Wikimedia Conference 2018/Thematic, regional, language-specific meetups/Wikisource Meetup

Participants

Nicolas Vigneron
Asaf Bartov WMF
Tanveer
Peter Meyer
Abbad (Levant)
Xelgen (Aleksey Chalabyan)
Peter Dewald
Jayanth Nath
Balajijagadesh
Ravishankar
Wojciech Pędzich
Satdeep Gill
Pierre-Yves Beaudouin from WM France

Notes

Most number of books downloaded from Wikisource is Polish Wikisource. Polish: crowdfund money, proofread books for school children requirement.

Education systems are an excellent way to share content

Sheer amount of text availability in Polish Wikisource

Asaf: explanations about Benyehuda project ( http://benyehuda.org/ online digital re-transcription, created before Wikisource in hebrew). Low literacy requirement for proofread using simple test editor. Wikisource proofreading is bit complex. Hebrew Wikisource is concentrating on more of religious texts (complementarity with the Benyehuda project).

Armenia: lot of encyclopedia of Armenia content is scanned in Wikisource.

Jayanth Nath: lot of old literature available in indic languages. collaborated with digital library of India. OCR was a improved by google OCR to Wikisource tool by T Shrinivasan.

Balaji: help required from mature Wikisource community to the growing Wikisource usergroup.

Ravi: Wikisource new project. University and government projects try to digitize the literature. limitation of OCR.

Satdeep: Punjabi Wikisource started last year. OCR not working for the Punjabi language. 3 days ago google's OCR started to give 99% accuracy.

Nicolas reports that there are now Wikisource sites in 68 languages. The 68th was just a couple of days -- the new Basque language Wikisource, https://eu.wikisource.org. There are some temporary technical issues (as usual when a project is open).

In Wikisource, there is much focus on cooperation. Every single page require cooperation of three people to make it final. so a lot of cooperation required among the community. most people don't enjoy adding Wikisource. lot more people may be interested to proofread. out community can be 10 times more .. easily, causually... campaings to use Wikisource edit. social wiki campaign.

Asaf: proofreading page is not enough. We need a vision document for Wikisource. We can imagine. We can build it. We can pitch it to the Wikimedia Foundation. This is a compelling vision.

Discovering own heritage.

Audio books in Wikisource. Wojciech: Wikimedia Polska gives micro-grants to prepare professional readings and interpretations of audio books. One effort put together Shakespeare's sonnets recorded by a famous actor. Asaf: LibriVox.org does this, creating audio files, then Internet Archive hosts the audio files. Is there any synergy with them? They do volunteer, amateur work of our kind.

How to Wikisource proofreading in mobile? (lots of technical points to solve).

Please include OPDS link to the Wikisource. OPDS is a standard for catalogs of electronic materials; reading application need them to use Wikisource.

Speech-to-text recognition for Wikisource can be solution for Wiksource proofreading.

Vision

Casual, Mobile first, wikisource editing.

Discussion

166 Wikisource related backlog in phabricator ( https://phabricator.wikimedia.org/tag/wikisource/ ), not much importance in phabricator.

some bugs are continuting for many areas.

strategic case for Wikisource for importance in the phabricator.

squiggle. transcribing one word at a time.

google crowd source app.

how do we do in the short term,

long term

how to give knowledge as service and equity with Wikisource. create this vision and present it to the foundation.

Next steps.

International Wikisource conference in France in 2019, probably focussing on tech. Need volunteers to help make this second Wikisource conference successful.

We discuss developers on Wikisource. User:Ankry from Poland and technical and interested in Wikisource.

A Wikisource conference in India is planned. Lot of Punjabi Wikisource. Bengali Wikisource developing help videos.

Wikisource proofreading

Wikisource stats -- how to update?

Spell checking; it's not available for Punjabi. Discussion of making it available based on Wikitionaries (or maybe based on lexemes in the d:Wikidata:Lexicographical data, a new project for words that will start next month on Wikidata).

Summarize: