marvellous XLIFF - but how can I explain why?
The more I look at XLIFF (and XML) the more marvelled I am about how straight forward it is. Still - it is a bit of a struggle to develop a way to present it to (often doubtful) “third parties”. To fresh are the memories of giving a presentation about Deja Vu at the NZSTI annual conference where I had not been 100% prepared and I had to change one of DV’s settings. The comment I heard afterwards from a colleague that I very much respect was “well - if Christof had to spend so much time on getting the settings right - how would I cope”.
The thing is, that any presentation works best if it runs smoothly and if it is well structured. That is particularly true for a concept that “looks” as scary as an xliff file might - even though it doesn’t compare with what a word doc looks in wordpad).
It seems to be a good approach to take the format apart and explain its parts - but maybe I can start not at the beginning with the XML declaration and all those formal XML related stuff. Instead I could start with an explaination of the translation unit tag. I’ve found a funny little example for an xliff file - where instead of working with two languages, they instead explain what would be written there (like “this is not English, but actually French” (this reminded me of a dubbed movie (it is a little bit like the movie clip that Tom pointed out on (http://www.lingo24.com/language_translation_forum/viewtopic.php?t=56).
So how do I think this could be useful to explain XLIFF? OK - I should say that there are tags (between those “<” and “>”) to label things - and there is the stuff that is labled by those tags - which in XLIFF is usually the source and the target text.
The cool thing is that the labels can be customised to your needs - so in our example below, they are customised to label content as source text, target text and alternative target texts (like fuzzy matches)
So if you are still reading this blog, please feel free to give me some feedback if this makes sense to you. And maybe have a look at this example and tell me if you think you can work it out? (you’ll note, that the labels are actually quite eloquent)
Thanks a lot!!
<trans-unit id=”n1″>
<source>[color=darkred:0f80bc1992]This is a sentence.[/color:0f80bc1992]</source>
<target xml:lang=”fr”>[color=darkred:0f80bc1992]Translation of “This is a sentence.”[/color:0f80bc1992]</target>
<alt-trans match-quality=”100%” tool=”TM_System”>
<source>[color=darkred:0f80bc1992]This is a sentence.[/color:0f80bc1992]</source>
<target xml:lang=”fr”>[color=darkred:0f80bc1992]TM match for “This is a sentence.”[/color:0f80bc1992]</target>
</alt-trans>
<alt-trans match-quality=”70%” tool=”TM_System”>
<source>[color=darkred:0f80bc1992]This is a short sentence.[/color:0f80bc1992]</source>
<target xml:lang=”fr”>[color=darkred:0f80bc1992]Fuzzy TM match for “This is a sentence.”[/color:0f80bc1992]</target>
</alt-trans>
</trans-unit>