Enhanced Translation

From Meta, a Wikimedia project coordination wiki

This article in other languages: Es/Traducción_facilitada

Motivation[edit]

It might be a nice feature if articles at wikipedia had the same content no matter which language they were first written in. Information might flow from one language to many other languages.

As things are now, an article in one language may have content that's not included in the same article in other languages. This happens at least in two situations:

  • Two (or more) "sister articles" (articles on exactly the same topic) are created independently in each language. Most likely their structure and content don't match. This is not easy to solve if the two versions are very different. The effort required to merge the two "content and structure" sets may be substantial.
  • There's an original article in a "first" language that is translated once into a "second" language. Later, the "first" article changes, and the "second" becomes outdated. This case is covered below in "A simplistic way to put it". Solving this issue alone would be great!
  • There may be other sources of mismatch.

Maintaining content equivalency in many languages requires a system to compare articles to those of other languages.

It also requires a way to see what changes have been done to the original between the moment when it was translated and now; that way the translator can easily focus on the changes that need to be done to the translator. This is in practice a diff of the two versions and is part of what wikimedia does already.

Enhanced Translation is that system.

A simplistic way to put it[edit]

  • On day 1, I translate a page with 3 lines from English into Spanish. I set the clock to zero, to signal content equivalence.
  • On day 7, I come back and see that the English page has been edited: line 2 has changed and there's a new line between lines 3 and 4. (Often there are many small changes, which makes those changes difficult to track.)
  • I want to compare the two English versions (day 1 and day 7) so that I will know what needs to be done to the Spanish page. When I'm done, I'll reset the clock to zero.
  • Wikimedia can already show a [list of recent changes], and a [comparison of any two versions].
  • We need a way to set a clock to zero (keeping the version number somewhere), so that we can then compare (using the version number as the "oldid" against the current version) the "zero" version with the "now" version (both in English) to make it easy for the Spanish (human) translator.

Development[edit]

Developers have yet to be contacted.

Features[edit]

Enhanced Translation aims to make translation between languages easier. Any feature that adds functionality in this manner should be included.

Edit Tracking[edit]

Although articles cannot be compared word-for-word, topics within the articles can be tracked and updates monitored so translators may update only the portions of the articles changed.

MediaWiki already contains a Revision History feature, which allows the comparison of different versions of an article. This feature could be used to compare changes to one article against changes in another, using Section Tracking as a reference. It would undoubtedly make translator's lives much easier, since modification dates could be supplied. This would facilitate content equivalency.

Language Disambiguation[edit]

On page en:Dog, there is, at the bottom, something such as this:

[[es:Perro]]
[[de:Hund]]

This markup connects pages of another language to the current page.

This can be used to create an internal disambiguation page of sorts that can be used to help the Enhanced Translation extension associate articles of different languages. Once this is accomplished, the extension can then take this information and create an "Edit Tracking" page.

From this "Edit Tracking" page, located in a "Language:" namespace, and linked to from in many different languages, such as "Lenguaje:Perro" or "Sprache:Hund", all these linked articles are collected and displayed.

Further, perhaps functionality could be added to allow one to type en:Perro or es:Dog, and reach an article in the desired language.

Section Tracking[edit]

Headlines of an article in one language should correspond to those of the counterpart in another language. What is needed, however, is a way to associate these sections. It would be easiest to add this information in a table on the "Edit Tracking" page, in this manner:

{|
|-
!en:Animal
!es:Animal
!de:Tier
|-
|[[en:Animal#Dog|Dog]]
|[[es:Animal#Perro|Perro]]
|[[de:Tier#Hund|Hund]]
|-
|[[en:Animal#Cat|Cat]]
|[[es:Animal#Gato|Gato]]
|[[de:Tier#Katze|Katze]]
|}
en:Animal es:Animal de:Tier
Dog Perro Hund
Cat Gato Katze

This could allow the extension to compare sections against each other, not just lines numbers.

Internationalization[edit]

A detailed method of how to prepare MediaWiki for separate, yet interlinked, sites is also necessary.

See also[edit]