You all are wonderful. Thank you all for the replies! I really appreciate
your taking the time out of your day to talk to me about this. You leave me
with a lot to think about. I will be looking for good ways to implement and
test your examples and advice in my wiki. I asked general questions, and I
got general answers. Maybe it would also be useful for me to be more
specific about my context.
My 19mb spaghetti wiki is ~6k tiddlers hardlinking to each other from their
bodies: https://philosopher.life/. Perhaps it's time for a massive
overhaul. I am fine with refactoring the entire thing by hand (although, I
think some of it could be automated), but I want to make sure I get the
most bang for my buck. I'm interesting to know, assuming you were in my
shoes/context, which low-hanging tagging or other mechanical improvements
you'd make.
In TW tags are mainly used to create dynamic lists. like the TOC or a
list of links.
If the tag is also a tiddler, those lists can be sorted, using the
list-field
With some valuable exceptions, I barely make use of dynamic lists. I create
lists by hand the majority of the time. Perhaps my link hardcoding is
wrong-headed. Part of my problem is that I don't always know how I want my
wiki to evolve, and maybe hard-coding has been a very flawed WYSIWYG coping
mechanism. Sometimes I'll bust something up into pieces and put the pieces
in different places. I feel like I'm building with lego with hardlinks.
less typing.
This seems especially true if you've automated tagging, although I think
automatically naming tiddlers is also possible (I've never figured that
out). For my logs, I often CnP an older link, edit the new one, and just
open the new link. I'm fairly quick on my keyboard with shortcuts, so this
hasn't seemed too painful to me. Tags seem like a good off-the-shelf answer
though. Perhaps it will save me a lot of time in the long run.
Once I have the hardlinks, it's easy to move them around them as sets,
manipulate the material, build other kinds of things with them. I often
have to lexically order my links by hand. Links feel concrete to me. I fear
I'm just rationalizing here or missing the point. Call me out on it.
If I had to assign most tiddlers to multiple groups most of the time, I
think tagging would awesome for me. Unfortunately, for whatever reason, I
don't often find the need for one tiddler to be found in more than one
category. When I do, I link or transclude by hand. It can be a very special
event for me in my wiki when I see a tiddler belongs in multiple places;
sometimes it screams something important to me. Perhaps tagging would make
this event more common.
As you pointed out, having a tiddler naming convention helps with
searching. ... But what if such a naming convention doesn't fit the
use-case ... Tags can be very handy here.
The vast majority of the time, my searches don't even use my tiddler naming
convention at all. I generally either know part of the name of the tiddler
I'm reaching for, or I'm just narrowing piles down by keywords found in
their bodies. Search is where I am least convinced that tags do anything
for me, but maybe that is just specific to my usecase. All the exceptions I
can think of are handled by my title naming conventions. I'm trying to
think of examples where having two names and search of the body is
required. Of course, I may be reasoning all wrong about my data in the
first place too.
:) ... Like good tiddler names, finding a good name that fits, is
hard. ... Several iterations may be needed. ... The advantage here is, that
refactoring tags, most of the times also leads to refactoring the content.
Which imo is a good thing.
Surely this is where tagging is most useful to me? This isn't ad hoc like
search; it's more about automatically constructing and reconstructing. I
love the idea that changing names actually changes the content; boy does
that sound powerful. I must be using and thinking about my TW all wrong.
So, I think of tags in TW (perhaps incorrectly) as assigning a tiddler
multiple names, categories, or properties. I've somehow avoided the need,
done something by hand that should have been automated, or completely been
blind to the need for such a thing. Naming is hard. I spend a non-trivial
amount of my time asking myself where things belong in my TW and why.
Sometimes this is exactly what helps me map into new territory!
When I don't have a very specific title in mind, my naming conventions go
something like this:
`YYYY.MM.DD -- FooTag: BarTitle`
It's been pretty useful to me. I get a decent timeline, I kind of always
have one tag, and I can reuse BarTitles freely (a fairly common problem for
me) because I know the other information is unique. Sometimes I just think
of each tiddler has a flat textfile, and I'm trying to pack as much
necessary metadata in the name as I can; that may be the wrong approach.
Generally, I do not seem to need to factor the names of my tiddlers. The
evolution of my naming conventions is actually something I aim to highlight
and preserve in my work (maybe tags will help me do that better). Sometimes
I build trees of linked tiddlers. In a sense, the names don't matter
because the treelink structures have done the major hierarchical grouping
work. Tags could eliminate the need to crawl that tree in some cases.
Alright, practical tagging possibility for my context number one! Thank you.
In my wikis I use 1-2 tags. But not every tiddler needs a tag. If a
tiddler has more tags, they tend to cause too much "noise", because they
show up in too many different lists.
Alright, so I think this suggests I shouldn't be tagging willy-nilly ever
(good, since I don't want to endure that chaos). If I don't have a clear
systematic reason and direct purpose already in mind, then don't tag.
If you don't want to use the "tag-space" to create an outline or
toc-like structure, there are other possibilities. eg: my tocP plugin,
which uses user-fields to create the connection between different tiddlers.
That plugin is very pretty. For how I normally use my wiki, I wish I could
click a button and it would add a formatted link in the body of the parent
and created the new tiddler. Unfortunately, I'd usually have to edit the
BarTitle in both the parent and child. Your system avoids that work. It may
just be a better way to go about doing this in the end.
When is it better to use tocP over the tag-space?
Every thing with the same tag can be said to belong in the same group.
This is a way to organise tiddlers, and you can quickly find everything
else in the same group.
I appreciate that. So far, I've not bumped into a problem without it. But,
perhaps there are possibilities here I'm missing out on that would I use if
I realized it.
Imagine Work and Personal tags
I feel like I accomplish this with my linking structures. However, tagging
might just be a better approach. I'm failing to come up with an example in
my own wiki, but that may just be my blindness here.
People often use tags to represent a particular kind of tiddler
task note booking journal
I can see that, although probably not well enough. This too is something I
feel comfortable doing by hardlinking tiddlers. Maybe I shouldn't. I'm
trying to see where tags are silverbullets here. They might just be another
good way of doing things in this case.
And they are often also used to represent status
new working completed cancelled archived
I will need to think more about this. Temporary flagging is usage I had not
considered, but surely I should! I usually just put such a thing in the
body of the tiddler, but maybe I shouldn't. Thank you!
And then there is peoples names as tags, subjects or categories
I've tried this (perhaps incorrectly?). This is actually where I stopped
using tags. Links and search seemed to solve it faster for me.
Tags are a quick and easy way to add a great deal of organisation even
processes to your information without much fuss.
I am excited to try it. Perhaps that's my problem here. Maybe I've been
fussing over tiddler naming conventions to the point that I've missed to
obvious answer in tags. What do you think?
Tags can be a switch, if tagged or not?
Reminds me of representing status. I'm trying to come up with a case where
I'd want this particular one though. It's interesting.
In my most recent large wiki I keep tags for ad hoc categories or
flagging and have moved most into other fields some of which are tag like
fields, Just to cope with what would be to many tags.
Neat. So, tags act as stand-in placeholders until you figure out how to
want to refactor into other metadata structures. That sounds awesome; I
feel stupid for now seeing how I would make use of it. I've been using my
TW for quite a while now, and you'd think that kind of thing would be
obvious to me.
their use and number is unlimited
I see that. In a sense, I'm looking to see how to cut salient sculptures
out of the unlimited marble slab. Ideally, I'd like to find ways to
cost-efficiently use tags while maintaining most of my current conventions
(or modifying those conventions in virtue of a useful emergence from adding
tags to the mix). Practically, I may be in for a complete refactor.
I am very against prescriptive rules in something like tiddlywiki. If you
don't have a reason to use tags than you don't need to use them. There is
nothing magic about them that you can't do with tiddler namespaces or
fields. If you are comfortable with using fields than the tags field has
restrictions that can make it less useful than other fields.
I hear that. Look, I want you to be right. I also realize a ton of very
intelligent people use this feature of TW, and I don't. That doesn't make
me wrong, but I have to bet against myself here. I think I'm obligated to
try and find a way to see how to use their reasoning in my own.
What are examples of restrictions on tags that might be worth avoiding?
So, there are no examples of tags being the best and irreplaceable tool
because they are just an implementation of what you can do with fields
built into the core.
That's a good point.
My assumption has been that that field invisibility while reading is useful
for hiding information/clutter, while tagging gives you some visible and
clickable interface to navigate and reason about tiddlers without opening
them. This seems kind of related to why I try to make my titles do as much
reasonable metadata work for me as possible. I also worry tags have a
tendency to hide the very metadata I want to see while reading through
lists of tiddlers.
Tags outperform linking with advanced search when you don't have time or
don't have the experience with data systems and coding required to set up
another system
Oh. Hmm. Well, I feel pretty incompetent here. I interpret you to be
straight up telling me I need to start using tags because I have no idea
what I'm doing (which is a very fair thing to tell me!).
I'm quite worried I'm using my TW all wrong. I'm not trying to assume the
grass is greener on the other side, but I know how fallible I am (I'm
straight up stupid too often, dude). Yes, it works great for the way in
which I use it, but maybe the way in which I use it could be radically
improved if I were thinking in terms of tags. I have "Just a bunch of
tiddlers" carefully linked to each other in their bodies. It's been
profoundly flexible for me, but my gut says I'm missing something (which is
why I'm looking).
Tags don't really help you model anything that you can't with other
fields, they are just a shorthand method for the same thing
Fair enough. I do hope to make use of the shorthand and the tooling
ecosystem so I don't have to reinvent too many wheels.
What makes a good tag is completely application specific and possible
specific to the person using it, there isn't a general answer to this
I appreciate your contextualism here. If you happen to look at my wiki, do
you have any gut instincts about what would make a good tag for me?
The number of tags in a wiki to make it effective is the same as above,
there isn't a general answer. I have wikis with no tags that do complex
things and I my bookmarks wiki has thousands of tags.
One of my goals is to keep a unified wiki (with the exception of an
encrypted wiki because individual tiddlers do not appear to automatically
lock themselves [for good reasons, I take it]). What are some obstacles one
would face into trying to keep all of their TWs unified into one? Are there
good ways to combat those problems while maintaining a unified wiki?
IMHO the key disadvantage of TiddlyWiki's fields mechanism is that field
values dosn't show up in the search results...
Interesting. Is there no way around this?
Tags provide the equivalent of folders, and allow semantic context to be
affixed to tiddlers without despoiling the title. While it's true that much
of what you do with Tags could possibly be done with fields, tags have been
bestowed with special visibility and powers right out of the box. They're
automatically part of searches. If you change a title that is used as a tag
TW will offer to change it in tag and list fields. If you click on a tag,
it will show you all tiddlers tagged with that tag, and allow you to change
the list order.
I've yet to find a good way to use the Update "in the tags and list fields
of other tiddlers" so far. When I modify a tiddler title, I may need to
modify all the links to it as well, which is usually trivial. Are tags
meant to bypass that kind of problem?
I might be weird, but I want my titles to be pack a lot of metadata into
them. Do you think I need to start offloading some of my title work into
tags? Should I duplicate some of it tags instead (I'm not sure how that
helps me)?
A common searching style for me when I can't find a tiddler, is to search
for a related tiddler that I can remember. Then click on a tag that might
relate it to the one I can't find and see a list of tiddlers -- one of
which is likely to be the right tiddler.
This is where my link tree structures really shine, imho. I spend a lot of
time thinking about the hierarchy of my links. I've basically forced a
virtual file structure on myself. My "Root" directory really does contain
links to everything else on the wiki, and there are reasons for why every
link is placed there.
The problem mechanically with using links is that there is no core
mechanism to update your links if you change your title. The only good work
around is to use PMario's unilinks, which allows you to make changes via
the subtitle field and never have to touch the title again.
This is the best argument in the thread by far!
Short-term, it's usually in the creation of the tiddler that I made a
mistake in the name and have to go change it in both places.
The lack of a core mechanism here isn't too frustrating for me. When I
must, I search and replace the .html file with a text editor. But, and this
may just be a quirk of my project, I usually leave the breadcrumbs of such
changes to give myself evidence of the evolution of my wiki. Transclusions
appear to solve my problems so far (but, it's possible that I'm missing
something that will eventually bite me in the butt). Still, it may be
better to go another route and build that evidence differently. I could,
for example, attach the new tags to everything with the old, and continue
on just using the new. What do you think?
I've been thinking about PMario's unilinks as well, although not for what
you suggested. I may end up moving to it just to handle: Gator, gators,
Alligators, alligators, alligator, etc. pointing to the same tiddler.
Perhaps I should use it for the purpose you point out. TW does seem to
favor using non-link mechanics for directories.
What is the reason TW's core doesn't offer chance to update links in the
bodies of all tiddlers?
--
You received this message because you are subscribed to the Google Groups "TiddlyWiki" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tiddlywiki+***@googlegroups.com.
To post to this group, send email to ***@googlegroups.com.
Visit this group at https://groups.google.com/group/tiddlywiki.
To view this discussion on the web visit https://groups.google.com/d/msgid/tiddlywiki/0a528457-975b-4df0-a890-0091afe7103e%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.