Yesterday one of my product managers told me that their product is in the final stages of being certified for sale in the EU, and asked me if we have a translation firm we work with since some EU countries require native language documents. Target is about 6 months from now.
As far as I know, my company has never had anything translated.
So far I only have one document that would be needed for this product, but it's in unstructured FrameMaker and makes fairly extensive use of variables. I think there's only one condition in use, though, for commenting/questions.
All graphics use callouts with a separate, FM table as a legend. I've also been working on consistency in information presentation and words by applying the principles of simplified technical English as far as I can and also using AutoText to insert standard language for procedures and descriptions. This is a long process, though, since I'm basically teaching myself how to do this and I'm also the lone writer.
What do I need to do to make sure the document(s) I supply can be used by a translation company as easily and cheaply as possible? Do most translation companies accept Frame or MIF files? What do I have to do about variables and conditions? I'm unlikely to be involved in selection of a translation firm, but on the off chance I am, what should I ask about and/or guard against?
I work with such a translation company.
I think they take MIF as input and produce as output a MIF with translated text.
They have people to translate them into European languages, Russian, and Japanese.
Maybe others as well, but those are the languages I know that they do.
Afterwards, it is necessary to hand the output off to someone (me, for example) to fix the output. For example, page breaks. Or graphics captions that are too long for the text box, and they need to be enlargened.
The customer in the end receives new FM and MIF and PDF files.
Discuss with the translation agency what they will do! Be in close contact! And always tell them that you will take care of all questions! Much better than getting a translation which must be changed.
All larger agencies will accept MIF. Never convert to Word!
I recommend that the translation agency collects special terms from your document and translates this and gets the approval from your local contact! This way it does not need to be discussed after the translation which term should have been used.
Discuss, if they will also translate variables and cross-refererence formats and automatic numbering and text from the master pages, etc.
I do the final layout myself. And I wrote a FrameScript script which afterwards applies the correct language (also in table formats), updates all variables from a list (as otherwise it happens that these variables are translated differently in different projects; should not happen, but it does), updates the cross-reference formats and automatic numbering (if such a format or paragraph was not in the text to translate, it will not be translated).
If the agency should do this, you have to provide also copyright files, TOC, index, etc. You might delete the content of the TOC or index. (Some agencies really translated also the TOC and index!)
Discuss, how the document will be proofread. Some agencies have a proofreading web-based tool, others give you the MIF files, you create the PDF and send it to the proofreader.
The proofreader should _not_ change everything what he/she does not like or would have phrased differently. Only real mistakes! There will be discussions with your proofreaders anyway. If you have requirements for your style, then you must discuss this!
Discuss, if the agency can do an pretranslation with a machine translation tool like DeepL or Google Translated. This will then be corrected by the translator. Usually this reduces the costs by 10 to 20 %. In my experience only larger agencies offer such a service.
I just tested how conditional text works in the CAT software I use. If you have conditional text, you will need to make all the conditional text visible ("Show All"). If it is hidden, it may be embedded in the MIF file somewhere, but it doesn't show up in the CAT tool). Here is an example of how it looks in a CAT tool when the condition is set to "Show All":
'Tag2' is the conditional text marker (I called the condition 'Yes').
(FYI, when translating, the tags are inserted in the same location in the target language to make sure the formatting remains intact. Variables, text formatting (bold, italics, etc.), cross-references, and markers, etc., all show up as tags, so the translator can't mess with them.)
Anyway, before creating a MIF file, set all conditional text to 'Show.'
The translation company I am using right now recieve my FrameMaker files exported to XLIFF. And I get XLIFF back and insert in the original docuements. Works fine, mostly.
Don't I have to be using structured FM with XML files for that?
Oh, this should be very helpful. Thanks.
Keep the suggestions and information coming. It's all grist for my mill and it is greatly appreciated.
Great video and information. Unfortunately, my CAT tool can't seem to handle the XLIFF 1.2 that is outputted by FM. Maybe in an update....
For clarity, do you suspect that it's a shortcoming of your tool, or of FrameMaker?
Probably the tool. It says it cannot handle XLIFF 2.0. (FM shows that it is creating a 1.2 file, but I guess that doesn't work either. But I only spent a few minutes testing, so there's always the possibility of user error. 😃
If it is a translation company that uses CAT (Computer-Aided Translation) software (Trados, Wordfast, etc.)--and all the big firms do--then you should be able to supply them with a MIF file and they can translate it without any major issues. (I use a CAT tool and round-trip from English-to-Spanish or English-to-Japanese.) I have never tried doing this with conditional text, but my guess is that the MIF file will make all the conditional text accessible to the CAT tool to enable it to be translated. (And CAT tools use markers for formatting, etc., so none of your settings and variables should be adversely affected--when you get the MIF file back and open it in FrameMaker, the variables, etc., should display just like they did before the translation. Ditto for Index markers--they show up as separate 'sentences' in the CAT tool I use (WordFast) and can be translated. In the example below, Line 8 has [Tag 1] (a cross-reference marker) and [Tag 2] (an index marker). The line below it is the index marker text.
Sure, there may be some glitches when round-tripping, but it should turn out much better than roundtripping to Word and back.
Definitely do not convert to Word.
Don't delete the contents of the TOC. They need to see how they are supposed to be formatted.
One problem that we have seen is where the source text contains forced line breaks. Then when it goes through automatic translation, it looks like two separate sentences, and the translation comes out wrong.
Note that each language has its own punctuation symbols and rules. For example, upside-down question marks, or << instead of quotation marks. Or in English, you don't leave a space before a colon, but in some languages, you do. Or spaces instead of commas in writing numbers.