


Dr Ross Moore

Mathematics Dept | 12 Wally’s Walk, 734
Macquarie University, NSW 2109, Australia

T: [phone] | F: [phone]<tel:%2B61%202%209850%209695>
M:[phone]<tel:%2B61%20409%20125%20670> | E: ***@mq.edu.au<mailto:***@mq.edu.au>

http://www.maths.mq.edu.au<http://mq.edu.au/>


[cid:***@01D030BE.D37A46F0]<http://mq.edu.au/>




CRICOS Provider Number 00002J. Think before you print.
Please consider the environment before printing this email.<http://mq.edu.au/>

This message is intended for the addressee named and may
contain confidential information. If you are not the intended
recipient, please delete it and notify the sender. Views expressed
in this message are those of the individual sender, and are not
necessarily the views of Macquarie University.<http://mq.edu.au/>

[Norbert Preining]

Hi Ross, hi all,


sorry for being late...

On Tue, 09 Oct 2018, Ross Moore wrote:
> * including an sRGB profile, and

I have uploaded a package to CTAN called "colorprofiles" which contains
sRGB.icc
FOGRA39L_coated.icc
It will be installed into TeX Live within the next 2 days I guess.

> * adjusting pdfx.sty for the new default name.

You can use sRGB.icc, this is the name as I included it.

I used the sRGB.icm from Argyllcms, and FOGRA39L_coated as sent by
Peter.

All the best

Norbert

--
PREINING Norbert http://www.preining.info
Accelia Inc. + JAIST + TeX Live + Debian Developer
GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13

[Ross Moore]

Hi Norbert,

On 9 Oct 2018, at 12:44 pm, Norbert Preining <***@logic.at<mailto:***@logic.at>> wrote:

Hi Ross, hi all,


sorry for being late...

On Tue, 09 Oct 2018, Ross Moore wrote:
> * including an sRGB profile, and

I have uploaded a package to CTAN called "colorprofiles" which contains
sRGB.icc
FOGRA39L_coated.icc
It will be installed into TeX Live within the next 2 days I guess.

> * adjusting pdfx.sty for the new default name.

You can use sRGB.icc, this is the name as I included it.

I used the sRGB.icm from Argyllcms, and FOGRA39L_coated as sent by
Peter.

Great.
When I get that, I can make the appropriate adjustments.

I’m a little surprised at the message we (the pdfx developers)
got reporting that someone could no longer process a document that
had worked previously.

Here is the message that was sent:


PDFX: Color Profile Problems on MacOS, Texlive 2018

To: Radhakrishnan CV <***@river-valley.org<mailto:***@river-valley.org>>, Thanh Hàn Thé <***@river-valley.org<mailto:***@river-valley.org>>, Ross Moore <***@mq.edu.au<mailto:***@mq.edu.au>>, Peter Selinger <***@mathstat.dal.ca<mailto:***@mathstat.dal.ca>>

Dear Authors of PDFX,

the recent update to (48612, 1.5.84) may have introduced a regression. Previously, e.g., on (44412, 1.5.84), documents would compile without errors, but know I get:

pdfx.sty:1301: No color profile found to use for RGB screen colors

I assume the defaults were used previously. Do I have to provide a color profile manually from now on or is this a new issue?

Best regards,
Steffen MÀrcker

There has been no update yet to pdfx.sty on CTAN or for TeXLive 2018, at least not by me.
I thought that the sRGB color profile was only to be removed from Linux distributions,
so why is this problem appearing under MacOSX, as stated in his header?

My experience is that if you install a new TeXLive distribution,
then the old one stays but kpsewhich may not find it.
So the simplest solution is to copy/move the old file into a texmf-local/ hierarchy,
until the updated pdfx.sty becomes available, using a new file name.

Is this not accurate?



All the best

Norbert


Cheers,

Ross


Dr Ross Moore

Mathematics Dept | 12 Wally’s Walk, 734
Macquarie University, NSW 2109, Australia

T: [phone] | F: [phone]<tel:%2B61%202%209850%209695>
M:[phone]<tel:%2B61%20409%20125%20670> | E: ***@mq.edu.au<mailto:***@mq.edu.au>

http://www.maths.mq.edu.au<http://mq.edu.au/>


[cid:***@01D030BE.D37A46F0]<http://mq.edu.au/>




CRICOS Provider Number 00002J. Think before you print.
Please consider the environment before printing this email.<http://mq.edu.au/>

This message is intended for the addressee named and may
contain confidential information. If you are not the intended
recipient, please delete it and notify the sender. Views expressed
in this message are those of the individual sender, and are not
necessarily the views of Macquarie University.<http://mq.edu.au/>

[Norbert Preining]

Dear Ross,

On Tue, 09 Oct 2018, Ross Moore wrote:
> There has been no update yet to pdfx.sty on CTAN or for TeXLive 2018, at least not by me.

No, the profile was removed from TeX Live, too, thus it did effect
MacOSX as well. The update to the xpdf package was not triggered by an
update by you at CTAN, but by the removal of the non-free profile.

The colorprofiles packages is already installed on CTAN and might show
up already tomorrow in TeX Live.

Best

Norbert

--
PREINING Norbert http://www.preining.info
Accelia Inc. + JAIST + TeX Live + Debian Developer
GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13

[Peter Selinger]

> Aaah, here it is: http://www.argyllcms.com/sRGB.icm
> but where is a simple License statement to be found?

Just for the record: a simple License statement is included as a
string in the file itself. The string is "Public Domain. No Warranty,
Use at own risk."

-- Peter

[Ross Moore]

On 9 Oct 2018, at 10:37 pm, Peter Selinger <***@mathstat.dal.ca<mailto:***@mathstat.dal.ca>> wrote:

> Aaah, here it is: http://www.argyllcms.com/sRGB.icm<https://protect-au.mimecast.com/s/cyG5Cvl1g2SW6qZwSX1WSB?domain=argyllcms.com>
> but where is a simple License statement to be found?

Just for the record: a simple License statement is included as a
string in the file itself. The string is "Public Domain. No Warranty,
Use at own risk.”

Beautiful.


-- Peter


Dr Ross Moore

Mathematics Dept | 12 Wally’s Walk, 734
Macquarie University, NSW 2109, Australia

T: [phone] | F: [phone]<tel:%2B61%202%209850%209695>
M:[phone]<tel:%2B61%20409%20125%20670> | E: ***@mq.edu.au<mailto:***@mq.edu.au>

http://www.maths.mq.edu.au<http://mq.edu.au/>


[cid:***@01D030BE.D37A46F0]<http://mq.edu.au/>




CRICOS Provider Number 00002J. Think before you print.
Please consider the environment before printing this email.<http://mq.edu.au/>

This message is intended for the addressee named and may
contain confidential information. If you are not the intended
recipient, please delete it and notify the sender. Views expressed
in this message are those of the individual sender, and are not
necessarily the views of Macquarie University.<http://mq.edu.au/>