Hi Paola

Number recognition and substitution is still a bit of an enigma to me! I would have thought that Studio would never use recognized numbers to leverage the segment. Indeed, if the number format is right, this is what happens:

In my case, 2:11 is ONE token, while 12,6 are TWO tokens - most likely Studio things one is a time or a date or something like that. Studio does not auto-substitute anything. I wonder why.

I did not try disabling date or time recognition (in TM settings), but you might try.

I tried TermInjector, and that works fine once you found out that the TermInjector rule does not recognize \d, but requires [0-9] instead...

Daniel

Hi Daniel,

I disabled the date/time recognition to avoid localizing the numbers (I had the same issue as you: the colon became a comma). What I wasn't able to achieve was to tell Studio to treat numbers as untranslatable items and ignore them completely.

I have tried to turn off Number Recognition (Project Settings / TM / Settings ) and also to add a variable [0-9]:[0-9] but I don't know these advanced functions well enough so nothing changed.

The quotes are in the following formats, and can have one or two digits, before and/or after the colon:

simple: (John 7:16)  <<< chapter seven, verse sixteen

range: (John 7:16-18) << chapter seven, verses sixteen to eighteen

specific: (John 7:16,28) << chapter seven, verses sixteen and twenty-eight

Paola

Paola said:

I want segment #6 to be translated automatically by Studio, based on my translation of segment #2.

Try turning off number recognition, but leave dates/times.  That format is clearly recognised as a time so that will solve that particular problem.

Paola said:

and also to add a variable [0-9]:[0-9]

Unfortunately that will not work.  Pattern recognition is not available for variables so it will be looking for that specific text.

Dealing with these is trickier:

Paola said:

range: (John 7:16-18) << chapter seven, verses sixteen to eighteen

specific: (John 7:16,28) << chapter seven, verses sixteen and twenty-eight

I'd try using something like Terminjector as suggested already, or use the Regex AutoSuggest Provider to handle the segments as you translate them.  The danger of using partial time and partial number recognition in the settings, with your text, is that part of the target will be correct and you could easily miss something.  So another solution would be this:

1. Turn off number recognition, but leave dates/times.
2. Create QA rules to find all the variants you have and this will find the ones that are not handled correctly so you can manually fix them

Unfortunately it's not "yet" possible to specify custom patterns for TM recognition unless you use the Terminjector, and this can be a little tricky to set up.  But once working it does a good job.

Paola

Paola said:

I want segment #6 to be translated automatically by Studio, based on my translation of segment #2.

Try turning off number recognition, but leave dates/times.  That format is clearly recognised as a time so that will solve that particular problem.

Paola said:

and also to add a variable [0-9]:[0-9]

Unfortunately that will not work.  Pattern recognition is not available for variables so it will be looking for that specific text.

Dealing with these is trickier:

Paola said:

range: (John 7:16-18) << chapter seven, verses sixteen to eighteen

specific: (John 7:16,28) << chapter seven, verses sixteen and twenty-eight

I'd try using something like Terminjector as suggested already, or use the Regex AutoSuggest Provider to handle the segments as you translate them.  The danger of using partial time and partial number recognition in the settings, with your text, is that part of the target will be correct and you could easily miss something.  So another solution would be this:

1. Turn off number recognition, but leave dates/times.
2. Create QA rules to find all the variants you have and this will find the ones that are not handled correctly so you can manually fix them

Unfortunately it's not "yet" possible to specify custom patterns for TM recognition unless you use the Terminjector, and this can be a little tricky to set up.  But once working it does a good job.

Thank you very much for your detailed answer, Paul Filkin!