Studio 2011: FrameMaker tagging
Thread poster: philwhite
philwhite
Local time: 21:08
German to English
Mar 20, 2012

Hello again!

I am currently banging my head against the desk trying to get very large FrameMaker projects working cleanly in Studio 2011.

I am facing two problems:

The first problem is that the manuals (well over 1000 pages in total) contain many segments which are identical except for the fact that one segment contains, for instance, an index marker and the other does not. The same applies to other markers, for instance for a list of abbreviations, and to special characters, in particular non-breaking spaces. If I completely translate a given manual, inserting or removing such tags as the fuzzy matches are offered to me, save the files and then create a new, identical project, I only receive 97% in the new project. This is largely due to the fact that only one version of these quasi-identical segments is stored in the translation memory, which means that I need to work through 3% of a completely translated manual (this represents 30 or 40 pages of the entire volume of manuals!) Inserting or removing tags, depending on which version of the segment is stored in the TM. If such segments are identified as being different in the editor, surely they must also be stored as separate segments in the TM. Is there any way to force the correct behavior?

The second problem concerns filtering. I have introduced a "Product" field as a pick list in which multiple values are not permitted. The appropriate product is then used for filtering and updating when I am translating a particular manual. As I work through, there are numerous segments which are identical in the different manuals, but despite setting the filter, the translation from the product I am translating still only gives me a 99% match, as I have the multiple translations penalty turned on. Surely, the whole point of the filter is that you can have multiple translations and select the correct one using the filter. To my mind, there is no point in using a filter if it does not work like this.

The combination of these two issues means that I have a significant overhead any time that I need to carry out a small update on a large manual. If the customer knows that there are no more than, say, 1000 words on new and changed material in the manual, I cannot charge them for the 5 or 6 thousand words that I actually need to process in one form or another.

Am I missing something fundamental with both of these issues?

Any help would, as always, be appreciated.


Direct link Reply with quote
 

SDL Community  Identity Verified
United Kingdom
Local time: 21:08
English
Using Filters and multiple translations Mar 21, 2012

Hi Phil,

On the index markers and multiple translations. I think you just need to save the segments as new translations so you don't keep overwriting them and end up with only one TU in the TM. So when you see these segments use ctrl+shift+u or whatever shortcut you have for this in Translation - Add as new translation

On the use of fields. This is tricky and I think we could have worded this better but I believe your problem is not releated to the use of multiple translations but rather the selection of the field as a filter in the first place... at least given your comments on being unable to have multiple translations this seems likely.

So, when you add your filter to penalise the TM results you don't want the window says "The following filters will cause filter penalties to be applied for non-matching translation units:". I find this a little confusing because it actually works as a double negative. So if you have two products, Fruit and Cars for example. And you want results using the field Fruit to be penalised then when you add this you would say the condition was NOT product="Fruit". So after adding Fruit you must tick the check box to say NOT.

If I do this then my results are all CM or 100%. If I don't tick the box I see a Filter mismatch penalty and get a 99% match, and this is what I think you are getting at.... I hope?

Regards

Paul



[Edited at 2012-03-21 08:21 GMT]


Direct link Reply with quote
 
philwhite
Local time: 21:08
German to English
TOPIC STARTER
The dawning of realization! Mar 21, 2012

Thank you so much for your help.

Yes, saving the translations as new units works in 99% of the cases I have encountered (exceptions being segments where numerals differ and the number of non-breaking spaces differ). Quite frankly it had never occurred to me to add these as a new translation unit, because the editor regards the source as being different, look up regards the source as being different, and it never crossed my mind that differently tagged segments would not be saved as separate segments.

As far as filtering is concerned, your option to use the NOT flag seems a little bizarre in the sense that in order to filter on Product1, I would have to in fact filter on NOT Product2, NOT Product3 ... NOT Product17, which seems an excruciatingly painful way of killing the cat (although my dog may well appreciate that).

No, in fact I have worked out what is going on. I should be using fields that allow multiple values. If I have an identical sentence in two products, and a filter field that does not allow multiple values, the value "Product1" will by default be overwritten by "Product2" if I confirm the segment in Product2. This means that in future translations of Product1, I receive a filter penalty. I would have expected a new translation unit to be created in this scenario. Clearly, I should be using fields with multiple values and applying the CONTAINS operator. This is preferable to creating new segments for every exact match subject only to a filter penalty.

As usual, with all the Difficulties I have encountered migrating from Workbench to Studio, the behavior in Studio is logical and more fit for purpose than in Workbench, but the documentation, for instance regarding the differences in filtering, is woefully inadequate. In the last couple of months, I have spent far more time trying to migrate large FrameMaker projects from Workbench to Studio than I have translating productively, which is surely not the point of the productivity tool. The only way in which migration was viable at all was by using my knowledge of regular expressions to manipulate and exported TM to account for the very considerable differences in tagging between Workbench and Studio.

But I see the benefits and potential of Studio and do not regret the 20 days or so that I have invested into the migration.


Direct link Reply with quote
 

Jerzy Czopik  Identity Verified
Germany
Local time: 21:08
Member (2003)
Polish to German
+ ...
What kind of tags are those? Mar 21, 2012

When migrating from Framnemaker 7 to 8 or newer, the format changed and the MIF files are now Unicode. However, the old FM files needed special characters for non breaking spaces or typographic quotes.
Should the tags around the quotes be just font changes, check the formatting in Framemaker - it can be, that just reformatting the document there will help. You could even replace all typographic quotes in the source by plain ones and still keep the right ones in target. BTW, when the tags around the quotes are just pairs, you can also simply remove them. As long you remove start AND end tag of formatting, the file shoud produce target without any problem.


Direct link Reply with quote
 
philwhite
Local time: 21:08
German to English
TOPIC STARTER
I have to use tags Mar 21, 2012

I have already captured the quotes and the few dozen other formatting changes between FM7 and FM10 and between Workbench and Studio. Even after several days identifying and scripting the systematic changes, we only get about 90% of the 1000+ pages of manuals we have translated. (Without scripting global changes to the old Workbench Database before it was imported, we had about 65%.) Two of us have now put aside a week to retranslate existing manuals to capture this remaining 10%.

The non-breaking spaces are crucial, as we have segments where the only difference is in a single numeral and the number of non-breaking spaces. By default, Studio displays the nonbreaking spaces as text and does not distinguish between one or multiple non-breaking spaces. By displaying special characters as tags (this primarily affects tabs and non-breaking spaces, I have a chance of hearing, the spaces (I am visually impaired and use a screen reader) and Studio identifies a non-match on the basis of different tagging. Otherwise, it puts in a translation with the wrong number of spaces.

Saving all the many variants of essentially the same segment as new segments causes all of them to be penalized during pretranslation. When working through the translation, The first such segment is offered correctly, and when I confirm it, all the other 30 or so segments also go in correctly at 100%. Without using tags, they went in wrongly.


Direct link Reply with quote
 

Jerzy Czopik  Identity Verified
Germany
Local time: 21:08
Member (2003)
Polish to German
+ ...
Studio displays non breaking spaces like Word Mar 21, 2012

When you work with Studio and select the option "Show whitespace" (Menu View), Studio displays the same sign as Word, namely the small circle in uppercase, similar to the sign for degreess °.
This is for MIF8 and newer. In MIF7 the non breaking space was a tag.
You can also remove the multiple translation penalty in project settings, but this might not be the best solution.

OTOH why do you need multiple spaces? Usually we do evertyhing to avoid them, as typographicaly they are not correct.


Direct link Reply with quote
 
philwhite
Local time: 21:08
German to English
TOPIC STARTER
Yes, I know Mar 21, 2012

Yes, I know that that option is available, and I have it turned on, but my screen reader does not speak the whitespaces to me. It speaks the tags to me. The option of displaying special characters as tags is far more appropriate. Also, lookup behaves differently depending on whether non-breaking spaces are displayed as tags or not (as I explained above).

As I said, the non-breaking spaces are crucial in this company's products. Certain parameters have to be entered with a certain number of spaces, and this has to be documented correctly. It has nothing to do with typography.
S128 (no space)
S 64 (one space)
S  8 (two spaces)

Without tags and with numeral recognition turned on, Studio treats the last two as identical, which is catastrophic.


Direct link Reply with quote
 


To report site rules violations or get help, contact a site moderator:


You can also contact site staff by submitting a support request »

Studio 2011: FrameMaker tagging

Advanced search







SDL Trados Studio 2017 Freelance
The leading translation software used by over 250,000 translators.

SDL Trados Studio 2017 helps translators increase translation productivity whilst ensuring quality. Combining translation memory, terminology management and machine translation in one simple and easy-to-use environment.

More info »
Déjà Vu X3
Try it, Love it

Find out why Déjà Vu is today the most flexible, customizable and user-friendly tool on the market. See the brand new features in action: *Completely redesigned user interface *Live Preview *Inline spell checking *Inline

More info »



Forums
  • All of ProZ.com
  • Term search
  • Jobs
  • Forums
  • Multiple search