I do a lot of software localization and I've run into a potential bug with MT. When the first word in the Source string begins with a placeholder (e.g. &Image Effects`SE) then MT shows no matches for any of the words in the string i.e. I get nothing for "image" and nothing for "effects".
Or is this a setting I need to switch? There's nothing obvious.
Hi Michael,
I'm not sure if this is a problem we have logged but a quick test gives some interesting results for me. Just a few tests and their results:
Using your source:
Changing the & symbol to a % symbol:
Changing the & to another letter:
So I think you're right that the $ symbol is causing a problem here and I'll follow this up. But in the meantime maybe you could use a workaround for this and search replace the source using the SDLXLIFF Toolkit using something unlikely to appear in other scenarios, maybe %%% instead of the $. Then search replace again afterwards?
I know this is far from ideal, and you would probably want to update your TM again when you were done so it didn't contain %%% in the source TUs, but if you use a scratch pad TM for updating during the translation itself then it might not be too much work.
But I will make sure this is logged.
Regards
Paul
Paul Filkin | RWS Group
________________________
Design your own training!
You've done the courses and still need to go a little further, or still not clear?
Tell us what you need in our Community Solutions Hub
Paul Filkin | RWS Group
________________________
Design your own training!
You've done the courses and still need to go a little further, or still not clear?
Tell us what you need in our Community Solutions Hub
Paul Filkin | RWS Group
________________________
Design your own training!
You've done the courses and still need to go a little further, or still not clear?
Tell us what you need in our Community Solutions Hub
Just slogging through 21k words and 90% of them start with <p> which means I'm not getting ANY matches for anything. Sure would be have been nice if ...
< and > are "<" and ">" in escaped HTML. Are you working on an XML file? <p> is <p>. I think you should be eliminating these characters at the parsing stage. Wrong file format?
Daniel
< and > are "<" and ">" in escaped HTML. Are you working on an XML file? <p> is <p>. I think you should be eliminating these characters at the parsing stage. Wrong file format?
Daniel
You'd be surprised at the number of big software houses that do their localization by sending out xml or html as text for translation ... yes, I agree it's not a good format to be working in but I can't change what the clients send out, sadly.
if it's a software house, then you get the files in source format, not as (already screwed-up) SDLXLIFF, I suppose...
I.e. you can easily process them as it should be, right?
I get them as plain text. I must return them as plain text. Other than Studio's automatic conversion where I can mostly rely on it re-converting it to what it was, I am loathe to start processing for the fear of messing up 21k words. Also, the client uses <ex></ex> to supply context information and if I somehow mess with the formatting so that all the code is hidden, I lose access to that.
I get them as plain text. I must return them as plain text.
Sure. That does not prevent you in any way to treat them in Studio as needed.
E.g. if they are in fact XML files, but with .txt extension, you can simply change the extension and let Studio treat them as XML files. Or, create a custom XML file type which will be configured to process .txt files...
Also, the client uses <ex></ex> to supply context information and if I somehow mess with the formatting so that all the code is hidden, I lose access to that.
So do NOT mess with the formatting then... ;-)
I don't see ANY problem...
Perhaps you can post a short sample...
Can't post a sample due to non-disclosure but they're not in a format that would allow me to just change the extension. I did consider (and try) that.
Not even a "lorem ipsum" anonymized piece, which would give a picture about the format? We don't care about the content in fact.
If not, then I'm afraid no one can help you really.
Translate
