Page MenuHomePhabricator

wikilinks shouldn't be added as such for titles with affixes
Closed, ResolvedPublic

Description

Author: sundarbecse

Description:
Some text in English Wikipedia of the form "<preposition> [[article name]]..." will get translated into "[[translated articlename with suffixal case marker]]". Since we will never have an article name with a case marker, this will forever be a redlink. It'll be annoying to have hundreds of such spurious red links in an article, more so, when there are relevant articles present.

The solution will be to do the following:
"[[translated article name|translatedarticlename iwth suffixal case marker]]"

An example:
English: "of [[bamboo]]" should be translated as "[[மூங்கில்|மூங்கிலுடைய]]" instead of "[[மூங்கிலுடைய]]".
A related issue is at http://sourceforge.net/tracker/?func=detail&atid=603140&aid=3017517&group_id=93107


Version: unspecified
Severity: normal

Details

Reference
bz27494

Event Timeline

bzimport raised the priority of this task from to Medium.Nov 21 2014, 11:24 PM
bzimport set Reference to bz27494.

sundarbecse wrote:

I realise that the same issue will hold for any affixal marker, not necessarily suffixes and case markers.

WBMSRI wrote:

We have fixed this issue in the following way. Assume that the source language (English) page has a word with link say [[SL|SW]] where SW is a word in the source language and SL is the title of the page being linked in the source language.

In WikiBhasha, SW will be translated to a target language word TW using machine translation (MT) and SL is transformed to TL (title of the inter wiki lined target language page) by Wikipedia API.

Here are the possible outcomes

  1. [[SL|SW]]: there is no inter wiki link for SL and MT could not translate SW
  2. [[SL|TW]]: there is no inter wiki link for SL and MT could translate SW to TW
  3. [[TL|SW]]: there exists an inter wiki link TL for SL and MT could not translate SW
  4. [[TL|TW]]: there exists an inter wiki link TL for SL and MT could translate SW to TW

WikiBhasha transforms the above possible outcomes in the following ways.

  1. [[SL|SW]]-> SW
  2. [[SL|TW]]-> TW
  3. [[TL|SW]]-> [[TL]]
  4. [[TL|TW]]-> [[TL|TW]], if TL!=TW and [[TL]] if TL=TW

sundarbecse wrote:

Good to see the fix. Thanks.

What happens in the cases when it is not of the form [[SL|SW]], but [[SL]] accompanied by a phrase to be translated as a whole?
Let's take the following example.
"There are twenty-eight states in [[India]]."

In Tamil, what will this become?
a) [[இந்தியாவில்]] இருபத்தி எட்டு மாநிலங்கள் உள்ளன.
b) [[இந்தியா]]வில் இருபத்தி எட்டு மாநிலங்கள் உள்ளன.
c) [[இந்தியா|இந்தியாவில்]] இருபத்தி எட்டு மாநிலங்கள் உள்ளன.

Of the above, (b) and (c) are acceptable, while (a) is not.

WBMSRI wrote:

We consider [[SL]] as one particular case of [[SL|SW]] where ‘SL=SW’.
For your example, we apply the option (c)
in [[India]] -> [[இந்தியா|இந்தியாவில்]]

sundarbecse wrote:

Thanks for the clarification.