[Wikisource-l] old toolbar is back

[ankry.wiki]

Just FYI.
 
We have restored in plwikisource the old, fast, light and efficient toolbar as 
a gadget:
  https://imgur.com/a/DFVhaeu
 
It fundamentally has the same functionality as the just removed build-in 
version (only the CodMirror icon is missing; but nobody in pl.ws seems to use 
this tool).
 
Regards
   Ankry___
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikisource-l

Re: [Wikisource-l] wscontest tool

[ankry.wiki]

Maybe a bit OT, but did anybody consider to analyze how many corrections / how 
much content a user changes his/her edits?
 
As we (in pl.ws) focus on text quality, it would be appreciated to score higher 
the edits with corrections than edits with no correction ("no errors found").
I understand that it is hard to create such tool, especially if automatic 
corrections (eg.  ->  or template parameter reorganization should be 
ignored), but maybe, somebody has an idea how to to do this?
 
Ankry
 
PS. We do not participate in this contest in its current form because some of 
our users are afraid that this will lower the text quality (users might prefer 
just to update the status without careful checking the content, just declaring 
that "no errors found"; and it is hard to identify such users/edits).
 
W dniu 2018-06-22 09:29:45 użytkownik Rachmat W.  
napisał:
Thank you, Sam!
 
We would like to utilize this tool in the Wikisource contest later. :)
 
Kind regards,
Rachmat
On Jun 22, 2018, 07:08 +0700, Sam Wilson , wrote:
Hi all,
After chatting with User:CristianCantoro at the Barcelona hackathon, I
had a crack at a web frontend for the wscontest tool. It is now online
at https://tools.wmflabs.org/wscontest/
It's not quite done, there's a bunch of things to be fixed, but is
functioning to some extent. I've started a list of its issues at
https://phabricator.wikimedia.org/tag/tool-wscontest/
--sam
___
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikisource-l
 
 ___
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikisource-l

Re: [Wikisource-l] Pages where template include size is exceeded

[ankry.wiki]

Hi,
 
Unfortunately, MediaWiki has some strict limits which are easily hit if 
creating large pages or ineffective code.
You can see parser profiling data (and limits currently set) in the bottom of a 
preview page.
 
We encounter this problem in plwikisource on some really large pages, like:
  https://pl.wikisource.org/wiki/Faraon/ca%C5%82o%C5%9B%C4%87
where there is a lot of transcluded text; and out solution here was to use 
substitution (see the code on the talk page of this page) and handle updates of 
substituted pages by bot.
 
But this is not your case.
 
IMO, your problem is that templates you use on the transcluded pages multiple 
times should be as simple as possible, avoiding calling other templates 
(especially if they also call templates, calling templates, ...). As you can 
see previewing this page in wikicode mode:
 
https://ta.wikisource.org/w/index.php?title=பக்கம்:எனது_நாடகவாழ்க்கை.pdf/19=edit
 
it generates over 110kB of transcluded code. A lot.
And I can see two possible solutions here:
- simplify the TOC templates (actually, it is more effective to use many simple 
dedicated templates than a single combo), or
- use a LUA module to process all necessary parameters and generate expected 
wikicode avoiding transferring parameters multiple times.
 
The latter solution is used in fr.ws & pl.ws; see
 
https://pl.wikisource.org/wiki/Szablon:SpisPozycja and
https://pl.wikisource.org/wiki/Modu%C5%82:Spis
(based on:
https://fr.wikisource.org/wiki/Module:Table )
 
In higly-used templates even removal of a not necessary space, semicolon, as 
well as shortening HTML ids, style names, template names or parameters names 
gives sometimes a significant effect.
 
Ankry
 
W dniu 2018-04-07 08:04:20 użytkownik balaji  napisał:
Hi,
     I have created a transclusion page in ta.wikisource where some pages are 
not displayed. A localised category which in English means "Pages where 
template include size is exceeded"
 
https://translatewiki.net/wiki/MediaWiki:Post-expand-template-inclusion-category
 
https://translatewiki.net/wiki/MediaWiki:Post-expand-template-inclusion-category/ta
 
The page is  https://ta.wikisource.org/s/8wuc
 
What to be done for all the pages to be displayed.
 
Regards,
J. Balaji
(User:Balajijagadesh)
 ___
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikisource-l

Re: [Wikisource-l] Frankenstein's 200th birthday

[ankry.wiki]

The oldest Polish translation was published in 1958.
As the translator death date is unknown (he may be still alive), we cannot 
publish it in plwikisource before 2099.
 
Ankry
 
W dniu 2018-04-05 01:09:41 użytkownik billinghurst  
napisał:
A thought bubble that I shared on enWS was whether the broader Wikisource 
community had an interest in getting English language and available 
translations online in 2018.  I thought it would be a bit of a wikimedia 
talking point if we could do soemthing for such an iconic work. 
 
For non-English languages are there scans available of public domain 
translations? I know that enWS and frWS have 1831 editions,  though not 
scan-supported. 
 
Regards, Billinghurst 
 ___
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikisource-l

Re: [Wikisource-l] IA Upload fixes

[ankry.wiki]

As published in 1914 it is PD in US.
Its copyright status in Gernany (published in Dresden) is probably unclear, so 
it should not be uploaded to Commons.
 
It can be uploaded locally to oldwikisource and, likely, to some other wikis 
that recognize US copyright status only.
 
Ankry
 
W dniu 2017-12-03 21:40:20 użytkownik mathieu stumpf guntz 
 napisał:
No. I don't have a book I would like to try it in mind, and I guess it would be 
pointless to import an random book on which I don't plane to work, wouldn't it?
Right I would have been interested to find works of Charlotte Pulvers, 
especially Elzasaj Legendoj, but actually I'm not even completely sure the book 
is in public domain as I didn't found her vital records.
Cheers
Le 03/12/2017 à 01:47, Sam Wilson a écrit :
I agree there might be some work to do at some future point when we're trying 
to import mixed-media works into Wikisource, but I'm afraid for now the sole 
purpose of ia-upload is to convert IA scans of books into DjVus. MP3s or any 
other non-text work cannot be converted with this tool.
 
Have you tried using it? Did you notice any problems?
 
— Sam.
 
 
On Sun, 3 Dec 2017, at 06:03 AM, mathieu stumpf guntz wrote:
Le 29/11/2017 à 08:47, Sam Wilson a écrit :
* Attempts to use the tool on items that don't have DjVu, PDF, or JP2 Zip files 
will now fail. I'm not sure what people who were attempting to make DjVus of 
MP3s were thinking, but they'll get a nicer error message now.
Apart from the file format, which was under some patent issues until last year 
and should be soon be available to Commons (if not yet), what would be the 
problem with uploading the file to Commons? All the more, some books might have 
sound component. My daughter have a whole library of that kind. She also have 
plenty of book with tactile data which often make me wonder how we would 
integrate such a book in Wikisource, as I'm not aware of any technology to 
render that kind of data linked to the material texture. Defining a file format 
to store that kind of data would be surely the easy part, all the more when 
"digital texture" is a topic already extensively covered in the literature for 
3D rendering purpose.
 
Just sharing some thoughts…
 
I dare say I'm broken something along the way, so do keep an eye out for 
weirdnesses and let me know. Thanks, sam. 
___ Wikisource-l mailing list 
Wikisource-l@lists.wikimedia.org 
https://lists.wikimedia.org/mailman/listinfo/wikisource-l
 
 ___
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikisource-l

Re: [Wikisource-l] Validated works on kowikisource?

[ankry.wiki]

The index status information IS available in MediaWiki, see:
 
https://en.wikisource.org/w/index.php?title=Special:IndexPages
 
 
Unfortunately, it is hard to use it as
* it seems to be available only as this HTML special page (no API interface, no 
LUA interface)
* if is available only for 1 index pages and it is unclear how data is 
sorted (i.e. what is the characteristic of pages that exceed the 1 limit); 
however it hurts only the two largest wikisources at the moment (fr with over 
15600 index pages and en with over 11500)
* the index status is often not up-to-date; index pages need to be purged then.
 
In pl.ws we use this data to update index status by bot, eg. on this page:
  https://pl.wikisource.org/wiki/WS:Wikiprojekt_Proofread
 
IMO, if this information is available via API or directly in wikicode, it will 
be more convenient to use, and maybe useful also for automatic Wikidata status 
update (assuming there is a link to index page there).
 
Ankry
 
W dniu 2017-10-12 10:31:29 użytkownik Gerard Meijssen 
 napisał:
Hoi,
Yes we need to have a better mechanism. However, there is nothing stopping us 
to publish all the books that are finished and ready to be read. That is the 
proof of the Wikisource pudding..
 
So is there a Wikidata query showing all the books that are finished and ready 
to read. With this we can advertise, having all books makes it complete. 
Knowing all books that are finished is secondary.
Thanks,
      GerardM
On 12 October 2017 at 09:23, Nicolas VIGNERON  
wrote:
Hi,
 
The validation-status data can and sometimes is stored in Wikidata. See how it 
is stored here : https://www.wikidata.org/wiki/Help:Badges and here an example 
of Wikidata query for all items about validated text : 
http://tinyurl.com/yaud3uoy
Thad said, most of the times Wikidata is not up-to-date, and there should be a 
tool to take care of that instead of human being.
Cdlt, ~nicolas
___
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikisource-l
 ___
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikisource-l

[Wikisource-l] content modef of [[MediaWiki:Proofreadpage index data config]]

[ankry.wiki]

Dear wikisource administrators!

I noticed that default content model (wikitext) of index configuration page:
  [[MediaWiki:Proofreadpage index data config]]
is wrong in almost all wikisources. As it contains index definition in JSON 
format, the content model should be JSON to avoid potential presentation 
problems (eg. in case when somebody wishes to add wikicode examples or HTML 
tags here).

Yesterday, I disputed this with Tpt on IRC, and he agreed with me that the 
content model "wikitext" is wrong here and told that it is difficult to change 
it from the ProofreadPage extension. It should be done manually, when 
ProofredPage is configured.

The change only affects HTML presentation of this particular page and is 
written in database. Page code accessible by API or visible for the 
ProofreadPage extension remains unchanged. So the change is safe and also fully 
reversible (if somebody wishes so).

The content model has already been changed in en (by Yann), fr (by Tpt), pl & 
mul (by me), hu (by Tacsipacsi). Nothing needs to be changed on wikisources 
that do not use this page (eg. ar, de, ko, sv).

INSTRUCTION (English interface assumed)
- open [[MediaWiki:Proofreadpage index data config]] page
- choose "Page information" from the left menu
- find the "Page content model" row in the "Basic information" table
- if "wikitext" is displayed as current model, click "change"
- choose "JSON" as "New content model"
- click the "Change" button to save the change.

You need to be admin to change content model.

The change may be also performed by a global admin or steward. But they need to 
be ensured that no bot/tool uses HTML code of this page directly.
(A bot should use API if really needs this page, but if you know any bot 
accessing HTML code ot this page, let us know)

The alphabetic list of wikisources that still need this change:
be, bn, ca, da, es, gu, hr, hy, it, is, mk, nl, no, or, pa, pt, ru, ta, uk, 
vec, zh

Ankry

___
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikisource-l

[Wikisource-l] new right for Wikisource admins in Mediawiki/ProofreadPage

[ankry.wiki]

Unsure if you noticed that admins are now able to override Page status change 
limitations in ProofreadPage.
 
There is a discussion in phabricator:
  https://phabricator.wikimedia.org/T167491
about this feature if you care.
 
Ankry___
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikisource-l

Re: [Wikisource-l] Wikimedia Strategy

[ankry.wiki]

W dniu 2017-04-11 14:06:02 użytkownik Nicolas VIGNERON 
<vigneron.nico...@gmail.com> napisał:

> 2017-04-11 13:17 GMT+02:00 David Starner <prosfil...@gmail.com>:
> >
> > On Tue, Apr 11, 2017 at 2:46 AM ankry.wiki <ankry.w...@onet.pl> wrote:
> >>
> >> I doubt we can find 1000 works with PD translations into each Wikisource
> >> language, including Latin and Sanskrit.
> >> It would be hard to find 10. Mostly ancient.
> >>
> >> Unlike Wikipedia, we present content that has already been created by 
> >> somebody.
> >> We are not creating that ourselves.
> >> (except few ws accepting Wikisource translations)
> >
> > How many Wikisources don't accept user translations? I'd guess that at least
> > half of them do.
> 
> Good question. We should store clearly this information somewhere (on 
> https://www.wikidata.org/wiki/Q19335648 and local pages ?).

We do:
 https://wikisource.org/wiki/Wikisource:Subdomain_coordination
At least 4 do not allow translations.

> > It may not be universal, but you'll never know how many of those works
> actually have PD translations until you actually search for them. A list can
> at least provoke the search.
> 
> Exactly.
> I can easily find to 10 works in most languages of the planet (The Bible, the
> Universal Declaration of Human Rights, Shakespeare, Conan Doyle, Dickens, 
> Stevenson,
> Verne, some important international treaty and publication from the Vatican ;
> it's already a lot more than 10 works available in more than 100 languages)

most != all   (Most Wikisource should have... != All Wikisource should have...)

> Speaking of the UN, the UNESCO created the Index Translationum
> ( http://www.unesco.org/xtrans/bsstatlist.aspx ) that can be helpful here.
> Cdlt, ~nicolas

> PS: Latin or Sanskrit are not the thoughest challenges, try Breton or Venetian
> :P (by the way, the UDHR exist in these 4 languages and 500 more ;) only the
> Bible has more translations).

I have intentionally chosen dead languages to point out that "all" should not
be the goal.

Concerning, UDHR, we have unclear copyright status even for Polish translation:
it is not considered to be an official legal act, no "official" translation; 
translated by a Foundation which say nothing about copyright. And even, 
translations of foreign legal acts are considered copyrighted in Poland 
(according to opinions we have).

Translation copyright problems may exist for many translations of Conan Doyle,
Dickens, Stevenson or Verne.
I also doubt we will get a Wikisource translation of "The Posthumous Papers of 
the
Pickwick Club" into eg. Lithuanian (while ltwikisource seems to be like
a single-user project - at least recently).

We can talk about 1000-100 "base" works in, maybe, 5-10 most active Wikisources.

Ankry

___
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikisource-l

Re: [Wikisource-l] Wikimedia Strategy

[ankry.wiki]

I doubt we can find 1000 works with PD translations into each Wikisource 
language, including Latin and Sanskrit.
It would be hard to find 10. Mostly ancient.

Unlike Wikipedia, we present content that has already been created by somebody. 
We are not creating that ourselves.
(except few ws accepting Wikisource translations)

Ankry

W dniu 2017-04-11 09:42:54 użytkownik mathieu stumpf guntz 
 napisał:
> Hi Nemo,
> 
> We may establish a list a the "1000 works that every Wikisource should 
> have" (with translation possibly needed).
> 
> What metric could we use to define such a list? Maybe reference 
> frequency, but it requires statistics whose availability is unknown to me.
> 
> Statistically,
> psychoslave
> 
> Le 29/03/2017 à 08:30, Federico Leva (Nemo) a écrit :
> > One issue sometimes raised about Wikisource is how we know that we're 
> > working on the "right" books. Internet Archive is planning to 
> > textbooks starting from those which are most frequently assigned in 
> > USA schools:
> > http://blog.archive.org/2017/03/29/books-donated-for-macarthur-foundation-100change-challenge-from-bookmooch-users/
> >  
> >
> >
> > I was surprised to learn a project like OpenSyllabus exists and works, 
> > I emailed them to ask what it would take to do the same for other 
> > languages/geographies.
> >
> > Nemo
> >
> > ___
> > Wikisource-l mailing list
> > Wikisource-l@lists.wikimedia.org
> > https://lists.wikimedia.org/mailman/listinfo/wikisource-l
> 
> 
> ___
> Wikisource-l mailing list
> Wikisource-l@lists.wikimedia.org
> https://lists.wikimedia.org/mailman/listinfo/wikisource-l
> 




___
Wikisource-l mailing list
Wikisource-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikisource-l