Re: Pontoon: Subpar matching in suggestions
Matjaz Horvat <[email protected]>
| Newsgroups | gmane.comp.mozilla.localization |
|---|---|
| Message-ID | <CAEkJQysOcGD4aDYJzZr4vUde+B0pKjKx2iPGRCuN5y2Yt4D7Nw@mail.gmail.com> |
Hi, Indeed, Pontoon's Translation Memory doesn't work at all for longer strings (bug 1469010). While that affects less than 1% of all strings in Pontoon, the longer the string, the more time can be saved using TM. We should fix this. I'm adding it to the Q4 list: https://wiki.mozilla.org/L10n:Pontoon#Other_bugs The problem with complex FTL strings is harder to solve, because FTL format cannot be fully converted to simple file formats (like TMX). Hence, it also involves changes to the data model and UX. See bug 1453341 for details. -Matjaž On Sat, Sep 1, 2018 at 5:45 PM Julen Ruiz Aizpuru <[email protected]> wrote: > > 2018-09-01 17:18 GMT+02:00 Jordi Serratosa <[email protected]>: > > > > In my opinion, the existing string is a great candidate for a machinery > > suggestion but alas! the translator has to retranslate the whole string > > again. This is a major time waster. I've seen the same behaviour in other > > strings too. > > > > I can only agree with the sentiment here; even sometimes 100% matches > are missed if they go above 255 characters [1]. > > Are there any plans to make fixing this a priority? It might be > obvious to say but it'd be a real time saver for translators, plus > it'd help with consistency. I hope something can be done soon to > improve the TM quality overall. > > Thanks for your time put into improving Pontoon for all! > > Regards, > Julen. > > [1] https://bugzilla.mozilla.org/show_bug.cgi?id=1468452 > _______________________________________________ > dev-l10n mailing list > [email protected] > https://lists.mozilla.org/listinfo/dev-l10n _______________________________________________ dev-l10n mailing list [email protected] https://lists.mozilla.org/listinfo/dev-l10n