AI is everywhere, and I think it should be first and foremost be used for bug tracking and solving

The essential is said in the title.

Some bugs and cryptic error messages exist since the inception of Trados Studio 2009, that is 15 years ago.

I do not know if tracking those bugs and errors is too much time consuming for RWS teams and thus not profitable enough, or if the software is that deep that it is impossible to dig in it to find the inconsistencies that cause those errors, but the fact is that some 10+ years old bug still exist. This morning again, I had the boxes below on my screen when I sat in front of my screen! Yet, nobody used my computer last night.


Multiple error dialog boxes overlaid with the message 'Une exception de type 'Sdl.Desktop.Platform.StudioOfflineException' a ete levee.' in Trados Studio.

The question I ask myself for months is: since AI is everywhere and purportedly so powerful, since vendors, includng RWS, sell us their stuff arguing it is AI-enhanced, why does not RWS implement some AI functionality to monitor, track and point the root causes of those errors, in order for RWS teams to fix them? We could them hopefully soon have a fully operational Studio, without bugs.






Generated Image Alt-Text
[edited by: RWS Community AI at 2:03 PM (GMT 1) on 25 Sep 2024]
emoji
Parents
  •  

    Just because the title of an error message looks familiar doesn’t necessarily mean it’s caused by the same issue as one from 10 years ago. We are continuously working on fixing bugs, often even before they’re reported, especially if you've agreed to share telemetry data with us. This data helps us capture error stacks, including those that don’t necessarily interrupt your work, and allows us to focus on the areas where we can make the most impact.

    It’s clear that experiences vary—some users rarely encounter issues, while others experience them more frequently. Naturally, forums like this tend to attract users who are seeking help, reporting bugs, or expressing frustration, which sometimes skews the perception. Even so, we take this feedback seriously and use it to help guide our development efforts. For instance, we employ topic modelling to identify recurring themes from forum discussions, and we've utilised OpenAI’s GPT-4 to help analyse and interpret the results. This provides valuable insights, though it can be challenging, as users may not always know the best way to frame their questions or what information to provide that is helpful for troubleshooting.

    We understand that the range of products we offer and the ongoing development efforts behind them adds complexity, but we are fully committed to addressing these challenges and ensuring the software works as expected. While you shouldn’t have to worry about the inner workings, we want you to know that these efforts are constant.

    We truly appreciate feedback like yours. It’s encouraging to see constructive thoughts on how we might address challenges, especially with the use of AI. Thank you for your engagement—it's great to see we are aligned in our approach to tackling these issues.

    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
  • Very informative answer,  , I've learned some stuff I didn't know.

    And,  wow, wouldn't that be great, if AI could inform the user accurately what has caused the same old error message, whether it's Trados or it's something on the user's setup that is triggering the issue. It could lead to longer-term solutions and better understanding by the user.

    Bon courage !

    Ali Smiley

    emoji
Reply Children
No Data