• State Suggested Answer
  • Replies 2 replies
  • Answers 1 answer
  • Subscribers 433 subscribers
  • Views 1340 views
  • Users 0 members are here
Options
Related

Issue with Alphanumeric Strings Recognition

Enrico Rosso

Hello everyone,

Alphanumeric Strings Recognition feature is for sure one of the most useful feature around in Trados Studio.

Sadly it recognizes as an Alphanumeric String even the [0-9]-[A-Z]+ regular expression, producing wrong results as you can see below:

Screenshot showing Trados Studio translation memory matches with incorrect alphanumeric string recognition, displaying '3-LENTA' as '2-LENTA' and '4-RALLENTA' as '3-VELOCE'.

The suggestion comes from a correct TU in the TM:

Screenshot of Trados Studio Edit Translation Unit window with a source text 'A43' incorrectly matched to target text '4-RALLENTA' due to faulty alphanumeric string recognition.

I tried unticking the "Count as one if words are joined by dashes" option, with no results.

This also produces wrong fuzzy suggestions, making the translation tricky and the word count unfair.

How do you manage this issue? Has anyone a good solution?

2 Years ago a nice idea was proposed: Being able to modify the regular expressions used for Alphanumeric String Recognition in each TM the same way we do for Segmentation Rules.

https://community.sdl.com/ideas/translation-productivity-ideas/i/trados-studio-ideas/regex-for-alphanumeric-strings-and-variables

Thank you to anyone willing to give suggestions and advices.

Generated Image Alt-Text
[edited by: Trados AI at 6:45 PM (GMT 0) on 28 Feb 2024]
emoji
Rate translation
×
Suggest better translation
×
Moderator UI
×
  • Get AI Suggestion
  • 0

    Enrico Rosso said:

    Sadly it recognizes as an Alphanumeric String even the [0-9]-[A-Z]+ regular expression, producing wrong results as you can see below:

    Screenshot of Trados Studio showing a list of terms with 100% match and CM status, including '2-LENTA', '3-VELOCE', and '4-RALLENTA'.

    I don't really understand this... if this is a filtered set of results why is this wrong as they all match your expression?

    Default settings on a TM should give this which seems to be correct:

    Screenshot of Trados Studio displaying a warning message 'Some segments contain formatting or tag errors' above a list of terms with 100% match and CM status.

    Have you tried reindexing your TM?  It feels as though it's getting confused.

    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

    emoji


    Generated Image Alt-Text
    [edited by: Trados AI at 6:45 PM (GMT 0) on 28 Feb 2024]
  • 0 in reply to Paul

    Hello Paul and thank You for answering.

    I'm afraid I wasn't clear enough in explaining my issue:

    The translation should be like that:

    2-LENTA = 2-SLOW

    3-VELOCE = 3-FAST

    4-RALLENTA = 4-SLOW DOWN

    Instead Studio considers them as placeables and just copy them to the target, accounting those segments as a CM or 100%, instead they're no matches.

    So the TM actually works as expected, but the mechanism generates an issue.

    If the character "-" wasn't considered in an alphanumeric string, it would work like a charm, i would have a number and a word for each segment.

    I can separate and join again the strings before and after translation, but this could add a lot of overhead to the translation process.

    I understand this kind of settings are like a blanket too short to cover everyone, I'm writing this in hope something can be done or planned for a future release, in order to always improve studio the way SDL keeps doing.