3

Hi all!

Happy New Year! I hope you all had a wonderful New Year's Eve and beginning of 2017.

I am posting today because I would like to ask for everybody's feedback on a dictionary prototype that we have created :)

You will find the dictionary here: https://gengo.com/t/bocabo

This dictionary uses past Gengo translations, and you can choose to search your own translations or all Gengo translations.You can also narrow down the search for a particular customer.

We would like to know whether this is something that you guys would find useful, and we would appreciate suggestions for improvement as well.

Thanks,

Lara

18 comments

  • 3
    Avatar
    Alexander

    Best wishes from me to everyone too!

    This online search tool is a great idea! In my first short test, it is fast enough to be useful. However, only the translation is shown, not the original context, which I think is a limitation.

    I would not exactly call it a dictionary. For general purposes, nothing beats a real dictionary that was crafted as such. I think the main benefit of this tool is that it enables different translators to make consistent choices when working for the same customer.

    I suggest to replace the field 'Only customer with id:' with a checkbox 'Only the customer id of the job I am currently working on'. I don't think I will ever want to use a different customer id, and such an option would save the time of manually finding the id and pasting it into the field.

    I think there is a bug when filtering both by customer id and My translations only. When I search for 'test' I get several results. When I narrow the search down to customer -123456 only, I get zero results. Makes sense, since I assume there is no customer with negative id. But when I narrow it down even further by also checking My translations only, I get several results again.

    Edited by Alexander
  • 4
    Avatar
    fc.spa

    Hi Lara. This could come in handy, but I would like to make a few remarks.

    After trying a few searches, I have noticed that the results are full of mistakes and errors. I don't know if it is possible or you have already considered the idea, but it might be better to omit at least the parts that have been reported by senior translators and include only the work of colleagues with a rating above a certain score. It is very important to have a certain degree of reliability: If I'm looking up something uncommon, and the few results I get (if any) can't be trusted, this dictionary could end up being misleading more than helpful.
    Moreover, having many incorrect translations leads to error propagation. I agree with Alexander about the possibility of consistency when working for the same customer; on the other hand, this means relying on other translators' skills, and considering the poor work I've come across, it's something that could be counterproductive and lead to a chain reaction of future errors.
    Being able to read sources would also be nice.

    Also, are you sure you want to call this a "dictionary"? Considering the above, it's something too fallacious to deserve this name. Aside from lexical nuances, I'm sure people will complain about receiving bad feedbacks on errors made using this tool: "Hey, I found this sentence using Gengo's dictionary, it's your fault!"
    Maybe something like "database" (if not "translation memory") would be more appropriate and save you some hassle.

    Just tossing ideas around, improved and taken cum grano salis it could be a useful tool.

    Thank you, and happy new year!

    Fabio

  • 7
    Avatar
    LBTranslations

    I have not looked at the dictionary but I have been working with TM on one of Gengo's major clients for a while now. Please note that I worked on this client for years prior to the introduction of TM.

    I don't even have to look at this dictionary to tell you that I will never use it. Already, as we predicted, the TM is chock full of errors. The TM creates work for translators instead of making the work easier. And of course, while being paid less than before. I am doing more work than before for less money as I have to continually correct the TM that never seems to register the correct translations, only the incorrect ones. And this dictionary is taken from the TM? That is utterly laughable.

    Please, please, please fix the existing problems and issues with the dashboard before starting something new. We have been telling you this for years! Please start listening to your translators before all of your experienced ones leave out of frustration at not being heard.

  • 4
    Avatar
    KevanSF

    Hi Lara, hi all!

    I'd like to add my best wishes to all for a happy, productive, 2017!

     

    I'm glad to see Gengo working to innovate and improve!

     

    I have to agree with LBTranslations above, in that the TM is often more hassle than it's worth when doing jobs for the auction client. Even when the translation is already filled in (meaning we get paid next to nothing), I really have to go over it with a fine-tooth comb. There are often silly typos in the text (misspelled words), and tiny errors such as black instead of white, cm instead of mm, 2.8 instead of 3.8, etc. Additionally, I often find awkward English phrasing or inconsistent, sloppy punctuation. Occasionally, there are flat-out wrong translations.  All this appears in the segments that are already filled out, just to be clear.

     

    In short, I have to put just as much time, care, and attention into these low- or no-pay segments as I do in coming up with a fresh translation. (Again, to be clear, I still enjoy working on the jobs, I'm learning a lot, and I absolutely appreciate both Gengo and that large-volume client for all the opportunities to work at something I enjoy. I'm simply voicing what I feel are valid observations in the hopes that it may help guide the PTB in Gengo to continue to innovate and improve their interface for the benefit of all!)

     

    Anyway, with that in mind, I'm hoping you can understand any resistance or negativity you might encounter from the translators regarding this new feature, which on the surface, seems like a souped-up version of TM (or a giant TM compilation).

     

    That said, I plan to maintain an open mind, and I'm confident that TM can be improved, and that this new database function could evolve into a useful utility, although I do worry that errors could simply be propagated, as mentioned above.

     

    Will there be a link from the dashboard?

     

    I agree with previous comments:

    —It definitely should have a more accurate name than "Dictionary"

    —There definitely needs to be a function for users to remove (or at least flag) faulty or low-quality translations. (Perhaps STs or others could periodically review the database and cull poor translation attempts.

     

    Best wishes,

    Kevan

     

     

     

  • 1
    Avatar
    chrisdechert

    I have mixed feelings about this database. On the plus side, I think it could be useful to jog my memory on a term that has just escaped my mind. I like it as a means of seeing how other translators are translating a difficult and ambiguous term in order to see if I can be doing a more accurate job.

    On the negative side, I am a bit resentful that all my hard and meticulous work at figuring out a very difficult phrase or word is just made available to anyone who wants to type the word in a blank. As well, I have to reiterate the problem of propagation of errors. Especially in specialized fields, for example the very specialized philately and numismatic texts that are currently being translated for an auction site, if one translator enters a reasonable but frankly wrong translation for a term, the likelihood is that this will continue to be used. I am certain of this, because there are already many examples with the already existing TM.

    As for improvements, I would recommend that the search term be in bold font in the results, to avoid having to search the entire translation for it. As well, the original text is really important to make the search results useful - for instance, as is done on the Linguee website. Consider the Spanish/French term talla/taille, which can mean size, waist, fit, height or cut. Without knowing how it is used in the original sentence, the possible translations are quite meaningless.

    Just as an incidental note, I noticed in the translations provided, the text that we fill in as individual boxes is written continuously. If this is the way that the items are returned to the customer, we should be making sure to put some punctuation at the end of each box in order to improve readability and professional appearance.

  • 0
    Avatar
    Paulo C.

    I speak for myself. Every single month I have a job evaluated by the Senior Translator. So, I don't have any chances for errors. However, I see deep errors every single day. 

    This tool should be only to search the my past jobs not for all. Or, if available for all, it should include all pairs, not only one. I work on other pairs and it would be great to see all pairs.

    Moreover, it would be great to have graphics about quality performances (also for glossaries) for every pairs. Based on that, the jobs should be prioritized to those who have best performances. The actual feed simply doesn't works for quality improvements.

    Have a great year! Rooster rocks!

     

    Edited by Paulo C.
  • 0
    Avatar
    bgazzal

    I'd need more time to play around with this but like the idea a lot, especially the "search my translations" function, which could be very helpful - I build up my own glossaries for certain technical jobs and this could be another way to access material, and the Gengo TM really should have a search function.

    However we would need access to both source and translation (+ some context, ie a couple of sentences) for this to be truly helpful - unless I missed something we can see the target results at the moment, and have no context. As others have mentioned, a Linguee like hit display would be nice - otherwise the Memsource search TM function was really good as well.

    cheers

     

    ben

  • 0
    Avatar
    Sebaey Mahgoub

    Thank you for your efforts.

    This is a very good idea.

    I have been thinking about something like this during the last few months.

    I have made something like this from my own translations.

     

  • 0
    Avatar
    Sergio

    Hi Lara and all who shared their feelings,

    firstly, I wish to reciprocate the wishes for a happy and productive new year.

    I basically agree about all that has been said on this new tool till now, and I must admit that it is exactly what I was thinking to need.

    In my opinion, the tool has a great potential in the search of one's previous translations and when you're trying to recollect that word, an idiom, those particular phrases you have already used previously, both to complete the job faster and to assure consistency to a specific customer.

    Personally, I prefer to work on any job by considering and interpreting it as it is, and, above all, I always try to reach my own translation with the minimum number of external influences (except those of dictionaries, obviously). So, I think that I would not be incline to consulting the other translators' versions, even though I should say it could be useful anyway (and in some critical circumstances) to have the possibility to access to them (but only to the correct ones, of course).

    I also want to add that I think we should be grateful to Gengo for developing this great opportunity for us to improve our translations... and ourselves!

    A last consideration for what regards the name of the tool: I also agree about the fact that it can not be really considered a "Dictionary", but a sort of "Memory" to which any translator can freely access to provide fluent, consistent, reliable and, in the final analysis, good jobs, while enhancing themselves.

    Bye!

    Sergio       

  • 0
    Avatar
    mnyxta

    Finally!

    I've really felt the need for a feature like this for quite some time. Since TMs are not used in my language pair, whenever I would see a familiar job, I wished I had kept better records of previous translations. Trying to find them in the "ARCHIVED JOBS" section can really be a hassle.

    The only thing that's missing from the new sservice is being able to view the source text. That's definitely a must.

    Thanks for the addition! 

  • 0
    Avatar
    kvstegemann

    Hi all!

    In addition to what was said before, I would suggest to call this thing "concordance search". It seems it does something similar as a concordance search in a standard CAT tool.

    To make such a thing really useful, it should be integrated in the translation workbench in a way that it can be used without touching the mouse. Translation productivity is always killed when you have to move your fingers away from the keyboard for whatever reason. The ideal workflow would be: while working in a segment, press a shortkey like Ctrl-K, the cursor jumps into the concordance search field, enter a search term, hit enter, see the list of matches, hit another shortkey (can be Ctrl-K) again and the cursor jumps back to the same position in the segment where I started. Switching between windows or tabs can only be done with the mouse and is therefore counter-productive.

    Regarding the TM I can only second what LBTranslations said. But this thing is different from a TM, a concordance search can only provide ideas and suggestions taken from previous translations. It can never be used for auto-filling segments because even an exactly identical string can have totally different meanings in different contexts.

    Have a nice year, everyone :)

    Edit: Oh, and while we're at it: What LBT said, please fix the existing bugs first. The job lists still show the wrong rates where the old rates still apply, and the list of completed jobs is still totally mixed up in terms of sorting. The validations (the yellow warning sign) create ten times more false alarms than useful ones, at least in my pair. And regarding the TM, I can only say I am happy that I hardly encounter it anywhere.

    Edited by kvstegemann
  • 0
    Avatar
    Bettina

    Dear Lara, 

    I wish you a very happy New Year 2017.

    I tried out the dictionary and I find it very helpful.

    Frequently, I come across a difficult term in translation that I know I have translated before and have spent quite a lot of time finding an appropriate phrase in my language, but don't remember what it was. These are often terms were the dictionaries I usually use don't help. I have thought about creating my own personal dictionary by writing down all these terms but found that it would be time-consuming to do. Thus, if Gengo would do it automatically, it would definitely help me in my work.

    Thanks for your effort in making our work easier for us!

    Bettina

  • 0
    Avatar
    Magda Falcone

    What a great idea and tool! ... I agree with the other translators (source text showed, client ID of my job, no "glossary" name & trustable results) and my contribution to this discussion is the following:

    Possible name > Term search bar (very simple and clear and this is what the tool has been designed for)

    New feature > "Exact match" field (with a checkbox)

    My kindest regards and Happy New Year Everyone,

    Magda

    Edited by Magda Falcone
  • 0
    Avatar
    George

    Hi Lara and best wishes to everyone for a Happy New Year.

    A few months ago I asked Gengo's support for the possibility to sort previously approved jobs based on various fields: customer ID, date, amount etc.The main reason for that request was exactly what the "dictionary" is trying to offer now: find previous translations of a term and maintain consistency. 

    I like the new tool and will use it. As mentioned by other fellow translators, it would be a lot more useful if the original text was also included. 

    My request to be able to sort previous jobs still applies. I may want to see my jobs and earnings from a particular client, my jobs of July 2015, all Pro jobs and other information based on various criteria. The present tool is a first and good step. However, my previous experience shows that tools like this are announced but disappear soon for no apparent reason. I hope this will stay.

    Kind regards,

    George

  • 0
    Avatar
    margaqf

    Hi Lara and all my fellow translators from Gengo. 

    First of all, I may share with you my best wished for this new year; let´s do our best effort and grow together. 

    My first contact with the new tool has been positive. Mainly a kind of autoupdating database aiming to help us to do better translations. I am afraid that, among a wide range of results, a translator that has doubts on a term could be more confused, since these results do not indicate which of them are correct and which not. The integrated Glossary and the Matches tool provide us translators –according to my point of view– with enough support, along with our own tools and the customer´s glossaries. 

    Anyway, this tool could be improved to develope into something else that is really useful.

    Best regards,

    Marga   

  • 0
    Avatar
    Kathleen

    Thank you for giving us the opportunity to provide feedback. I must say, as it stands now, I would only use this for looking up my own work.

    I agree with others who say that context is key, and so is the assurance of quality. I also would find it extremely helpful to bold the search term in the resulting text. Given these three improvements, there would be much more to cheer about. Long live revision!!!

    Happy New Year to all translators and all Gengo staff!!

    Kathleen

     

  • 0
    Avatar
    martha.dayo

    i like this! another thesaurus-like function for me, similarly to eijiro et al. maybe not useful as a dictionary per se, similarly to eijiro et al lol.

    i like the one phrase/sentence return for other translators work. but for my work it returns an entire text, so that's not helpful. if it could pick out just the pertinent phrase/sentence for "my work" as well that wld be fine.

    i looked up for example ものづくり which is always a bit of a pain. i know in the past i had settled on a good "explanatory" translation but i never remember it, and any time it comes up i have to search back through past jobs, so this memory is totally up my alley.

    surely loaded with errors but surely useful when one already understands the meaning.

     

    hi lara!!!

  • 0
    Avatar
    mirko

    This "concordance search" tool needs to show entire translation units, that is to say, source+target, otherwise it's relatively useless. Perhaps you could also find a way to integrate it directly within the translation space, in an unobtrusive way (e.g. a collapsible tab, a popup, or something like that), and maybe also add a shortcut to do a search on a highlighted word/sentence, a bit like in a CAT tool.

Please sign in to leave a comment.