Searchable database of comics?

[louisxiv]

Jun 12, 2013 9:34:57 GMT snipertom said:

because people won't automatically know what a char or loc is?

Fair enough on the abbreviation, but I already tell 'em to click on the pretty coloured blobs so I don't think adding a para to the howto:

"Some tag descriptions are further flagged with to words in [] indicate their general type, so you can search for a class of tagged thing by using the text search box. Search for:
[who] - to find tags associated with people
[where] - finds places
[what] - finds things"

is to too much to ask people to absorb. They may even notice it for themselves.

The nice thing about it is that we can experiment with these to find if they are actually useful without the overhead of rewriting any step. Just include 'em in the existing tag definitions.

As I've asked before - as we've got text search/select now do we need specific tag processing anyway?

[louisxiv]

Character Index updated: 969/1208 pages indexed, 304 tags defined.

Corrects last night's little problem of duplicate input, also adds tag occurrence counts as [321] and id attributes on all tag defs (lowercase tag name) and page entries (p123) for crosslinks and references.

[snipertom]

I think the tag selectors are useful for a bunch of reasons:
- they function as an autocomplete for the search
- it allows you to explore things that you don't know the name for
- it allows you to explore related things from the same posts easily (eg having searched for the etherium, maybe now you're interested in blinkerstone)
- they power the "scroll to first instance" thing
- it means you don't have to use the keyboard, which from a UI perspective is great and means you're not interrupting exploration

[philman]

Yeah still in snipergirl's camp here. Clickable tags and typable tags are both useful for different reasons.

And I wouldn't be opposed to the [who] [where] etc style of descriptors, but still like the idea of being able to organise tags by class easily. If we had descriptors llike that in the tag description, but made a short clickable list at the top of the page (going across the page rather than down it in list form maybe) then people could find all the places, people, themes, etc. Using descriptors like that would make it easier to have a few extra classes too. I still think it would be useful to be able to class all the themed tags, such as forest or robots for example.

Also we seem to only have 7 chapters left to go! 9,11,13,15,17,19 and 31. I'll go through and do a couple of them tonight, don't know if anyone has started any of them yet though.

[snipertom]

Finally finished 18, doing 9 now.

Whee! past the 80% mark

[louisxiv]

Character Index updated: 1002/1208 pages indexed, 306 tags defined.

(gun4py.txt up there now too)

[louisxiv]

Snipergirl & philman's points taken that tags and search serve overlapping but different purposes.

[snipertom]

Chapter 9, 11 are done now which means we're up to 85%!

[louisxiv]

Character Index updated: 1035/1208 pages indexed, 315 tags defined.


(I spent some time this evening playing with using internal linking for quick navigation of chapters: results are fine, but I'm not entirely happy with the input side, too much embedded html is creeping in and it is too easy to break: looking at lighter weight markup and formatting (via Markdown) now.)

[philman]

31 (long chapter!) and CityFace 2 done.

Tried to merge but snipergirl obviously had done something while I was editing and github went crazy telling me I have a detatched head or something

[philman]

and now it won't let me do anything again unless I delete all the files and redownload them, removing the last chapters work. argh.

[philman]

Ok, I have the files, with both my changes and Snipergirl's most recent ones, but it still won't let me merge or sync anything. I don't even know what's wrong. Just keep getting messages about Detatched Head eb0a542

[snipertom]

ok, what you do is-
1) copy your current altered files "index.txt" and "alltags.txt" into a new directory
2) delete everything including hidden files and folders in the old directory
3) github will say that the local copy is missing, so select "clone again" and it'll redownload everything
4) copy back "index.txt" and "alltags.txt" from the new directory (ie that has your changes in it) into the original one (where you cloned it again)
then you should be able to upload.
basically all the version information is in those .git* folders and files, once they're deleted it doesn't show "detached head" anymore however github needs *some* version information, hence why it has to redownload everything. unfortunately it needs the folder to be empty before it will clone it again.

[snipertom]

Jun 13, 2013 22:28:11 GMT louisxiv said:

Character Index updated: 1035/1208 pages indexed, 315 tags defined.


(I spent some time this evening playing with using internal linking for quick navigation of chapters: results are fine, but I'm not entirely happy with the input side, too much embedded html is creeping in and it is too easy to break: looking at lighter weight markup and formatting (via Markdown) now.)

Sounds intriguing and exciting! Which bits are you linking and how? Look forward to checking it out!

[philman]

That was my thought after I had shut my laptop in anger and gone to bed in a huff last night too snipergirl!

Ill have to wait for tonight now since I'm on a phone now

[snipertom]

Jun 14, 2013 7:49:02 GMT philman said:

That was my thought after I had shut my laptop in anger and gone to bed in a huff last night too snipergirl!

Ill have to wait for tonight now since I'm on a phone now

Oh github ate my chapter 9 last night and I almost did the same thing. RAR!

EDIT: Chapter 13 just uploaded. So 1066/1210 are up which means 88% done, plus when Philman's ch31 is up that is 92% gosh golly YOU GUYS HAVE BEEN AMAZING

[jasmijn]

By the by, this thread can lose the question mark in its title, what with the searchable database of comics actually existing and everything.

[philman]

Jun 14, 2013 9:38:19 GMT jasmijn said:

By the by, this thread can lose the question mark in its title, what with the searchable database of comics actually existing and everything.

Might be a good idea to put a large prominent link to the current version of the searchable database at the top of the opening post too, so people clicking here for the first time see the current version and don't get put off by thinking it is still pie in the sky!

EDIT: Also I expect the initial tagging to be finished this weekend I guess, given we have only a couple more chapters to do! Then starts the more onerous process of checking everything. Do we have an outline plan for how we're doing that yet?

[snipertom]

Egads, I need to finish the tag-helper-thingy. SHould be able to do that this weekend so that we have something that will at least make it slightly easier/less painful to do it.

Basically at the moment it has a dropdown list of all the comics w/ descriptions at the top; then the selected comic on the left; then the tags currently used in big on the upper right, then other tags (with tag counts) below that.

My only concern is in regards to it showing comics but not ads etc and whether that would be unethical/illegal/break tom's heart. however i only plan to distribute this to maybe 5-10 people tops and not make it publicly available so...?



EDIT: yes, yes I know it's ugly at the moment.

[snipertom]

OK, most of the above works now;
Need to add a "copy" button still
"113" ends up often causing a glitch as it gets alphabet sorted before a lot of the comic numbers (should be easy enough to fix though)
obviously need to make things more readable
It autofills with the already existing tags but there's no way to deselect those tags at the moment (might be useful though if we're only ADDING rather than removing tags?)
Also can't add a new tag as yet

And of course, you need to have a current copy of "index.txt" and have python installed so as to be able to run my python script

[louisxiv]

Character Index updated: 1066/1209 pages indexed, 318 tags defined.

[philman]

updated Ch 31 and cityface 2 from last night. not doing anymore tonight tho.

You program/script thing you put int he post above looks interesting snipergirl, but I agree, having the comic image in the page does seem wrong, I can't imagine Tom or other comic maker really like their images being put on other pages like that, even if it is only a private page. I guess if it is only used by a few people, temporarily while we're working on the project?

[snipertom]

alternative would be to embed the whole page as an iframe; in that case we have all the ads and everything, shouldn't be a big deal to get that in there!

EDIT: iframe now working, doesn't look any worse than the original version!

EDIT EDIT: 94% done! Only 2 chapters to go!

EDIT EDIT EDIT (HETTY!): this is what the tag-helper looks like now:

[louisxiv]

Character Index page updated: 1147 pages indexed, 324 tags defined.

Spotted some problems

Tag error:
BobEyes and Bobeyes - capitalisation must be consistent or you get two tags

Index errors:

#it is not the meddling - showing as a tag? Probably an extra tab at the end of the index.txt line p817?

Duplicate pages: 788 & 789

(Sorry: normally I'd just tweak the page appropriately then submit a fix to GitHub, but I've just cycled 72 miles so I'm a bit too cream-crackered to trust myself to fix them properly)

[snipertom]

fixed all the above.

72mi is a long way! good ride?

EDIT: I've also gone through all the undefined tags and defined them, and consolidated wherever there's an unnecessary duplicate tag (eg "bone" and "bones")- plural has been preferred throughout for consistency

[louisxiv]

Character Index page updated: 1145 pages indexed, 348 tags defined.

Thanks for the fixes snipergirl.

The ride is a local fun run, I've done it a few times before, no problem. I'm no sport cyclist, just riding a commuter bike.
Unfortunately traffic and other circumstances kept speed way down at first which I wasn't expecting and was underdressed, so I chilled down enough (cold wind, rain) to make the first half of the run miserable and exhausting to the extent that I very nearly bailed out.
When I could let rip later, riding in front of the wind, it was rather more fun

[louisxiv]

I’ve added Markdown and Smartypants processing to the latest gunN.py, allowing for simpler markup and consistent use of smart quotes, ellipses and en and em dashes amongst other things.

So I’ve run the page through the new process and put it up on the site for people to check out. So far I’ve only noticed an issue with the text search: searching for something “'s” (with a dumb apostrophe) does not find something “’s” (with a smart right-single-quote).

[louisxiv]

Quick Markdown markup:

(very simplified - see Daring Fireball for fuller details)

For _italics_ surround the thing you want italicised between underscores, and for bold use **two asterisks** around your strongly emphasised word or phrase.

For a link to a tag, e.g. in the tag descriptions to cross-link Bob (Sutton)’s description the markup is

Bob Sutton, a gardener/maintainer of artificial [habitats](#habitat) and keeper
of [Young’s Park](#youngspark) with his wife, [Marcia](#marcia)

– which makes use of the tag internal links I set up the other day: to link to a tag just use #tagname as shown, similarly to a page use #p1234.


I’ve uploaded an version of the index page with some example crosslinks.

(One tiny problem-ette is that browsers put the linked item at the top of the screen when jumping to it, where snipergirl’s scroll-up bar then drops down over it. I've tweaked the css so the topscroll bar is now just a blob on the right when it appears.)

[louisxiv]

HTML musing:

I was already thinking about Definition Lists for the tag listing, but I see they've been redefined as Description Lists for HTML5 so it seems like a semantic HTML no-brainer.

But if so, what does that require from the .js and .css to present/select things as we do heretofore?

[louisxiv]

Another musing: can we / should we get this discussion/thread pinned so people can find it easily? Or…

I just doubt that discussions over the how-to's and fine detail of tagging and presentation are of interest to general readership, so perhaps a new thread just announcing updates, radical format changes?