Translation:<text> tags and segmentation in MemoQ TMS

Please post all questions relating to Help & Manual 6 here!

Moderators: Alexander Halser, Tim Green

Post Reply
Andreas Boettcher
Posts: 21
Joined: Fri Jun 07, 2013 7:52 am

Translation:<text> tags and segmentation in MemoQ TMS

Unread post by Andreas Boettcher »

Hi,

I'm using Kilgray MemoQ to translate HuM topics. One thing I'm having problems with is segmentation caused by <text> tags.
The problem is easily illustrated when applying a style to individual words within a sentence, which in turn has a style. I'd expect the sentence to form one segment. Formatting within the sentence would ideally be represented by tags in the TMS.

Example text:
This is a dummy text. It serves to show how text formatting information is stored in the XML file.

Example XML (translate attribute removed for easier scanning):
<para styleclass="Normal">
<text styleclass="Normal">This is a </text><text styleclass="Text Style">dummy</text><text styleclass="Normal"> text. It serves to </text><text styleclass="Character Style">show</text><text styleclass="Normal"> how text </text><text styleclass="Character Style">formatting</text><text styleclass="Normal"> information is stored in the XML file.</text>
</para>
Segmentation is fairly excessive, since MemoQ interprets every <text> tag as structural, thus forcing a new segment. I could tell MemoQ to interpret <text> tags as inline. This would create a huge number of other problems though.

Is there a way to avoid the described segmentation problem (other than not applying formats to individual text elements, obviously)?
User avatar
Tim Green
Site Admin
Posts: 23156
Joined: Mon Jun 24, 2002 9:11 am
Location: Bruehl, Germany
Contact:

Re: Translation:<text> tags and segmentation in MemoQ TMS

Unread post by Tim Green »

Is there a way to avoid the described segmentation problem (other than not applying formats to individual text elements, obviously)?
There are two solutions for this: 1) Use styles instead of manual formatting. 2) More importantly: Don't use a translation tool that forces you to work on the individual XML tags -- any tool that forces you to do that is counterproductive and will waste time, work and money. If translation memory isn't important to you it's easiest to use Help & Manual itself as the translation editor. The Basic version is fine for that -- it can edit Pro and Server projects without restrictions, it just can't create them (but the translator doesn't need to do that).

If you need translation memory, either work with a translator who uses a more professional tool like SDL Trados (the industry leader) or you can also use the ICanLocalize translation service. It comes with free translation memory software that has full support for Help & Manual projects, and also does all your project management and billing management. You can locate a translator through the service, then it is free for you (they take a cut from the translator for the mediation). Or you can work with your own translator using the service, with the "private translator" option. Then you pay a small percentage fee for using the service.
Regards,
Tim (EC Software Documentation & User Support)

Private support:
Please do not email or PM me with private support requests -- post to the forum directly.
Andreas Boettcher
Posts: 21
Joined: Fri Jun 07, 2013 7:52 am

Re: Translation:<text> tags and segmentation in MemoQ TMS

Unread post by Andreas Boettcher »

1) I am using styles. Granted, I am applying them sort of manually within styles, but that should be possible, shouldn't it? How else am I going to do highlighting for text items such as company terminology or product names?

2) SDL Trados is NOT an option for me or the company I work for. Having said that, SDL Trados uses inline tags and structural tags, just as MemoQ does. And just like MemoQ, SDL Trados seems to rely on the concept of structural tags to determine segments. I believe that the described segmentation issue would therefore also occur in SDL Trados. As a result, I assumed that there is a solution to my problem, other than buying new software.

I am mildly surprised to be confronted with what looks like a sales pitch when all I did was asking for a solution to a problem that seems to have been solved for users of SDL Trados.
User avatar
Tim Green
Site Admin
Posts: 23156
Joined: Mon Jun 24, 2002 9:11 am
Location: Bruehl, Germany
Contact:

Re: Translation:<text> tags and segmentation in MemoQ TMS

Unread post by Tim Green »

Hi Andreas,
I am mildly surprised to be confronted with what looks like a sales pitch when all I did was asking for a solution to a problem that seems to have been solved for users of SDL Trados.
It hasn't been "solved" for Trados users. Trados is just a usable translation memory tool that doesn't force you down to the source code. And that was hardly a sales pitch -- we have nothing at all to do with the Trados company. They don't even give us any information, let alone money. We have to figure everything out ourselves. The ICanLocalize service also has no connection with us. They are just two examples translation memory tools that provide a usable user interface when working with Help & Manual. If you want a sales pitch, try using Madcap Flare. They produce their own (not cheap) translation memory tool that they want you to buy. 8)

Our choice was to provide 100% XML source so that you can choose your own tool. However, if you are using a tool that forces you to look at individual XML tags then that's sub-optimal and you need to find something better. We can't do anything about that since we only generate the XML. And we really can't provide individualized options for individual tools -- that would go far, far beyond our development time capacity.
I am using styles. Granted, I am applying them sort of manually within styles, but that should be possible, shouldn't it? How else am I going to do highlighting for text items such as company terminology or product names?
If you're formatting individual pieces of text inside paragraphs then that text will have additional tags. That can't be avoided, unfortunately. That is simply how XML works. If you can't use a translator-oriented tool like Trados, Nero Across or ICanLocalize then the alternative is to use Help & Manual itself as the translation editor, in combination with the Project Synchronization tool for updating your translations. Working directly on the XML code is never a viable option. It may seem cheap at first, but you will always damage the XML in the process and then the hours you spend locating and fixing the problems created will eat up your savings many times over.
Regards,
Tim (EC Software Documentation & User Support)

Private support:
Please do not email or PM me with private support requests -- post to the forum directly.
Pavel Doronin
Posts: 4
Joined: Wed May 27, 2015 2:44 pm
Contact:

Re: Translation:<text> tags and segmentation in MemoQ TMS

Unread post by Pavel Doronin »

Hello Andreas,
I had the same problem with memoQ. Actually, it does no matter which tool do you use: Trados, memoQ, Across or any cloud-based tools like Memsource or Smart CAT (I tried many different tools). Trados is not 'more professional' as Mr. Green says. memoQ is even more flexible. As soon as you have many styles, conditional tags and complex markup, XML output from H&M is generally not localization friendly if we speak about translation with external tool. I have spent weeks to find an easy solution for handling XML topics from H&M with memoQ. I can share my experience how to create custom XML filters with memoQ and set proper segmentation.
Phil Turner
Posts: 2
Joined: Wed Jan 13, 2016 8:42 am

Re: Translation:<text> tags and segmentation in MemoQ TMS

Unread post by Phil Turner »

Hi Pavel,

I know it's been a while since this topic was started, but I'm now experiencing the same problems as Andreas. MemoQ starts a new segment with each XML tag. You said:
I have spent weeks to find an easy solution for handling XML topics from H&M with memoQ. I can share my experience how to create custom XML filters with memoQ and set proper segmentation.
Can you help me with this? I have no idea what to do in MemoQ to get this working. Though I think that it's rather about the XML filters that should be set correctly than about segmentation rules.

Regards
Phil
Andreas Boettcher
Posts: 21
Joined: Fri Jun 07, 2013 7:52 am

Re: Translation:<text> tags and segmentation in MemoQ TMS

Unread post by Andreas Boettcher »

Phil,

You are correct, you will have to use a custom XML import filter to get useful segmentation in MemoQ.
This comes at a price though, which is a ridiculous number of tag markers in the translation editor and consequently your translation memory.

If you can live with that, you can use the attached XML filter I created to import H&M topic files. The filter interprets <text> tags as inline. As a result, <text> tags will no longer affect segmentation.

To keep your TM clean, you'd have to manipulate the topic XML prior to import.
The way I see it, you'd have to identify every <text> tag that shares a common set of attributes and replace it with a unqiue tag.

<text styleclass="Normal"> becomes <textnormal>
<text styleclass="Bold"> becomes <textbold>
etc.

Only then could you modify your import filter more flexibly. Of course, you'd have to revert all the replaced tags after exporting your target file.

I've come to the conclusion that Help and Manual relies on an XML structure that just doesn't work well with modern translation environments and workflows.

edit: Seems like neither mqres files nor zip files are allowed. Just PM me your e-mail address and I'll e-mail the file to you.
Phil Turner
Posts: 2
Joined: Wed Jan 13, 2016 8:42 am

Re: Translation:<text> tags and segmentation in MemoQ TMS

Unread post by Phil Turner »

Hi Andreas,

thanks for your response. And sorry for my late response, I hadn't seen yours. Someone already sent me a working filter configuration file in the meantime, and it works. But you are right ... lots of tags in the editor and the in the TM. I suppose I do have to live with that as long as I'm working with H&M. I don't want to mess around with the files to be honest, so I'll just keep all the tags. Works for me. If I ever switch to a different documentation tool, I'll probably just do a clean-up of the TM and everything should be alright. One other thing came to my mind: Do you have a configuration file that makes the preview work?

Regards
Phil
Pavel Doronin
Posts: 4
Joined: Wed May 27, 2015 2:44 pm
Contact:

Re: Translation:<text> tags and segmentation in MemoQ TMS

Unread post by Pavel Doronin »

Guys, sorry, I've missed your messages. If somebody still needs my help with memoQ filters, feel free to connect with me via LinkedIn, I'll be glad to help you.
Pavel Doronin
Posts: 4
Joined: Wed May 27, 2015 2:44 pm
Contact:

Re: Translation:<text> tags and segmentation in MemoQ TMS

Unread post by Pavel Doronin »

Me and my colleagues have created a short and simple guide for creation a well-tailored memoQ filter. Please find it here: https://yadi.sk/i/hYvvWgRgrguLb
I hope it helps.
Post Reply