Oct 28, 2013

Want a revolution? Try memoQ 2013 Release 2.


OK, so I'm exaggerating a bit. And even though the new version of memoQ was officially released today by Kilgray, it really is still beta software. But damned good beta. I expect that there will be more of interest to individual translators added in this version of memoQ than in any other version I've seen up to now. Lots of T's to cross and i's to dot still, but there is great promise, and it's worth having a look now at the future of memoQ.

I'm not talking about changes to the memoQ Server. There are lots of those in this version, and for a change many of them actually seem to be helpful to translators working on the server and less focused on slicing and stuffing linguistic sausage faster like many of the 6.x server features introduced. The rollout webinar with István Lengyel and Florian Sachse of Kilgray showed enough of why memoQ Server users should be pleased. But they could have filled the hour and three quarters with nothing but presentations of new or improved functions for the rest of us and still not run out of material. Since I still have a project to finish tonight, I'll just hit a few of the highlights that I'll probably return to later as the features stabilize and are truly ready for productive work.

Language recognition
memoQ now intelligently recognizes the language(s) of the source text. This is a small convenience in setting up projects perhaps, but for those occasions when a source language has many passages in another language or more than one other language, these other language segments can be identified automatically, copied source to target and locked. I can think of more than a few patent dispute translations where this would have been helpful.

Startup Wizard
A new feature under the Help menu gives a quick, friendly guided tour of important settings that are often overlooked that are hard to find for new users and many experienced ones. This is actually one of my favorite new features and possibly the best help I've seen yet for making a better start with the software.

Better Microsoft Word spelling integration
Custom dictionaries can now be imported from Microsoft Word with greater ease. Users can now also choose Microsoft Word for dynamic marking of possible spelling errors (unknown words). This is a good thing for those of us who hate Hunspell. Oh, and those pesky doubled words are caught now.

More stuff with Microsoft Word...
like exporting tracked changes between translation versions to a DOCX file (sans formatting I think), exporting target comment to a DOCX file (alas! in writing the specification Kilgray failed to consider that one might want to select which comments get exported and possibly suppress all the comments, but I'm told this will be remedied quickly), font substitution in DOCX files (this was a major WTF feature for me, but if I understood correctly, there is some way I can use this to protect text formatted a certain way, such as code in a programming guide - if that's true, this is cool) and...

the TM lookup tool,
an external application which runs in Microsoft Word and any other environment and allows you to look up text copied to the Clipboard in selected memoQ TMs. Too bad they didn't include termbases in this new feature. Yet.

New filters and processes
like direct import of InDesign files with a preview using the free online Language Terminal integration, Adobe InCopy and some file formats that must be pretty damned geeky because I've never heard of them.

Why am I excited about
a plain text view which is about as exciting as lukewarm, unspiced pea soup. Well, because it's absence has been driving me nuts for years now. It's in this version.

Meanwhile, back at the termbase
great things are happening with new import options that are still a wee bit buggy but will get very good very soon. Until now memoQ could only import terms as TMX and delimited text. New options include Excel (at last!), MultiTerm XML and TBX. It was child's play for me to tweak a couple of TermStar MARTIF exports from STAR Transit to import those terms, because TBX is a dialect of MARTIF and STAR's MARTIF is very close to TBX. Extra effort? About 2 minutes of search and replace so I'm hoping Kilgray will go the extra five yards and touch this import option down.

The addition of the MultiTerm XML import option means that memoQ users can now roundtrip data from memoQ to partners using SDL MultiTerm and back for termbase updates. Unfortunately at the moment, the only meta data transferred in the import is the definition field, but efforts are in progress to support at least the MultiTerm fields memoQ exports to XML with Kilgray's own definition. That was simply forgotten at specification time (oops). But still, this will be serious headache relief for those of us who work in teams with SDL Trados users and want to share terminology in the most effective ways.

Is that all?
No. This new version of memoQ is like a very messy Christmas where one can easily lose the overview of hat's under the tree with all the wrapping paper and bits of ribbon cluttering the floor. As it gets cleaned up, we'll all notice a good bit more, and I suspect that Santa's Hungarian and German helpers will be slipping a few more things under the tree that they might forget themselves until some user trips over them. There has been so much effort put into consolidation and improvement of existing features that it's simply too much to keep track of. I've made a list and checked it more than twice and still find things to add. But I'll end with another look at something I've already blogged about, that groundbreaking

Monolingual Project and TM Update
with edited files in any target format. It still has a lot of little quirks, especially with some formats, but here I expect a lot of improvements. I've made a little demonstration video and put it on YouTube; it shows the reimport of edited translations to update the translated file and the TM in memoQ, and it shows two different ways to look at tracked changes before revealing the dark secret of Row History Recovery which I think Kilgray didn't realize was possible. Well, damnit, they should have made it a feature with a button anyway.

 
(View this in full screen mode by clicking the icon at the lower right of the video window.) 

Oh yes, and one more cool little thing about this release that I forgot to mention...

... the quickstart shortcut to creating memoQ projects
in the context menu by right-clicking on a file. I'm not much into single-file projects any more and prefer to use "container" projects for customers or categories instead, but it's still a nice little addition that can save time once in a while:



12 comments:

  1. Thanks so much for this, Kevin - I missed the memoQ webinar live so it's great to find out about the new features. Your video on how to import a revised version of the target file is really useful. I've been trying it out now as I write, on a 28k book I translated. Unfortunately it hasn't worked - the two sides bear no resemblance to each other and the whole thing is full of pink cross-links. Perhaps the two-column table format in Word has confused it. Oh well, I'll just find a more straightforward text to try it out on! But in principle, a promising feature.

    ReplyDelete
    Replies
    1. You aren't trying to import a bilingual RTF are you? I saw someone misunderstand earlier today and try to use this feature with a bilingual file rather than an edited monolingual one. (This reminds me of the legitimate scenarios to test with bilingual source files, but that's a different can of worms.) Make yourself a small sample for testing at first and bear in mind that this feature will involve and improve a lot in the near future. Only a small part of the full specification has been implemented, partly for reasons of available resources, but perhaps also because Kilgray didn't expect the new feature to be this well received. (In the background I hear a chorus of legal and financial translators singing "I told you so!")

      I totally missed out on Florian's comment in the webinar that the alignment window will be upgraded at some point to mark differences, good to see that this is a priority. What I didn't show in the video is how one can also use the filter settings just to show the change-marked segments. That combined with switching on the tracked changes to compare with the last exported version can be rather helpful to get an overview in a big document. The preview (if available) will give you the context that is lost in the translation grid when the filter is applied.

      Delete
    2. "Involve"? I meant "evolve" of course. Evil autocorrect elves strike again. Or Freud's Little Helpers.

      Delete
    3. Kari, I just read your more detailed description of those files in the Yahoo list. OMG. Interesting scenario. I think I would precede it with a lock on the other language using the recognition feature, but this'll give me something strange to test on a smaller scale in any case.

      Delete
  2. Thank you very much for this detailed overview! Out of interest, could you also explain how you converted the TermStar Martif files to import them in MemoQ? This sounds really promising...

    ReplyDelete
    Replies
    1. Not today, Greg, but soon probably. I have asked a few people to give me some small, non-proprietary samples with synonym structures and meta data that I can use for demo and testing; my clients would not thank me to show the process with the data I have at the moment. Basically I just looked at the differences between a TBX that worked and STAR's MARTIF structure. Two minutes of search & replace later the data imported just fine to memoQ. It was easy enough that it seems to me Kilgray could add TermStar-dialect MARTIF imports easily enough. But it probably won't happen soon if the people who can use it don't make that clear to Kilgray Support. There are a lot of priorities to juggle right now, and those who are silent about their real needs for real projects may have a long wait. The product planners are pleased to hear feedback; without it they really can't be expected to do much. So speak up, please!!! to "support@kilgray.com"

      Delete
  3. Wow! Thank you so much for your articles on 2013 R2, Kevin! I've been reading them in preparation for taking the leap from 6.0.59. It looks like I'll find lots of great new features. And thanks to your helpful comments and videos, I'll have a fighting chance of actually finding and using them! Thanks again!

    ReplyDelete
  4. Thank you, Lisa. There have been many steady improvements since this post first went up, and I've discovering new gems all the time that Kilgray never got around to mentioning. And as of Build 53, it looks like most of the irritating issues have been resolved. More or less on the schedule I expected.

    ReplyDelete
  5. Dear Kevin, thanks so much for your post. I am particularly interested in the language recognition feature. However, it looks like memoQ now doesn't support the scenario you've described, i.e. it cannot identify segments in a different language, copy them to target and lock them. Perhaps, I am missing something.
    Another thing that can be quite useful is for memoQ to be able to import Word changes from a Word document with the track-changes mode enabled and export the translations of these changes so that they show as the translated changes. Now this doesn't seem to be supported, either. But I could be wrong for all I know:)
    Arkadi

    ReplyDelete
    Replies
    1. Perhaps you are missing something? If you are using a recent version of memoQ, certainly. The language recognition feature comes into play in several different ways. For the copy/lock behavior, try having a look at the bottom of the dialog that appears for Operations > Lock/Unlock Segments.... I discussed the details of this in a blog post and/or video a while ago, pointing out some of the limitations I think, but I can't recall exactly where. There is about a 2% error rate in some of the tests I've run with German/Dutch/English, but that's good enough for processing to produce a reasonable cost estimate if you have a 50,000 word document with mixed-up language that you want to quote without investing a lot of time. (Of course, with larger documents there are often risks from embedded text, bitmap graphics, etc. so I often run a parallel volume check by OCR to confirm with reasonable approximation any counts obtained by the word count function of a software program based on the original file.)

      Delete
    2. As for the tracked changes, there are several ways to accomplish that without much ado in many cases. But until the full, original specification for the monolingual review feature is implemented, I really don't have the energy to discuss the details of dealing with the difficult editing examples.

      Delete

Notice to spammers: your locations are being traced and fed to the recreational target list for my new line of chemical weapon drones :-)