How do I use XML files converted from FM into OmegaT?

[Denjiba_manual]

How do I convert FM files into XML (or any other file format that sustains the file format which can be used into Omega T) so that it can be imported into the open source translation software OmegaT?

I an able to use [Save as] and create a XML file but an error message appears when I try to import it into Omega T and I am unable to translate the file.

The reason I’m trying to use XML files for translating in Omega T, is because in the past the person who is in charge of editing the FM file would create a text file with sentences that need to be translated. I would then import the text file into OmegaT to translate and send the text file with the translated sentences to the editor. This would then be edited into a FM file.

This would be no problem if the content that needs to be translated was little. But in some cases there would be a large amount of content that needs to be translated which would take more time to edit the FM files.

What I’m trying to find out is if I could use a file format that can be used into Omega T and then be able to convert that file into FM without the editor re-edit the file.

Regards

Denjiba-manual

[Jeff_Coatsworth]

Wouldn't you just save as MIF and send that to the translation company?

[Bob_Niland]

I have zero experience with this, but the advice so far seems a bit off the mark. As I read the basenote, there is no external translation service engaged.

How do I convert FM files into XML (or any other file format that sustains the file format which can be  used into Omega T) so that it can be imported into the open source  translation software OmegaT?

According to its own page, and Wiki, OT supports a limited number of XMLs. Any default XML that Frame currently generates is probably not one of them, nor is MIF. You might be able to define the output transforms so that Frame emits something close to a supported input XML, but then will you be able to round-trip the translated text back into Frame?

OT also supports HTML and XHTML for input, so outputing those from Frame might be worth a try, but you'll face the same round-trip challenge.

I am able to use [Save as] and create a XML file but an error message appears when I try to import it into Omega T...

What is the error?

Have you tried to "register the format file extension into the preferred file filter"?

Possibly OT could easily be trained to work with MIF, and just process the actual text strings.

What OT brings to to DTP is growing adaptive dictionaries that aid translation. If you can't solve the round-trip problem, you might as well just copy the entire Flow A out of the Frame document into a plaintext file. Translate it (probably leaving the source language in place for location alignment purposes), then paste it back and apply formats as needed.

_________

Or use a different translation tool that is Frame-aware (and I wouldn't know of any).

Probably won't be free.

[Jeff_Coatsworth]

I don't know about you, but I'd never trust my help files to some freebie translator software - have you ever seen the results of a Google Translate page (if you know another language)? You get what you pay for IMHO ;>)

[Bob_Niland]

I don't know about you, but I'd never trust my help files to some  freebie translator software - have you ever seen the results of a Google  Translate page (if you know another language)? You get what you pay for  IMHO ;>)

I quite agree. My experience in this is limited to mocking up what two of our covers would look like in Mandarin, using Google. Apart from being a font nightmare, it was clear that the opportunities for mistranslation were legion. Alas, I hear that the "professional" translation of my documents resulted in key operating facts being incorrect. I keep telling management we need to pay for two translations, from two different providers, one BACK into English, so we can get a rough damage assessment.

That said, OmegaT doesn't seem to be a "machine translation" tool. Out of the box it evidently translates nothing. It appears to be a collection of tools, and an adaptive memorizer, that allows easy re-use of translations already typed in by the professional translator.

[Bob_Niland]

A search of the web indicates that others have asked about this before, and there's even an enhancement request into the OT project to support MIF (but it's 2 years old, and doesn't seem to be going anywhere).

When other people have asked on various forums, they get an answer similar to this thread:

• suggestions for a path into OT (such as MIF > MIF2go > Word > OXML > OmegaT) ...
• ... but pointing out that it's a one-way trip.
  

[JohannesKrueger · Accepted solution]

Regarding the "Save as XML" function in FrameMaker:

You probably know that in FrameMaker, you can work with "structured" and ordinary documents.

When using "Save as XML" with an ordinary FM document, the resulting XML is merely an output format and not suitable for re-opening in FM. You can translate the resulting XML file, but once you get the file back into FM, you will have to re-apply all the formatting and put the original figures back into place. (At least, this is what I know from FM8.)

Otherwise, when working with a structure application in FM,  you can save structured FM documents as XML for editing. This means, you can edit (translate) the XML document outside FM, and open the edited (translated) XML file in FM.

Regarding OmegaT:

I see two ways of translating a FM document with OmegaT. For both, you'll need the Okapi framework to convert your files to an intermediate format.

One way requires that your document is a structured FM document, which you can save as XML for editing. The XML document has probably none of the document types that OmegaT can translate right away. You can, however, use a utility from the Okapi framework to extract translatable text from the XML document into a XLIFF file. Later you can merge the translated text back into the original XML file, which you can bring back into FM.

The other way is similar, but starts with a MIF file from FrameMaker. In this case, it doesn't matter whether the FM document is structured or not. Obviously there is a MIF filter for the Okapi framework, which allows you to extract translatable text from MIF into an XLIFF file. Just as above, you later can merge the translated text back into the original MIF file. However, the MIF filter is in beta state, so don't expect it to work flawlessly.

More on the Okapi framework can be found at http://okapi.opentag.com.

All that said, I've only played around a bit with OmegaT and Okapi. At work, I'm using an ancient version of Trados.

Johannes

[Denjiba_manual]

Thank you everyone for your answers.

After talking it over with my superiors, XML is something we need to work on in the near future and not something we can implement right this minute. Currently we are using Adobe Framemaker 10 (apologies for not saying this earlier) so I don’t know if the answer JohannesKruger gave is 100% but is definitely something to look into.

In the meantime we are using HTML from the help files of our software which is pretty much the same as the paper document to translate, instead of copying and pasting word strings from FM to text and then using the text to translate and editing back it into FM.

Regards

Denjiba-manual