RSS
From My Wiki
RSS interoperability session
3 scenarios -
- True mutlilingual site. Feed is simultaneously translated into a number of languages
- Publishing something in their home language - sometimes want to translate to another language (not really a parallel site) - community members help with translation -
- Aggregators - pulling information from other sites and other languages - aggregating, translating, and then feeding them out to different languages
RSS was never designed for translation. What kinds of tricks can be used to make it work well. Every item in an RSS field called "GID" globally unique identifyer. RSS feed has GID with suffix with the language. The system doesn't need to know anything about the source material. And whoever is getting the feeds doesn't have to know what the process of translation was.
Brainstorm on used cases - different kinds of apps that can be used with RSS.
Hooking a machine-based translation system into the process, etc.
Issues with encoding. Planet, for instance does not use Unicode, uses Latin1 Users must also have readers that support unicode. There are tools that allow you to move between character encoding - must imbed information about the encoding.
Good practice - explain to users why they might not be able to read text that's in unicode if they are using a very old browser/OS.
Example - building a Russian site pulling in info from many sites - build in unicode, then down-convert to other encodings.
Not talking about RSS as a consumer service - but as a quick and simple way to move data between systems - more as an internal tool.
How to divide RSS feeds to volunteers? Show what's in the queue, and let the users pick what they will translate. The users will tell you what's interesting.
An idea of using machine translation to translate just the first paragraph, or an abstract, to see if it's worth full translation. It might be possible to provide parallel feeds of original and machine translation.
Wikis and wiki links as a translation tool.
Using RSS - need to build other tools around it - like queue management. The positive thing about RSS is that tools are available - like getting a newsreader.
One possible thing - use RSS to send out phrases to be translated to 100 people who are domain experts. SMS is another possibility. Finding volunteer jobs that large groups of people can do. Marginal cases - technical terms don't exist yet - so some people have to make up words for things. There are also regional differences - like the tradition in Serbia to borrow technical words, and the Croatians make new domestic words. Icelandic - make new translated terms - it is illegal to use English words.
