9

Hi guys,

I noticed that the glossary is displayed in a new way in the workbench. I mostly do jobs for a large auction house which has a relatively large glossary, with the new layout I have to scroll past the glossary suggestions before being able to see any TM suggestions. This means that there is no way I can compare a TM suggestion with the source text without scrolling up and down all the time. I do have to mention that I am working on a laptop, probably this is less of a problem for people working on desktops with large screens at high resolutions. Either way, I think this new layout could cause quite a few errors due to the difficulty of comparing TM suggestions (partial matches) to the source text.

My second problem with the workbench is that it can still get very slow when working on larger collections. Often the letters I am typing appear on the screen with a delay of about a second, which can also cause mistakes when moving between segments of a job. Often the first letter I am typing is added to a previous segment, for some reason the workbench inputs the 'new' letter before moving to the next segment, even though Tab was pressed before typing the next letter. I have tried using both Chrome and Firefox, the problem seems to be even worse in Chrome. I currently have a laptop with 8gb of RAM, I feel that should really be enough for doing text-based work. Any other translators that have the same problem and/or a solution?

Best,

Philip

10 comments

  • 7
    Avatar
    Lidia

    I do a lot of jobs for the same client, and I completely agree. Being able to see the category in the glossary without an extra click is not nearly worth it for me if the trade-off is having to scroll for TM matches. Perhaps a better solution would be letting the translator pick a category for each job with one click in the beginning, and then displaying only the glossary matches for the selected category and the uncategorized ones.

    I have also noticed a few issues that appear to be bugs (I primarily use Chrome on Windows and on an Android mobile phone):

    1. Clicking on a word in the source text no longer inserts the first glossary match for that word in the target text; one has to click on the glossary (and, therefore, keep the glossary section open, which may be problematic on small screens and sometimes results in lagging)

    2. Clicking on the comment icon or the glossary match icon sometimes results in inconsistent behavior (for example, the glossary section opens instead of the comment section)

    3. On a mobile phone, the comment/glossary sections would not close sometimes; I have to reload the dashboard to get rid of them.

    I hope this can be addressed by the development team. Frankly, given my pattern of use, this latest set of updates introduces a number of serious inconveniences and no benefits. I am not sure if I am a typical user or not, but I would prefer it to be rolled back completely.

    Edited by Lidia
  • 0
    Avatar
    philip.lardot

    I was not expecting this post to actually achieve any change, but a reply within three weeks would certainly have been nice...

  • 2
    Avatar
    Lara Fernandez

    Hi @philip.lardot,

    I'm sorry that I was unable to post on here - my apologies!

    However, I did pass your feedback along, and this week we've deployed changes based on it, where you can now scroll through the glossary independently from the source text. Did you notice the change, and has it made it easier to work? If you've had the chance to work on any glossary-enabled job this week, I would love to hear any further feedback you may have.

    Thanks!
    Lara

  • 2
    Avatar
    Lidia

    Hi Lara, thanks for following up. It is wonderful to know that you are always listening!

    Being able to scroll through the glossary is helpful, but I still miss the ability to click/tap on a word in the source window and have the "default" translation filled in. This makes things much faster on a mobile phone when I don't have the glossary pane open.

    Another issue I noticed is that with this new update, each glossary word can be tapped only once. Whenever I need to use the same glossary word more than once, whether because the source text has it this way or because the sentence structure in the target language calls for it, I have to type it.

    I also thought of a new feature that might be helpful: when we click a glossary word and it is inserted, it would be nice if the software added a space after it. This way, if several glossary words occur in a sequence, we could just keep clicking on them without switching back to the translation window to add spaces. I know this wouldn't work in situations when a glossary word is followed by a comma or period, or in languages where there are often grammatical suffixes at the end, but for English, having a space would be more often helpful than not.

    On a separate note, I am not sure about other languages, but the ES->EN-GB glossary is overdue for a clean-up. As it is, about half of the top matches I am shown are not applicable to a typical job, so I have to either scroll or type in the right word. There are issues with inconsistent capitalization (for example, the word "brilliant" is spelled "Brilliant", and I have to change the "B" every single time); plural or singular forms missing for very common words; and first/default choices that are rarely used or simply incorrect, at least for the categories that are currently in use (for example, "alto" which virtually always means "height" in context, is translated as "alto", the musical instrument). I used to report these things, but barely anything happened, and now I can't even find the report button for the glossary. Can we have that feature back please (assuming, of course, that someone is actually going to look at these issues)?

    I know this looks like nitpicking, but I estimate that these extra keystrokes add up to cut my hourly pay rate by a quarter to a half---without any benefit for the customer or to Gengo. Given the pricing for the TM jobs and knowing that some people actually depend on these jobs to make a living, it is important to fine-tune the system to achieve maximum efficiency and minimal waste. So, once again, thank you for looking into this!

  • 1
    Avatar
    Lara Fernandez

    Hi Lidia,

    Thanks for providing further feedback on the issue :)

    Regarding the glossary functionalities, I will pass this feedback along and see what can be done -- as always, I can't promise immediate changes or improvements, but I'll make sure that the team is aware of the translator experience.

    As for the clean up, I will bring this up as well. Just to make sure -- we recently implemented a feature that allows translators to flag poor TM or MT suggestions. If I understand correctly, the glossary works separately from these, and is lacking this feature at the moment? 

    Thanks,

    Lara

  • 1
    Avatar
    Lidia

    Hello Lara,

    Yes, you are right: the feature for flagging poor TM and MT suggestions is there, and it is actually awesome: very quick and easy to use. However, there is no way to report incorrect glossary entries. If I remember correctly, there used to be a Google form, and there was a link to it next to the glossary. It was a bit of a hassle, but at least there was a way.

    Thanks for checking!

  • 0
    Avatar
    Lara Fernandez

    Hi Lidia,

    I've checked with the team on the glossary clean up issue, and it seems like there shouldn't be such a thing as a "polluted" glossary, given that these are provided by the customers, and don't include data entered by other translators (like TM) or by machines (like MT). Based on your language pair, I believe you're speaking about the glossary of a specific customers (let's say C) -- I will ping the account manager for this glossary to see if we can fix things like that B for brilliant (which I can 100% understand how it could be annoying!) For other things like singular and plurals, the team has confirmed that those are not normally provided by the customer in their glossaries - the glossary should be a base reference and should be modified accordingly.

    Will post again when I have updates regarding possible improvements to the display :)

    Thanks,

    Lara

  • 0
    Avatar
    Lidia

    @Lara,

    Thank you for the clarification. I understand that the glossary is not affected by what translators type in, but it still might be worth a second look based on how the client's content has evolved over those years. For instance, we have quite a few glossary entries left over from categories that are not currently used, such as the following:

    • diseño = layout (from "Cars"; something like "design" or "pattern" would be more in line with current usage)
    • años = years old (from "Wine"; in the vast majority of cases, it simply means "years")
    • volante = steering wheel (from "Cars"; the only jobs where I have actually seen this word are about antique watches, and it means "balance wheel". This one probably causes actual quality issues because while "años" is easy enough to figure out, watch parts may not be something many translators are familiar with, and if they uncritically use the glossary match, it will result in an incorrect translation)
    • Romana = steelyard (in the actual jobs, it only ever refers to the Roman Republic or the Roman Empire) 

    In addition, there are phrases with common words and odd syntax, such as "conjunto de = set with" (everyone knows what a "conjunto" is, and don't people usually say "set *of*" anyway?). Some of these can probably just be removed to minimize the scrolling.

    What I mention is by no means the complete list, just a few examples that popped into my head based on the jobs I saw in the last couple of days. Hope this helps provide a more nuanced picture of the translator experience.

  • 1
    Avatar
    Lara Fernandez

    Hi Lidia!

    Thanks so much for the elaborate reply -- I will pass this feedback to the person in charge of this customer's glossary, to see if they can bring it up with them :)

    Have a great week!
    Lara

  • 1
    Avatar
    Lynn Harding

    My issue with the changes to the way the glossary works is that it automatically opens even when there aren't any glossary entries, so I'm constantly having to close it to keep it from needlessly taking up space. Can you make it so that it only opens by default if there's actually something in it? 

Please sign in to leave a comment.