tag:blogger.com,1999:blog-20155610.post7728934755375382581..comments2024-03-06T02:46:19.929+00:00Comments on Translation Tribulations: What good is memoQ fuzzy term matching?Kevin Lossnerhttp://www.blogger.com/profile/14727800526216764023noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-20155610.post-81323704336989659652013-07-25T15:46:19.590+01:002013-07-25T15:46:19.590+01:00700,000 entries? I can't conceive of how that ...700,000 entries? I can't conceive of how that would even be remotely useful, as it would tend to drown what I need in "noise". The technical and legal glossaries I began 13 years ago in Trados and Déjà Vu probably have around 50,000 entries (I haven't bothered to check the count in few years), and that is really too much. These days my most valuable terminologies a QA tools with no more than a few hundred entries typically. Ah well, whatever I think of the mass data approach for getting work done, it is useful for stress testing. I remember years ago arguing with Kilgray developers that TM performance needed to be upgraded to handle my personal compendium of some 300,000 segments, back when they could not believe anyone would work with more than 50,000 segments in a TM. Now as you know they can swallow and use those 2 million TU TMX files from the EU DGT, though memoQ still gets indigestion if it sees that data in other formats.<br /><br />If you can afford the downtime to wait for the term editing operation to complete, I am very curious how long it actually takes. This may turn out to be an experiment like that tar drop that finally fell after a great-grandfather's lifetime of waiting.Kevin Lossnerhttps://www.blogger.com/profile/14727800526216764023noreply@blogger.comtag:blogger.com,1999:blog-20155610.post-70765131880564788422013-07-25T15:35:50.341+01:002013-07-25T15:35:50.341+01:00Hmm. I just posted a comment on my iPad but it see...Hmm. I just posted a comment on my iPad but it seems to have disappeared. Probably not a good idea to post from inside the Pocket app on the iPad.<br /><br />Anyway, yes, my TB is rather large. It weighs in at around 700,000 entries. However, since it didn't actually crash memoQ I suppose I might as well just try to be patient and wait.<br /><br />Michael Michael Beijerhttps://www.blogger.com/profile/12826804655385764008noreply@blogger.comtag:blogger.com,1999:blog-20155610.post-56894516052887106172013-07-25T13:17:29.936+01:002013-07-25T13:17:29.936+01:00@Michael: What was the size of the termbase you we...@Michael: What was the size of the termbase you were trying to do this with? The ones I have tested so far had a few hundred to a few thousand terms in them, but knowing your addiction to "big data" I suspect I'm off your scale by some orders of magnitude. Although memoQ handles some (not all) data operations faster than DVX or SDL Trados, when I am messing with very large data quantities (like XLIFF files with 100,000+ segments), I am able to choke the program.<br /><br />@Anette: Sometimes there might be good reason for exact matching of term, especially if there are similar terms with which it might be confused. Also, the compound word feature is currently still only active for German, so I can imagine cases (and I have them in German as well), where the 50% prefix setting might work better for an individual term. QA may also be an issue if fuzzy term matching is ever added to that - it could be disastrous if the target text check became tolerant of typos or bad spelling.Kevin Lossnerhttps://www.blogger.com/profile/14727800526216764023noreply@blogger.comtag:blogger.com,1999:blog-20155610.post-90600406029951379312013-07-25T10:22:26.887+01:002013-07-25T10:22:26.887+01:00Hmm, so I tried the tip in your second video, to c...Hmm, so I tried the tip in your second video, to change all my old termbase entries to ‘fuzzy’ in one fell swoop, but it didn't work. <br /><br />Last night, I selected all the entries in my main TB and then clicked on Fuzzy and, lo and behold, this morning memoQ is still churning away (‘Not responding’). Suppose I'll have to kill it and try again. I will report back later.<br /><br />Michael Michael Beijerhttps://www.blogger.com/profile/12826804655385764008noreply@blogger.comtag:blogger.com,1999:blog-20155610.post-76777262843570881282013-07-25T09:09:25.275+01:002013-07-25T09:09:25.275+01:00Interesting. I've been wondering about the ide...Interesting. I've been wondering about the idea with this feature myself. So what would be the reason not to set all the TBs with fuzzy matching? Anettehttps://www.blogger.com/profile/10150587643608721875noreply@blogger.com