|
|
Post by louisxiv on Dec 7, 2013 9:46:37 GMT
Yes, sorry I meant to say AitF to be in its correct order in the timeline. If other spinoffs appear they can be inserted in internal chronological order, or at point of appearance on-line (an appropriate chapter boundary anyway) if we have no other information.
File format:
I want to move away from a single line per page as that makes the file difficult to read for simple human beans like me. As it is I work on my copy with tabs set -wide- so I get clear-ish columns for better check-at-a-glance, but have to keep switching line-wrap on and off. It is unweildy.
The comic page type data line strikes me as a reasonably robust format, self-contained, easy to understand (?). I'm not too concerned about the extra typing for adding each item as I just copy, paste & edit anyway.
At the price of adding more structure to the file itself we could go down to pure type:value pairs and be a bit more rigid about formatting over multiple lines: so comicname:pageno pair defines the page to which subsequent type:value pairs apply, until the next comicname:pageno (or eof).
So the example becomes
gun:1824
tag: Annie
tag: CherryTree
tag: Habitat
tag: HabitatSymbol
desc: (slow zoom out from the tree)
snark: Here she is. There is where Annie was. Right here.
gun:1823
…
(leading whitespace for legibility)
My concern here is the extra structure to the file itself, with processing of lines of data being far more dependent on other lines, makes processing complex and the file itself less robust. The existing format, and the comic page type data have the advantage of being sortable, which years of faffing with text data file has convinced me is an important property (1). That being said though I do like the look of the whitespaced type:value pairs. Very readable.
(1. I suspect this preference masks an atavistic fear that someone is going to drop the text file and shuffle it.)
((Err, yes. I did use punch card decks, back in the day. Why do you ask?))
|
|
|
|
Post by louisxiv on Dec 7, 2013 9:47:37 GMT
It would be cool if there were an option to view it chronologically on gunnerkrigg.com, too. Yes indeed. |
|
|
|
Post by snipertom on Dec 7, 2013 13:10:26 GMT
Yes, sorry I meant to say AitF to be in its correct order in the timeline. If other spinoffs appear they can be inserted in internal chronological order, or at point of appearance on-line (an appropriate chapter boundary anyway) if we have no other information. File format:
I want to move away from a single line per page as that makes the file difficult to read for simple human beans like me. As it is I work on my copy with tabs set -wide- so I get clear-ish columns for better check-at-a-glance, but have to keep switching line-wrap on and off. It is unweildy. The comic page type data line strikes me as a reasonably robust format, self-contained, easy to understand (?). I'm not too concerned about the extra typing for adding each item as I just copy, paste & edit anyway. At the price of adding more structure to the file itself we could go down to pure type:value pairs and be a bit more rigid about formatting over multiple lines: so comicname:pageno pair defines the page to which subsequent type:value pairs apply, until the next comicname:pageno (or eof). So the example becomes gun:1824
tag: Annie
tag: CherryTree
tag: Habitat
tag: HabitatSymbol
desc: (slow zoom out from the tree)
snark: Here she is. There is where Annie was. Right here.
gun:1823
…
(leading whitespace for legibility) My concern here is the extra structure to the file itself, with processing of lines of data being far more dependent on other lines, makes processing complex and the file itself less robust. The existing format, and the comic page type data have the advantage of being sortable, which years of faffing with text data file has convinced me is an important property (1). That being said though I do like the look of the whitespaced type:value pairs. Very readable. (1. I suspect this preference masks an atavistic fear that someone is going to drop the text file and shuffle it.) ((Err, yes. I did use punch card decks, back in the day. Why do you ask?)) We could format it as JSON, or the coffeescript version thereof, which is i think identical to what you'd propose using whitespace coffeescript.org/#destructuringDid I mention I love whitespace?  Advantage of using the coffeescript version of JSON is that it can be automatically compiled into JSON object(s) which can then be sorted, and python has native support for JSON |
|
|
|
Post by louisxiv on Dec 7, 2013 18:56:21 GMT
The important question is whether the format – JSON / coffeescript / etc – it human-read & write-able as well as simply processable.
Will take a look at coffeescript link when I'm capable…
(Well, Saturday pm, old friends on a pre-xmas visit, and a pub, are all implicated…)
|
|
|
|
Post by louisxiv on Dec 8, 2013 0:36:14 GMT
Hmm. Well, I have a temporary bodge for listing AitF.
I’ve put in a dummy page number, for which the gunnerkrigg.com site will return the current main comic page if the resulting invalid link is followed. Then I put the true link in the comment in the usual markdown format. Thus:
aitf1-3 ch-31b Annie Anwyn Coyote Gillitie #[Page 3](http://www.gunnerkrigg.com/extracomics/comic.php?c=Annie%20in%20the%20Forest%20Part%201&p=3) Look everyone, here she is! Our visitor from the Court!
Ugly as sin, but it’ll have to do until I update the script.
Only three pages and the cover tagged so far, but I’ll try to fit the rest in real soon.
|
|
|
|
Post by louisxiv on Dec 14, 2013 12:07:47 GMT
Index page updated: 1321 pages indexed, 425 tags defined That's Annie in the Forest pt1 tagged in-line, and the main comic up to date (until Monday). snipertom: the github repo is still a bit behind - the latest couple of patches to alltags.txt and index.txt (showing as patch-12, patch-13 in the network graph) roll up my previous changes (patches 14 to 16 - which I have not yet found how to delete).
|
|
|
|
Post by louisxiv on Dec 24, 2013 18:19:08 GMT
Just had a report that the index page was broken as of 2013-12-24 16:07 GMT. I'm not seeing problems at 18:07 - has anyone else noticed errors on, particularly, the javascripts?
|
|
|
|
Post by louisxiv on Dec 24, 2013 19:07:25 GMT
Just had a report that the index page was broken as of 2013-12-24 16:07 GMT. I'm not seeing problems at 18:07 - has anyone else noticed errors on, particularly, the javascripts? Ah-ha! A typo by the person reporting the problem had revealed a SNAFU of mine on a different page. Now fix'd |
|
|
|
Post by louisxiv on Jan 1, 2014 16:18:12 GMT
Character, etc, index page updated: 1329 pages indexed, 425 tags defined. (Includes Annie in the Forest part 1) No substantive progress on the index file format change discussed previously: I had pencilled in some work on this for the holiday, but other stuff was more pressing. I am still thinking about it and it is still in the ToDo pile. |
|
|
|
Post by snipertom on Jan 20, 2014 11:47:26 GMT
Last 2 pages up.
Wondering if I should tag AitF2 since I have it... or would that be too much of a spoiler?
|
|
|
|
Post by GK Sierra on Jan 20, 2014 17:08:08 GMT
Wondering if I should tag AitF2 since I have it... or would that be too much of a spoiler? If you wanted to play it safe, then it would probably be best to wait until Tom puts it up on the website. I don't know anything about Github, but it doesn't seem like there would be a "spoiler" feature readily available. |
|
|
|
Post by louisxiv on Jan 20, 2014 19:04:17 GMT
Last 2 pages up. Wondering if I should tag AitF2 since I have it... or would that be too much of a spoiler? Tag it locally in readiness, but don't drop it into the index until the comic comes online. The index is to visible, linkable stuff. Anyone can check and correct us if necessary. Tagging stuff that isn't so available loses that a little. |
|
|
|
Post by louisxiv on Jan 20, 2014 20:25:56 GMT
Page updated.
I've simplified the tag name: ROTD rather than RealmOfTheDead – and taken out some redundant Court tags because I thought we were only using it for significant external views, not day-to-day rooms?
Also: think I may finally have the hang of the shallow end of the GitHub pool now - I did this as a full repo Fork/Edit/Pull request rather than editing the files directly in GitHub (which does the same, but more behind the scenes?).
|
|
|
|
Post by snipertom on Jan 21, 2014 0:45:01 GMT
louisxiv GK SierraLocal tagging ftw, agreed  louisxivI'd been assuming all the non-specific Court places should be tagged as "court" due to each page needing to have a location. It doesn't really matter a huge amount in some ways because you can assume that anything that isn't tagged with a location is in a non-specific location of the court anyway! BTW guys I've FINALLY gotten around to starting to fix up handler.js so that it a) works on mobile and b) can handle selecting more than one tag at once. I'm not putting it up on github just yet for a bunch of reasons including that there will need to be some (relatively minor) site structure changes. The gist of it is that each tag needs to be wrapped in an <a href="#" class="tag apples">Apples</a> kind of thing (and the js will obv preventDefault() so that it doesn't just jump to the top of the screen). Anyway I'll keep you posted!
|
|
|
|
Post by louisxiv on Jan 22, 2014 0:29:18 GMT
[…] assume that anything that isn't tagged with a location is in a non-specific location of the court anyway Yes, that’s pretty much the default position. I’m sure we discussed and agreed it here a while back. Ah, now that sounds good :) Fortuitously, I'm just setting up to do some work on the index & tags processor script, at last, with the intent of generalising it so the same script handles Gunnerkigg, Misfile, and perhaps other comics if someone fancies taking a crack at them. Starting with making the input and output files params rather than hard coding, and then file format changes discussed a little while back. Anyway this makes now a relatively good time to look at changing the tag wrapping. Suggestion: once you've got multi-tag selection, how about inverted tag selection – the “all but these” selection? |
|
|
|
Post by jda on Feb 11, 2014 20:58:57 GMT
Hi, guys! First of all, I'd like to say :WOW!, what a great job you have made tagging almost everything taggable in the Comic. And then: I am thinking on a way to provide us wannabe translators of GKC a solid index of all the chapters, pages, panels, etc, AND the full transcribed text, so it may be of help to anyone translaing to a new language.
So, would anybody here kindly suggest how can I get such an index OR even if is there any full text transcription already on development?
Thanks a lot for your time!
|
|
|
|
Post by louisxiv on Feb 11, 2014 23:06:44 GMT
I have had a brief look at transcribing the text and making it searchable. I have put in the first 50 pages with the intent to use it for testing, but the index script rewrite is rather further ahead in the priority queue at present. 1 Annie My name is Antimony Carver. I would like to share with you the strange events that took place while I attended school at...
1 Title Gunnerkrigg Court Chapter 1: The Shadow and The Robot
2 Annie Gunnerkrigg Court does not look much like a school at all.
2 Annie It closer resembles a large industrial complex than a place of learning.
2 Annie Within the first week of my attendance, I began noticing a number of strange occurrences.
2 Annie The most prevalent of these oddities being the fact that I seem to have obtained a second shadow.
2 Notice NO RUNNING
3 Annie At times I would see this creature jump from shadow to Shadow
3 Annie once in a while hiding in the dark recesses of the Court.
3 Annie This made difficult my attempts to convey the situation to members of the faculty
3 Annie as my uninvited guest seemed to shy away from attention.
3 Annie More often than not, however, this Grim Spectre matched my every step.
3 Notice Study
3 Annie One evening, during solitary study
3 Annie I decided to confront the being.
3 FX snapThe intention is to feed it into a small php search function so people can look for pages containing "shadow" and get the word in context, a page number (with link), and who said it, e.g.: 1 Title Gunnerkrigg Court Chapter 1: The Shadow and The Robot
2 Annie The most prevalent of these oddities being the fact that I seem to have obtained a second shadow.
3 Annie At times I would see this creature jump from shadow to Shadow
There is a text search page for Misfile which works that way. The obsessive analysers and commentators on the forum there love it… Anyway, if anyone does have, or generates, a more complete transcript I’m very interested too. |
|
|
|
Post by louisxiv on Feb 11, 2014 23:39:33 GMT
I'm sure you could "crowd source" a dialogue transcription effort in the same way the first push on the index was done.
|
|
|
|
Post by exdevlin on Feb 12, 2014 0:04:56 GMT
I wouldn't mind re-reading through the chapters and doing a bit of chronicling as I go.
|
|
|
|
Post by snipertom on Feb 19, 2014 16:20:38 GMT
Just calling in here to say that I'm still working on things! Haven't disappeared, just been busy!
|
|
|
|
Post by louisxiv on Feb 19, 2014 23:09:18 GMT
|
|
|
|
Post by snipertom on Feb 23, 2014 14:24:20 GMT
 hehe! Well, I should have a little demo for you guys to play with soon, which will be followed shortly by one that is actually very similar to the old site. In other words- the search box in my demo can search for more than one thing, and multiple tag selection is not far away In regards to being able to search for 'not' a certain tag, I was wondering if cycling through 3 different states might not be a bad idea- unselected, selected ('and'), crossed-out ('not') Thoughts?
|
|
|
|
Post by louisxiv on Feb 23, 2014 19:10:16 GMT
(:)) hehe! Well, I should have a little demo for you guys to play with soon, which will be followed shortly by one that is actually very similar to the old site. In other words- the search box in my demo can search for more than one thing, and multiple tag selection is not far away :D In regards to being able to search for 'not' a certain tag, I was wondering if cycling through 3 different states might not be a bad idea- unselected, selected ('and'), crossed-out ('not') Thoughts? Sounds cool! It'll be interesting to see how well it works UI-wise for the 'not'. Looking forward to it. |
|
|
|
Post by louisxiv on Feb 23, 2014 20:49:19 GMT
Index page updated: 1351 pages indexed, 450 tags defined snipertom isn’t the only one who has been busy… This update was generated from the new-look index.txt by the new cig.py (Comic Index Generator – hmm, though maybe I’ll rename that Series Index Generator just for the Captain Scarlet reference… ). For comparison here’s the old index page and the old index.txt. Please shout (then post) if you see anything odd on the new page. In a few days, if no major SNAFUs crawl out of the conversion, I’ll push this up to GitHub and put cig.py up there somewhere too.
Slightly technical stuff: As mused upon earlier, I have changed index.txt to a simpler, more flexible format: page:845
url:http://www.gunnerkrigg.com/?p=845
tag:ch-32
tag:AntimonySymbol
tag:Dorm
tag:Rey
tag:SubDorm
desc:(Reynardine is waiting, looking outside, sadly, in the Submarine dorm)
page:844
url:http://www.gunnerkrigg.com/?p=844
tag:ch-32
tag:Cover
tag:Flowers
desc:Chapter 32:<br><br>**From the Forest She Came**<br><br>_Icon: SubDorm_
page:AitF-32
tag:ch-31b
tag:Annie
tag:Gillitie
tag:Irial
tag:JumpersPool
tag:Kamlen
desc: Annie, this here is my love, Irial
url:http://www.gunnerkrigg.com/extracomics/comic.php?c=Annie%20in%20the%20Forest%20Part%201&p=32
page:AitF-31
tag:ch-31b
tag:Annie
tag:Gillitie
tag:JumpersPool
tag:Kamlen
desc: Annie!
url:http://www.gunnerkrigg.com/extracomics/comic.php?c=Annie%20in%20the%20Forest%20Part%201&p=31
page:AitF-30
The big change, really, is the addition of the url: type for each page, rather than assuming a single url structure, so it is easy to handle diversions like AitF and comics which use date and page in their urls – forex: Jay Naylor’s Original Life: archives/2014/0 2/493.html (um, some nsfw thereabouts though that page is safe).
|
|
|
|
Post by Daedalus on Feb 24, 2014 1:36:35 GMT
You aren't lying :\
That is certainly not a site I ever expected to see linked to here  |
|
|
|
Post by snipertom on Feb 24, 2014 10:51:57 GMT
You aren't lying :\
That is certainly not a site I ever expected to see linked to here a furry comic? |
|
|
|
Post by louisxiv on Feb 24, 2014 11:49:15 GMT
You aren't lying :\
That is certainly not a site I ever expected to see linked to here a furry comic? Not Safe Fur Work |
|
|
|
Post by snipertom on Feb 24, 2014 12:24:19 GMT
Mostly done, except for the internal links (which should be trivial), and I haven't implemented the unselected->selected->'not' thing just yet. You can preview this here: codepen.io/snipergirl/pen/bHGnt(Basically click on the title of each tab to show the 'compiled' html/css/js) There are some structural changes to the html that make it work on mobile devices (essentially replacing "span" with "a href='#'"), as well as general improvements to the handler.js I've combined event.css, all.css and some css that was in the html of the page itself into the css you see in the 'css' tab so that it's all in one place and thus less confusing. The HTML is only referring to weird amazonaws links because that's where assets are stored on codepen. |
|
|
|
Post by Daedalus on Feb 24, 2014 14:03:35 GMT
You aren't lying :\
That is certainly not a site I ever expected to see linked to here a furry comic? I have no problem with furry comics. Not even most of the nsfw ones. However, Jay Naylor in particular does some pretty damn creepy stuff in the adult section of the site. |
|
|
|
Post by snipertom on Feb 24, 2014 14:19:11 GMT
I have no problem with furry comics. Not even most of the nsfw ones. However, Jay Naylor in particular does some pretty damn creepy stuff in the adult section of the site. Oh god I just read the encyclopaedia dramatica entry for him. |
|
hehe!
