Topic on ଉଇକିପିଡ଼ିଆ:ଆଲୋଚନା ସଭା

Aaharoni-WMF (ଆଲୋଚନାଅବଦାନ)

Sorry about not being able to write in Odia. I hope someone can translate this!

Content translation has been successful in supporting the translation process on many Wikipedia communities, and we want to help additional wikis with potential to grow using translation as part of a new initiative.

Content translation facilitates the creation of Wikipedia articles by translating content from other languages. It has been used already to create more than half a million articles. In addition, the tool provides mechanisms to encourage the creation of good quality content, preventing the publication of lightly edited machine translations. In general, our analysis shows that the translations produced are less likely to be deleted than the articles started from scratch.

Odia Wikipedia editors have used Content translation to create more than 1500 articles. Given the size of the editing community, we think that there is potential to use translation to create more articles, expand existing ones, and attract new editors that learn how to make productive edits. Translation can help the community to reduce the language gap with other languages and grow the number of editors in a sustainable way. In order to achieve this goal, we want to collaborate with you to make Content translation more visible in the Odia Wikipedia and support new ways to translate.

As a first step, during the next weeks we plan to enable Content translation by default on the Odia Wikipedia. That will make it easy for users to discover the tool through several entry points. However, users not interested in translation will still be able to disable it from their preferences.

Please feel free to share any comment in this conversation thread.

Thanks!

Psubhashish (ଆଲୋଚନାଅବଦାନ)

Hi @Aaharoni-WMF Thanks for sharing these updates. I have personally used Content Translation and have been also fixing the issues that are created by some of the editors, particularly the newer ones. So, please allow me to share these experiences here as some of them need to be fixed before rolling out largely:

  • CT adds a </nowiki> after many wikilinks and one has to use source editor to manually clean that up.
  • Infobox is generally quite messed up as the new lines get converted into no space create a big chunk of wikicode instead of an organized Infobox (meaning the lines do not appear as lines when edited with CT and they become a large chunk of text) making it really hard to update/translate content inside the Infobox.
  • When CT really helps new editors to create content in a short span (hence it reduces the stress and removes an entry-level barrier). However, almost all new editors so far have not checked the simple MoS that Odia Wikipedia has. In the recent past the community has seen no human effort being made to change the machine-translated text. IMO a notice has to be displayed before the editors start translating suggesting them to a) read the MoS, and b) make manual corrections to fit the Odia sentence structure and grammar before submitting the 'Publish' button. This is not an Odia specific issue but a Machine Translation issue. There has to be some human editing as Odia has a long way to go in the neural machine translation highway.

I personally would love to see more and more editors use both Google Translate and Bing Translator to create content but we have seen extremely low quality content created so far, mostly because of the lack of manual editing. I guess displaying notifications and making changes in the UI could help deliver that request as we assume good faith. In short, I welcome this move in my personal capacity as an editor here.

Aaharoni-WMF (ଆଲୋଚନାଅବଦାନ)

Thanks for the comments! Very useful.

About nowiki: Content Translation (CX) uses Visual Editor (VE) as the rich text editing component, and the </nowiki> thing is an issue that is common to Visual Editor and Content Translation (or, even more precisely, the internal Parsoid component, which both VE and CX use). The same problem also happens right here, in Flow discussions, because it also uses Visual Editor and Parsoid. You should document somewhere in your help pages for editors that when using Visual Editor, Content Translation, or Flow, the editor must note where does the link begin and end. While editing, a link is highlighted in blue. If the writing caret is within the link, then everything that is typed will be within the link, and "nowiki" will not be added. If the caret is not inside the link, but immediately after it, the link will not be highlighted in blue, and then "nowiki" will be added to prevent the inclusion of text after the link ("trail") in the link. The editor can move the caret back and forth using the arrow keys on the keyboard.

About Infoboxes: This is an unfortunate issue with templates in general. They are different in every wiki, so it's hard for Content Translation to adapt them cleanly. It may work well if you have TemplateData in the template here, where parameter names are mapped as aliases to the parameters in the source language (in your case it's most often English).

About MOS and machine-translated text: CX does show a notice with style suggestion and a link to further info when a user begins a translation: "Make the text read naturally. Machine translation is useful but you'll need to revise the text to make it accurate." (It currently appears in English because the user interface is not fully translated to Odia yet. You can —and should!— do it in translatewiki. Here's a direct link to localize Content Translation into Odia.) In addition, as I mentioned in the opening message, CX also shows a warning when it identifies that a user is trying to publish an article without editing the machine-translated text. It can also be adjusted further for your language. Read the page Help:Content translation/Translating/Translation quality for information about this. On the community level, publishing machine-translated text without appropriate editing should be forbidden. As you say, the text must be edited before pushing the Publish button. When you identify a user who publishes machine translations without editing, you should warn them on the talk page, and even threaten them with a block if needed. As for MOS, how do you deal with it for usual editing that doesn't involve Content translation?

We don't support Bing translator at all at the moment. This may happen at some point in the future

Hellohappy (ଆଲୋଚନାଅବଦାନ)

ବ୍ୟବହାରକାରୀ:Aaharoni-WMF : I translated "World Diabetes Day" from English Wikipedia to Odia ବିଶ୍ୱ ମଧୁମେହ ଦିବସ . However non of the references got copied from the source automatically . I had to manually copy each one from the source again . Could you please check what went wrong . Some of my other colleagues also complained about the same happening to them on different occasions .

Aaharoni-WMF (ଆଲୋଚନାଅବଦାନ)
"Improving the translation support for the Odia Wikipedia"ଙ୍କୁ ଉତ୍ତର