Proposal: Wiki Page Name Standardization

Do you have a fully fleshed idea that you think Baka-Tsuki should adopt? Post it here.

Moderators: thelastguardian, Fringe Security Bureau, Senior Editors, Senior Translators, Alt. Language Translator/Editor, Executive Council, Project Translators, Project Editors

User avatar
cloudii
Project Translator
Posts: 512
Joined: Mon Mar 25, 2013 4:30 pm
Favourite Light Novel: Ahouka!
Location: awkward buttface
Contact:

Proposal: Wiki Page Name Standardization

Post by cloudii »

EDIT: Skip to this post: viewtopic.php?f=16&t=8839&p=225814#p225814

The material here in the first few posts of this thread won't work, so ignore it for now.

Original Post:
---------------------------------------------------------------
Hello everyone,

As we all know, TLG installed the Language Extension to the Wiki with the last update.

Many of us have been confused about what it could be used for, but it occurred to me that we could use it restructure Baka-Tsuki to the standard alt. language organization on all other MediaWikis. If you don't know what I'm talking about, you know how on Wikipedia, you can click on the change-language button, and any page will show up in a different language?

Image

With the Language Extension, this capability would be fully available on Baka-Tsuki. (Notice how there's a change-language button at the top of BT right now?) You could set your language preference, and it would selectively/automatically display pages in the priority of your preferred language. Of course, this doesn't work right now... @____@

In order to do this, we would need to restructure Baka-Tsuki's alternative language structure to conform with MediaWiki standards.

Meaning, instead of making the alternative language pages their own page, the standard in MediaWiki is to use a sublanguage extension. For example, instead of:

Code: Select all

(French) http://www.baka-tsuki.org/project/index.php?title=Un_Simple_Sondage
(English) http://www.baka-tsuki.org/project/index.php?title=A_Simple_Survey
MediaWiki's standard is to do the following:

Code: Select all

(Spanish) http://www.baka-tsuki.org/project/index.php?title=A_Simple_Survey/es
(French) http://www.baka-tsuki.org/project/index.php?title=A_Simple_Survey/fr
(English) http://www.baka-tsuki.org/project/index.php?title=A_Simple_Survey
What are the advantages of this?

1). It makes use of the Language Extension, because that's how it functions. If pages aren't organized that way, the extension is useless.

2). The Language Extension has some code and functions dependent on this organization. For example, if you set your language preference to French, and when you click on the Sword_Art_Online page, the Wiki will automatically display Sword_Art_Online/fr if it exists. Otherwise, it will display the English. Note***

Any opinions?

Note***: Technically, you actually also need the Extension:AutoLanguage. IDK if TLG installed it. But if he installed the extension that allows you to set your language preference... it's kind of useless unless you install it together with the AutoLanguage extension...(?)
Twitter: @cloudiirain | BT Userpage | OreShura Translator | Biblia Editor (@HereticLNT) | Clockwork Editor (@HereticLNT)
User avatar
Mystes
Heaven's Blade Successor
Posts: 15932
Joined: Thu Aug 05, 2010 6:54 am
Favourite Light Novel:
Contact:

Re: Proposal: Alt Language Page Infrastructure Change

Post by Mystes »

Speaking of which, can we establish a standard for the series titles?

There are some fully-translated ones, some furigana-read ones, etc...
Kira0802

#campione at rizon for some #campione discussions~~ And other stuffs.
User avatar
darkoneko
Astral Realm

Re: Proposal: Alt Language Page Infrastructure Change

Post by darkoneko »

That looks like a pretty nift idea.

Oh, by the way I've started revamping the category system.

-- darkoneko
KuroiHikari
Fish Miner
Posts: 822
Joined: Fri Apr 16, 2010 1:01 am
Favourite Light Novel:

Re: Proposal: Alt Language Page Infrastructure Change

Post by KuroiHikari »

Kira0802 wrote:Speaking of which, can we establish a standard for the series titles?

There are some fully-translated ones, some furigana-read ones, etc...
I thought we're good with letting the translators decide that.
User avatar
cloudii
Project Translator
Posts: 512
Joined: Mon Mar 25, 2013 4:30 pm
Favourite Light Novel: Ahouka!
Location: awkward buttface
Contact:

Re: Proposal: Alt Language Page Infrastructure Change

Post by cloudii »

Kira0802 wrote:Speaking of which, can we establish a standard for the series titles?

There are some fully-translated ones, some furigana-read ones, etc...
In my opinion, if we end up implementing this system, the main location of the page should be the romanji.

That way, it doesn't discriminate between languages.

However, I don't care if we have redirects that take The World Only God Knows -> Kami nomi whatever whatever. I also don't care if the Heading 1 title of the page is "The World Only God Knows". That can be project preference.

But if the URL is going to be shared across languages, I think its more fair to go with the romanji.

Just my opinion though. xD
Twitter: @cloudiirain | BT Userpage | OreShura Translator | Biblia Editor (@HereticLNT) | Clockwork Editor (@HereticLNT)
User avatar
cautr
Tsuruya Gigglenator
Posts: 740
Joined: Mon Sep 16, 2013 10:41 am
Favourite Light Novel:

Re: Proposal: Alt Language Page Infrastructure Change

Post by cautr »

Kira0802 wrote:Speaking of which, can we establish a standard for the series titles?

There are some fully-translated ones, some furigana-read ones, etc...
I'd like series title in Jap. put in roumaji as the default. Exception: Series titles in katakana, that are meant to be English anyway (Hundred and the like).
Translator at guhehe.TRANSLATIONS

"Life's a tale, told by an idiot." -Shakespeare
User avatar
Misogi
Supreme Lord Temporal
Posts: 4119
Joined: Sun Mar 18, 2012 5:19 am
Favourite Light Novel:
Location: France

Re: Proposal: Alt Language Page Infrastructure Change

Post by Misogi »

I'm not fond of having the project page's language based on the wiki's.

That may be a wrong opinion, but still.
Twitter : @MisogID
----
Sans plume, ni sanité (Without feathers nor sanity), a blog reeking of sickness. Wipe your minds before entering.
User avatar
cloudii
Project Translator
Posts: 512
Joined: Mon Mar 25, 2013 4:30 pm
Favourite Light Novel: Ahouka!
Location: awkward buttface
Contact:

Re: Proposal: Alt Language Page Infrastructure Change

Post by cloudii »

Misogi wrote:I'm not fond of having the project page's language based on the wiki's.
I'm not quite sure I understand what you're saying? @____@;


Anyways, I did so more serious reading, and it looks like the Extension:AutoLanguage isn't being maintained and won't work with our newer version of MediaWiki. T________T;

I did some cursory searches, Extension:Polyglot does the same thing and appears to work with newer versions (based on feedback), but is experimental. It's designed to work in tandem with Extension:LanguageSelector that TLG does have installed (which is that language selector you see at the top).

Anyways, MediaWiki isn't my forte, but that's my report on the automatic redirecting language settings. Honestly, the code itself is pretty short... I glanced at the php @___@; It may be implementable if someone went and contacted the maintainer... possibly...?

Regardless, the Translate Extension that TLG did install will only be useful/meaningful if start naming our pages in the format:

Code: Select all

page/$lang
So I'll leave this thread hanging for now.
Twitter: @cloudiirain | BT Userpage | OreShura Translator | Biblia Editor (@HereticLNT) | Clockwork Editor (@HereticLNT)
User avatar
EusthEnoptEron
Project Translator
Posts: 836
Joined: Mon Jul 13, 2009 10:39 am
Favourite Light Novel: Ahouka!
Location: Switzerland

Re: Proposal: Alt Language Page Infrastructure Change

Post by EusthEnoptEron »

(This opinion is written without deeper understanding of how the extension works -- please bear with me.)
If I understand Misogi correctly, he has the same concern as I do: linking the overall wiki language with the project language. E.g. my language setting is Japanese and I wouldn't want to get out of my way setting it to German just to see a German project (of course, you could access it directly). Likewise, if my language were set to German, I may not want to automatically arrive on the German project page when following the project link in the navigation. In other words, it generates somewhat untransparent behavior (under which conditions/settings do I arrive on which page?) that may wind up being more confusing than helpful.
Alternative language projects aren't simple 1:1 translations of their English counterpart, but full-fledged, independent entities in the Baka-Tsuki universe (IMO).

That being said, I'm all for those naming conventions.
User avatar
cloudii
Project Translator
Posts: 512
Joined: Mon Mar 25, 2013 4:30 pm
Favourite Light Novel: Ahouka!
Location: awkward buttface
Contact:

Re: Proposal: Alt Language Page Infrastructure Change

Post by cloudii »

EusthEnoptEron wrote:(This opinion is written without deeper understanding of how the extension works -- please bear with me.)
If I understand Misogi correctly, he has the same concern as I do: linking the overall wiki language with the project language. E.g. my language setting is Japanese and I wouldn't want to get out of my way setting it to German just to see a German project (of course, you could access it directly). Likewise, if my language were set to German, I may not want to automatically arrive on the German project page when following the project link in the navigation. In other words, it generates somewhat untransparent behavior (under which conditions/settings do I arrive on which page?) that may wind up being more confusing than helpful.
Alternative language projects aren't simple 1:1 translations of their English counterpart, but full-fledged, independent entities in the Baka-Tsuki universe (IMO).

That being said, I'm all for those naming conventions.
Very valid point. I just thought it would be nice to be able to click on the sidebar (or in the future, the categories... to cut down on the number of language-centric categories we need) and have it arrive at your language page instead of the english default. @___@; But I'm not an alt language individual, so I don't know how much more or less convenient that is. This should be something the alt. language staff should decide on.

It was my understanding that the redirect behavior only occurs if you don't provide a "/" or "/$lang" (I may be wrong though ^^; I'm kind of dumb with code). Hence, there would be no unexpected redirects if the interwiki links on pages were more specific. ie: [[Oreshura_Volume1_Chapter2/fr]] or [[Oreshura_Volume1_Chapter2/]] (slash makes it behave like forced "en"). In that kind of situation, the behavior would be identical to how links work right now. But that might not be immediately intuitive to our members if we made that change.

Anyways, let's drop the idea of the redirecting extension. We don't even know if it'll work, anyways.

Should someone email/ask Oni/TLG if it's okay to change the naming structure so the Translate Extension can actually function properly? At the very least, the chapter pages between projects have to have the same title for TLG's extension to work.

EDIT/CORRECTION: It looks like the Polyglot Extension doesn't currently work for interwiki links (won't apply redirects) ... so that pretty much makes it useless, (unless we decide to keep the sidebar with full URL links... which according to KH the sidebar is change anyways, so...)? XD Well, anyways, that pretty much kills that idea until some developer decides to work more on it. XD
Twitter: @cloudiirain | BT Userpage | OreShura Translator | Biblia Editor (@HereticLNT) | Clockwork Editor (@HereticLNT)
User avatar
cloudii
Project Translator
Posts: 512
Joined: Mon Mar 25, 2013 4:30 pm
Favourite Light Novel: Ahouka!
Location: awkward buttface
Contact:

Re: Proposal: Wiki Page Name Standardization

Post by cloudii »

Posting again for a more thorough proposal:

Linking this old thread (locked).

Purpose:

Standardizing page names make it a lot easier for API parsers. A lot. Really a lot. @_____@;

Can you imagine trying to extract chapter/volume information when there's so much variety:
  • Some pages have a space, dash, paren, or something, while others don't. EG: Oreshura:Volume_1_Chapter_1 vs. OniAi:Volume1_Chapter4? Normally I would try using split('_') or something like that to start the text parsing, but it wouldn't work very well with the lack of standardization. T_____T
  • Some pages uses abbreviations for the series title, and others don't. EG: (Some of Oreshura's pages are abbreviated, and others are the full Ore_no_kanojo_blahblah)
  • Its difficult to filter to for alternative language translations. eg: (Francais) vs (francais) vs (french) vs (Français) are all different things to a parser...........
It's just not very pretty without standardization.

Cloud's Point:

Let's make a few guidelines for file naming conventions. This would apply to images as well. It would also go along with the network of HELP pages I'm producing, so that way new members would most likely see the recommended naming structure (along with the recommended Project Page format, etc) when they're just starting out.

That way, it makes the lives at the BT app development side easier. Also, it makes everything a lot more predictable. Hmm?

Proposal: Use Language Codes

This would solve the above third point. If we start using (fr), we won't have to worry about: (Francais) vs (francais) vs (french) vs (Français).

My possibility would be something like this. You get the idea:

Code: Select all

Gekkou|fr:Volume1_Chaper3
Proposal: All Pages in a Project Have the Same Prefix

Solves point two. This way, when we're looking for pages in API, we just have to search for pages that start with : "Sword_Art_Online_(fr):" and feel comfortable that we're not missing any pages that were named differently.

I know that some titles are really long, and we want abbreviations. However, these abbreviations should be plainly standardized. It gets troubling when there's several variations of the abbreviations. So here's my proposal:

A MAXIMUM of two prefixes for all pages in a project. The full prefix is the title in romanji (unless the title obviously should be English, like Accel World). And then there's an optional abbreviated prefix, like OreShura. The abbreviated prefix needs to be standardized. It'd be a pain to see "Oreshura", for example.

Also, the project URL (aka prefix) should be all latin characters. Can't encode some of those weird characters.

Proposal: (Hypothetical) Possibly Use Hierarchal Structure(?)

Did you know this works on MediaWiki? Something like:

Code: Select all

Sword_Art_Online_(fr)/Volume_1/Chapter_3
The cool thing is that we get this thing on the header:
Image

Automatic, built-in navigation! Which is pretty cool, in my opinion.

Anyways, in order for this to happen, TLG needs to enable the subpage setting for the main namespace. It's currently only on for the the User namespace and usertalk namespace. But it's a thought.
Twitter: @cloudiirain | BT Userpage | OreShura Translator | Biblia Editor (@HereticLNT) | Clockwork Editor (@HereticLNT)
Cthaeh
Yuki-Nagator
Posts: 647
Joined: Sun Nov 11, 2012 6:54 pm
Favourite Light Novel:

Re: Proposal: Wiki Page Name Standardization

Post by Cthaeh »

TL;DR version

I have a bot that could actually make the changes to the wiki. Is it worth doing? What specific things do we want to do with standard page names?


Full post

I've been thinking about this a bit recently, so resurrecting this proposal (actually I was going to start a new one before finding this thread):

One of the issues I might have had with the original proposal is there wasn't a clear way to go about changing all the pages currently on the wiki (particularly if you include changing all the links on individual pages). However, I recently realized that renaming every page on the wiki actually wouldn't actually be impossible to do. A bot should be able to move every relevant content page (and update links on pages) in under two days or so, at least that's the pace of the bot I used in my Preview->MTL page renamings. There would also be prep work (writing code to generate the new names and then double checking that output), but it should all be manageable.

Of course renaming every page would have consequences, the main downside I see being it would either inflate the wiki with redirects, or break any links coming to the wiki from external sites. On first thought, I might compromise between the two by removing redirects for individual chapter pages, and leaving the ones for the project pages.

The question is are there enough upsides to make it worth it? Cloudii was advocating based on a few points, one of which was making API parsing easier. Would it actually help people writing apps at the like (can anyone having written or writing an app comment)? Is there anything else specific that we'd be able to do because of it?

I guess the general consensus in the previous round of discussion seemed to be in favor already.


The one benefit that got me thinking about this is making the recent pages list sortable by language (see this thread). Which would require the "/$lang" format as cloudii was describing, but would work with our current extensions (I think it doesn't even require Translate, it just needs CleanChanges); you could also modify the extension to use a different "$lang" position if you wanted.
User avatar
cloudii
Project Translator
Posts: 512
Joined: Mon Mar 25, 2013 4:30 pm
Favourite Light Novel: Ahouka!
Location: awkward buttface
Contact:

Re: Proposal: Wiki Page Name Standardization

Post by cloudii »

I would say that there are more downsides than upsides, but you really need to get Nandaka and the other existing parsers of BT to comment on this before proceeding with this, because it's very likely we'd have some broken apps from a change like this. ^_____^ Their existing apps have priority over anything new that could be made.

Page name standardization would help with reducing the complexity of navigation bars at the bottom of each page.

It would not substantially help with the API unless every project was truly consistent, meaning the URL has some consistent natural number representation of the ordering of the chapters. That's actually the main trouble of the API. It's possible to get a list of all pages related to a series, but it's impossible to reliably know what order the chapters go in.

For instance, you can fetch via existing API all the mediawiki links on the SAO page:

SAO: Volume 1 Illustrations
SAO: Volume 1 Prologue
SAO: Volume 1 Chapter 1
SAO: Volume 1 Chapter 2
SAO: Volume 1 Interlude
SAO: Volume 1 Chapter 3
SAO: Volume 1 ANOTHER INTERLUDE
SAO: Volume 1 Weird Chapter Name
SAO: Volume 1 Epilogue
SAO: Volume 1 Author's Notes
SAO: Registration Page
SAO: Recent Updates
SAO: Volume 1 Chapitre 1 (Francais)
SAO: Volume 1 Full Text
User:Js06

This information gets returned as a randomly ordered list. Since order is not preserved, the API is basically useless and it is preferable to parse HTML <li> tags to acquire the sequence of consecutive chapters.

But most of the developers have already figured out how to parse the HTML... so is there even value to standardizing something like this? :3
Twitter: @cloudiirain | BT Userpage | OreShura Translator | Biblia Editor (@HereticLNT) | Clockwork Editor (@HereticLNT)
User avatar
Nandaka
Digitalizer Editor
Posts: 535
Joined: Thu Aug 09, 2012 9:46 am
Favourite Light Novel: Ahouka!
Location: Singapore
Contact:

Re: Proposal: Wiki Page Name Standardization

Post by Nandaka »

Hiya, Nandaka here.

Beside the order of the chapter, I need to get the structure of the volumes.

Currently, the application will parse the project html page to get the book/volume structure and chapter
https://github.com/calvinaquino/LNReade ... r.java#L77

roughly, the logic are like these:
1. Find h2 element with project page or some keyword (e.g. *by*, Side_Stor) in it, this mark the beginning of chapter listing.
https://github.com/calvinaquino/LNReade ... .java#L214
https://github.com/calvinaquino/LNReade ... .java#L195
2. Find h3 element after the starting h2 to get the volume marker.
https://github.com/calvinaquino/LNReade ... .java#L235
3. Find li element after the starting h3 to get the chapter list until it find another h3 element which mark the end of the chapter in that volume.
https://github.com/calvinaquino/LNReade ... .java#L497

If you can provide an API that can give the structure, the I can just use that :P

Other scenario is by using custom attribute to mark the beginning of the structure (e.g. bt-data/class?), maybe like this?

Code: Select all

<h3 class='volume'>Volume 1< /h3>
<img class='volume_cover' src='...' /img>
<ul>
  <li class='chapter'>Chapter 1</li>
  <li class='chapter'>Chapter 2</li>
  <li class='chapter'>Chapter 3</li>
</ul>
<h3 class='volume'>Volume 2< /h3>
<img class='volume_cover' src='...' /img>
<ul>
  <li class='chapter'>Chapter 1</li>
  <li class='chapter'>Chapter 2</li>
  <li class='chapter'>Chapter 3</li>
</ul>
Visit my blog for other applications.
User avatar
Shadowys
Project Translator
Posts: 246
Joined: Sun Dec 30, 2012 5:15 am
Favourite Light Novel: Zero No Tsukaima
Location: Somewhere in Malaysia
Contact:

Re: Proposal: Wiki Page Name Standardization

Post by Shadowys »

I realise I'm a little late into the discussion, but anyway this would impact the way my api handles alternative languages in the future, so I guess I'll put in my 2 cents worth.

Frankly I have not found a proper use in the API if the pages are standardized, as it has been so haphazard the way it is now that I have not thought of using the page title to actually do something (I've noticed the page titles have been renamed, but the links in all the pages have not).

My attempt at cleaning up the BT API is using a very forgiving regex parser to match anything that remotely looks like a volume, and has to be a top level li (a attribute provided from parsed wikitext), example here (https://baka-tsuki-api.herokuapp.com/ap ... o_Tsukaima), and is highly dependent on the order in which the links appear (i.e. top down)

https://github.com/Shadowys/btapi/blob/ ... #L382-L385

If the page links are standardised then it would be great, but for titles like "Illustrations", how would that fit into the numbering system?
Winter's the time of the year,
when the cold chill the skin,
from the very within,
but you grasped my hand,
your eyes shedding a frozen tear.
Our eyes met,
and warmth filled the air.
Post Reply

Return to “Proposals and Suggestions”