Translatable texts should be full phrases including punctuation

Translatable texts should be full phrases including punctuation

Issue ID:534
Issue Category:bug
Component:translation
Priority:normal
Status:fixed
Assigned:brion
Milestone:0.6
Keywords:translation

== Principles ==
1. Translatable texts should be full phrases or sentences.
2. Punctuation should always be included in the translatable string.
3. When a variable is necessary, this should be embedded in the full phrase, not assumed to be appended at the end of it.

== Rationale ==
Different languages have different word order. The only way to fully support that is to have a full phrase to translate, not two "halves" of a sentence (embed a variable token in the full sentence); or something that assumes a variable "at the end": it may not come at the end in the target language and trying to force it can lead to incomprehensible text. Not splitting up a sentence automatically implies a variable "token" should be used where appropriate, to allow the translator to use natural word order.

Punctuation should be included (including any full stop of a sentence) because different languages have different rules for punctuation; using the correct punctuation format is part of the translation process.

== Current problems ==
In most cases there is no problem with the English texts as in the translation files, but there are a few problems. Notably:

1. "My text and files are available under" and "except this private data: password, email address, IM address, phone number."
Two problems here: a link to Creative Commons is used "between" these two halves (no natural word order possible) and this embedded link is not followed by a space (but a translated text cannot start with a space!): the result is not only possible wrong word order, but text is not properly separated from the link. A single sentence with a variable for the link would solve both problems here.
2. "Unless otherwise specified, contents of this site are copyright by the contributors and available under the" and ". Contributors should be attributed by full name or nickname."
Note that these are two sentences, and the full stop that starts the second string actually belongs to the first. This could be one translatable string, or two (one for each sentence) but there should be a variable token for the CC link followed by a full stop for the first sentence in any case.
3. "Member since" and "The subscription has been authorized, but no callback URL was passed. Check with the site's instructions for details on how to authorize the subscription. Your subscription token is:" both assume a variable at the end.
They should be included via a variable token to allow appropriate word order.

Another issue caused by "missing variables" relates to the CC link that should be embedded in cases 1. and 2.:[[BR]]Since the CC licenses themselves have been translated in a large number of languages, a localized version should link to the corresponding localized CC version if that is available. This implies that the placeholder for the CC link should contain the fully-qualified URL so the translator can modify this to link to an appropriate version.

Legacy Data

This issue was migrated from another tracking system. The legacy data at time of import is provided below as a reference.

Ticket ID: 
534
Reported by: 
marjoleink
Owner: 
brion
Status: 
closed
Resolution: 
closed
Type: 
bug
Component: 
translation
Priority: 
3
Version: 
undefined
Milestone: 
0.6

Updates

#1

TranslateWiki guys are bugging us about this too. :)

Added to new i18n guidelines page:
http://status.net/wiki/I18n_guidelines#Sentence_splitting

There's a fair number of these to fix up; most of them are labeled as "lego" in the notes here currently:
http://translatewiki.net/w/i.php?title=Special%3ATranslate&task=reviewal...

#2

Most of the above are now fixed in 0.9.x.

A couple notes on ones mentioned about that I'm not changing:

The 'Member since' is split over two elements in what's basically a little table of ("some property", "the property value") pairs, and can't really be split up any more than 'User ID', 'Subscriptions' etc surrounding it.

The subscription token message is likewise broken up with the token in a separate block quote element designed to make it standalone; it can't really be merged into the sentence.

Login or Register to modify this issue, or to receive updates by email.

You can also subscribe to the RSS feed for updates to this issue.