My favourite conference by far is the ATA… and not just because the location changes every year and it’s always somewhere fantastic! This year was in Miami and the location was no exception… a great choice and huge numbers of translators interested in translation technology and how they can get more from it. My own involvement at these conferences is always very satisfying because I get to spend the full conference speaking to people, from the moment the doors open until they close, about technology. This year kept me particularly busy with two pre-conference sessions, mostly on Studio and MultiTerm; the “Toolbar” which is open all conference and is an area manned by technical representatives from all the tools vendors present; and a conference presentation I’ve wanted to deliver for a long time. This article is about that presentation, “XML, XPath, XSLT… the ‘X’ Files?”
I was a little worried I’d be talking to myself given it was the last session on Saturday afternoon but I was very pleasantly surprised to have a good showing of willing victims, and I hope I got my message across which is that these technologies can be helpful if you learn a little and are using translation tools that can take advantage of them. I had a brief chat with Jost Zetzsche afterwards who thinks these are difficult things to learn and the way forward is for tool vendors to simplify the way files are handled to take away the need for these technologies. I half agree with him… I do think that improvements in the translation tools should allow translatable text to be extracted in a simple way. But the other half of me, perhaps even more than half of me thinks that it’s not that simple and I’ve seen many files from customers where the complexity around what text needs to be extracted for translation means we’re a long way off providing an interface to deal with all eventualities in a simple way. So a good solution until we have this utopian capability is to make use of technologies that are used on a regular basis by the same people creating these X-Files for us to work with. A few simple lessons can take us a long way and I hope I succeeded in getting this part through!
So rather than repeat the contents of the presentation in this article I’ll refer you first to an article I’ve written in the past on XPath, and link to the slide deck and a recording of the presentation where it’s probably a little easier to see the smaller texts as I went through the examples. I’ve also put the sample files on a server where you can download them and have a play.
Thanks for everyone who attended… hope to have the chance to see you again next year!
If you just want to see the presentation then here it is:
(151030) The X files from Paul Filkin
But if you would like to watch the full presentation then I’ve recorded it and placed it here:
Length: 58 mins 36 seconds
Hi Paul,
Thanks very much for sharing the nice video!
Now I understand better how to extract content from XML files.
Btw, when I saw slide 8, I wished Studio’d had the “Examples” feature of DejaVu, that would’ve made my life easier dealing with XMLs, although I reckon that the tools mentioned in the video are very useful.
I agree with Jost that the challenge is creating a tool to easily parse these XML files.
XPath and regex don’t scare me, but I still see XML files as little bastards: sometimes I just give up fine tuning the prep step because I can’t get what I really want. So my guess is that XML is just a terrifying matter for many translators.
… Jesús Prieto …
Hi Jesús, that feature is quite nice (memoQ not Déjà Vu) but to be honest not that useful because it can only really tell you the stuff that’s easy to do anyway and Studio makes this pretty simple after you import the XML and just use the “element and attribute” options instead of XPath. You won’t find it useful for conditional extractions or things requiring operators but I guess it gives you the full path and you can then tailor it, but I much prefer the features in the Visualizer and XMLQuire because these show you in realtime whether you’ve found what you need even with really complex stuff.
Glad it was helpful for you though… in fact I feel I achieved something to hear this from an experienced person like you!!
Hi Paul,
The feature I mentioned is from Deja Vu (right hand side of slide 8 of the video or slide 7 of the slideshow).
Ah, I see what you mean. The same feature is in memoQ. I agree it’s quite nice because you see all the elements and attributes in one place. Studio is similar in fact because you can import the XML and you get all the attributes and elements in the filetype too, but it doesn’t list them all out like this. I prefer the Studio approach which gives a consistent mechanism for these simple rules and also more complex XPath constructs.
Hello Paul,
Great article but unfortunately the subject of file type settings is completely new to me and I am puzzled.
At least I found out that missing file type settings are the basis of my problem with some Heartsome files in Trados Studio 2015. The translation does never show in Heartsome 7 or 8.
Heartsome is gone already but a client continues using this so it is mandatory for me and the tool is a pain. Trados recognizes the files but not the suggested fuzzy matches that are contained in the .xlf file.
I tried to set up a file type but I encounter many unknown settings. I have a lot of .xsd files with settings in the Heartsome setup files but I do not know which to chose.
Would it be possible for you to have a look at the attached file and let me know how I can deal with this.
Thank you very much for your articles and your help.
Best regards,
Freundliche Grüße
Renate Lange
4000 Paraguari, Paraguay
Mobile: +595 981 748 963
Hi Renate. Sure, just email it to me – pfilkin@sdl.com