Questions about TRANSLATION process
Moderators: Alexander Halser, Tim Green
Questions about TRANSLATION process
The dam is about to burst, as I've been asked to make my largest manual available for translation. I have some experience with this from a former life when I used FrameMaker and I sent the FM files to a service that used translation memory software. But that was then, this is now. This will be my first crack at it in H&M. I have not yet spoken with the translation service (someone else did the vetting) but I know they use some kind of translation memory software. Exact type TBD. For the purposes of the following questions, let's assume their system is compatible with H&M.
1) According to the H&M documentation, my project should be in uncompressed .HMXP format. Are there any other format issues I should be aware of? (For example, I haven't followed up on the TidyXML feature. I'm guessing I should convert to TidyXML before proceeding. Please confirm.)
2) Most of the snippets I use are snippets from within the project, so they should work without issues. But there are a couple of snippets that are outside of the project files, from a central snippet repository. How do I deal with that?
3) If the translator is using translation memory software, what exactly will they send back to me? Will it be a translated .HMXP project that will then live as a parallel version of the original English version? And do I just compile and publish it in parallel with compiling and publishing the English version?
4) If YES to #3, what's the normal way to manage revisions? I would speculate the following: Revise the English version, make a copy of the project, and send to translator. Translator's software can parse the project and identify changes, and will send back a NEW parallel translated project. That parallel translated project REPLACES the existing (older) translated project. Is that correct?
1) According to the H&M documentation, my project should be in uncompressed .HMXP format. Are there any other format issues I should be aware of? (For example, I haven't followed up on the TidyXML feature. I'm guessing I should convert to TidyXML before proceeding. Please confirm.)
2) Most of the snippets I use are snippets from within the project, so they should work without issues. But there are a couple of snippets that are outside of the project files, from a central snippet repository. How do I deal with that?
3) If the translator is using translation memory software, what exactly will they send back to me? Will it be a translated .HMXP project that will then live as a parallel version of the original English version? And do I just compile and publish it in parallel with compiling and publishing the English version?
4) If YES to #3, what's the normal way to manage revisions? I would speculate the following: Revise the English version, make a copy of the project, and send to translator. Translator's software can parse the project and identify changes, and will send back a NEW parallel translated project. That parallel translated project REPLACES the existing (older) translated project. Is that correct?
e
d
============================================
Ed Hawco
d
============================================
Ed Hawco
Re: Questions about TRANSLATION process
Additional question:
The documentation for TidyXML says "You can only get the full benefits of Tidy XML if you use styles instead of manual formatting." I do use styles for everything at the paragraph level (and for tables), but within paragraphs I don't use styles for simple bold and Italics. Is that considered "manual formatting?"
In other words, when I switch from regular XML to TidyXML, and I going to have a problem with all those bold and Italics items?
Update: I ran a test conversion, and the manually-applied bold and Italics are formed as <text style="font-weight:bold;"> and <text style="font-style:italic;">. Is that considered "manual formatting" or is that considered "using styles?"
The documentation for TidyXML says "You can only get the full benefits of Tidy XML if you use styles instead of manual formatting." I do use styles for everything at the paragraph level (and for tables), but within paragraphs I don't use styles for simple bold and Italics. Is that considered "manual formatting?"
In other words, when I switch from regular XML to TidyXML, and I going to have a problem with all those bold and Italics items?
Update: I ran a test conversion, and the manually-applied bold and Italics are formed as <text style="font-weight:bold;"> and <text style="font-style:italic;">. Is that considered "manual formatting" or is that considered "using styles?"
e
d
============================================
Ed Hawco
d
============================================
Ed Hawco
- Tim Green
- Site Admin
- Posts: 23184
- Joined: Mon Jun 24, 2002 9:11 am
- Location: Bruehl, Germany
- Contact:
Re: Questions about TRANSLATION process
Hi Ed,
Don't worry about bold and italics, those are basically unavoidable tags. If you did them with styles you'd also have tags; they'd just say the style name instead of bold or italics. Swings and roundabouts -- nothing lost, nothing gained.
Don't worry about bold and italics, those are basically unavoidable tags. If you did them with styles you'd also have tags; they'd just say the style name instead of bold or italics. Swings and roundabouts -- nothing lost, nothing gained.
That's helpful for translators working directly on the XML, for example with translation memory programs. It doesn't matter if the translator is using Help+Manual as the translation tool.1) According to the H&M documentation, my project should be in uncompressed .HMXP format. Are there any other format issues I should be aware of? (For example, I haven't followed up on the TidyXML feature. I'm guessing I should convert to TidyXML before proceeding. Please confirm.)
Include the XML snippet files for the translator to work on. Then create a new repository for the translated version and add the translated files to it when you get them back, and link that repository to the translated project when you get it back.2) Most of the snippets I use are snippets from within the project, so they should work without issues. But there are a couple of snippets that are outside of the project files, from a central snippet repository. How do I deal with that?
Exactly. You send them the entire project and get the entire project back, translated.3) If the translator is using translation memory software, what exactly will they send back to me? Will it be a translated .HMXP project that will then live as a parallel version of the original English version? And do I just compile and publish it in parallel with compiling and publishing the English version?
Precisely. You just send the translator the new version of the project. The translation memory then pre-processes it for them: Every sentence that has already been translated gets replaced automatically. Changed and new material gets flagged and the human translator then works on that.4) If YES to #3, what's the normal way to manage revisions? I would speculate the following: Revise the English version, make a copy of the project, and send to translator. Translator's software can parse the project and identify changes, and will send back a NEW parallel translated project. That parallel translated project REPLACES the existing (older) translated project. Is that correct?
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.
Tim (EC Software Documentation & User Support)
Private support:
Please do not email or PM me with private support requests -- post to the forum directly.
Re: Questions about TRANSLATION process
Thanks Tim. I'm glad my assumptions where correct. Feeling confident!
e
d
============================================
Ed Hawco
d
============================================
Ed Hawco
- Tim Green
- Site Admin
- Posts: 23184
- Joined: Mon Jun 24, 2002 9:11 am
- Location: Bruehl, Germany
- Contact:
Re: Questions about TRANSLATION process
By the way: Ask the translator about costs for doing the pre-processing with the already-translated text from the translation memory. I've heard reports of some translators charging juicy fees for running the new project through the pre-processor with the existing translations from memory. Just a cautious heads-up...Ed Hawco wrote:Thanks Tim. I'm glad my assumptions where correct. Feeling confident!
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.
Tim (EC Software Documentation & User Support)
Private support:
Please do not email or PM me with private support requests -- post to the forum directly.
Re: Questions about TRANSLATION process
Ummm... Huh? I'm not sure what you mean. the text isn't translated yet. Do you mean when they're doing a revision? (Also, what is the "pre-processing?")Tim Green wrote:Ask the translator about costs for doing the pre-processing with the already-translated text from the translation memory.
Another question: Is it useful to send the translator the entire project, or do they only need the "Topics" folder?
e
d
============================================
Ed Hawco
d
============================================
Ed Hawco
- Tim Green
- Site Admin
- Posts: 23184
- Joined: Mon Jun 24, 2002 9:11 am
- Location: Bruehl, Germany
- Contact:
Re: Questions about TRANSLATION process
Ed Hawco wrote:Ummm... Huh? I'm not sure what you mean. the text isn't translated yet. Do you mean when they're doing a revision? (Also, what is the "pre-processing?")
That's relevant when you do the revision. Then you send them the complete revised original project, in the original language. The translation memory tool then "pre-processes" this before the human translator starts work on it. During this, the tool automatically translates all known sentences from its translation memory database, which was stored when the original version was translated. Some translators reportedly charge quite hefty fees for this automatic procedure.
Always send the entire project folder. There are texts in the main .hmxp file and the \Maps folder contains the table of contents file(s), which also need(s) to be translated.Another question: Is it useful to send the translator the entire project, or do they only need the "Topics" folder?
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.
Tim (EC Software Documentation & User Support)
Private support:
Please do not email or PM me with private support requests -- post to the forum directly.
Re: Questions about TRANSLATION process
OK, got it. Given that I revise this manual on average of once a month, that pre-processing thing really could be an issue! Thanks for pointing that out.
You've said that if the translator is using translation memory software that they don't have to use H&M. I'm assuming they will be smart enough to KNOW there is text bits in the project file and the maps folder that needs to be translated. Is that a safe assumption?
You've said that if the translator is using translation memory software that they don't have to use H&M. I'm assuming they will be smart enough to KNOW there is text bits in the project file and the maps folder that needs to be translated. Is that a safe assumption?
e
d
============================================
Ed Hawco
d
============================================
Ed Hawco
- Tim Green
- Site Admin
- Posts: 23184
- Joined: Mon Jun 24, 2002 9:11 am
- Location: Bruehl, Germany
- Contact:
Re: Questions about TRANSLATION process
It doesn't hurt to point it out, Murphy is everywhere. You should also give them the SDL Trados configuration files for Help+Manual that you can find in the \Templates\Translation\ folder in the HM program directory.Ed Hawco wrote:I'm assuming they will be smart enough to KNOW there is text bits in the project file and the maps folder that needs to be translated. Is that a safe assumption?
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.
Tim (EC Software Documentation & User Support)
Private support:
Please do not email or PM me with private support requests -- post to the forum directly.
Re: Questions about TRANSLATION process
Hello. This translation project has ballooned and is getting huge! More documents, more languages...
Anyway, I got the first project back from the translator, and first impressions are good. However, there are a few issues:
- Although I sent a note along with the project saying which build conditions do NOT need to be translated, all text has been translated. This means we're paying for translation of writer's notes, stuff that is intended for English only, etc. Should I take this as an indication that the translator is not very experienced with this kind of project? Or is parsing build conditions within the translation memory software particularly difficult? (This is important because one of the documents to be translated has three versions built in and only one needs to be translated -- roughly half of the document.)
- The HTML project GUI still has elements of English, such as the text over the Search field ("Enter one or more keywords to search ('*' and '?' wildcards are supported):") and all tool tips ("Next Topic," "Previous Topic," etc.) Where does that text live within an H&M project? Is this also a sign of inexperience on the part of the translator?
FYI, I sent the entire uncompressed project, including the Maps folder, Baggage folder, etc. The project's Config > Common Properties > Language Settings is correctly set to "French (Canada)."
Please advise...
Anyway, I got the first project back from the translator, and first impressions are good. However, there are a few issues:
- Although I sent a note along with the project saying which build conditions do NOT need to be translated, all text has been translated. This means we're paying for translation of writer's notes, stuff that is intended for English only, etc. Should I take this as an indication that the translator is not very experienced with this kind of project? Or is parsing build conditions within the translation memory software particularly difficult? (This is important because one of the documents to be translated has three versions built in and only one needs to be translated -- roughly half of the document.)
- The HTML project GUI still has elements of English, such as the text over the Search field ("Enter one or more keywords to search ('*' and '?' wildcards are supported):") and all tool tips ("Next Topic," "Previous Topic," etc.) Where does that text live within an H&M project? Is this also a sign of inexperience on the part of the translator?
FYI, I sent the entire uncompressed project, including the Maps folder, Baggage folder, etc. The project's Config > Common Properties > Language Settings is correctly set to "French (Canada)."
Please advise...
e
d
============================================
Ed Hawco
d
============================================
Ed Hawco
- Tim Green
- Site Admin
- Posts: 23184
- Joined: Mon Jun 24, 2002 9:11 am
- Location: Bruehl, Germany
- Contact:
Re: Questions about TRANSLATION process
Hi Ed,
It's hard to say what the problem is with the text that shouldn't have been translated. It won't have been translated automatically if it was the first time, that would be a human translator. You'd need to check with them first to see if there was a misunderstanding about what should and shouldn't have been translated.
Those "GUI" elements you refer to aren't in the project, they are in the WebHelp skin. The search interface texts from Zoom actually already have translations in language files that you can select. Open the skin in Help+Manual and go to Configuration > Publishing Options > WebHelp > Full Text Search. In the lower right area there are load and save icons where you can load and save language files. French is definitely available, along with most other western European languages. There are a couple of interface texts that belong to the skin itself rather than Zoom, including "Enter one or more keywords". Depending on the skin those will be in either the text variables in the skin (Configuration > Common Properties > Text Variables) or in the source code of the search template, in the same Full Text Search section referred to above.
It's hard to say what the problem is with the text that shouldn't have been translated. It won't have been translated automatically if it was the first time, that would be a human translator. You'd need to check with them first to see if there was a misunderstanding about what should and shouldn't have been translated.
Those "GUI" elements you refer to aren't in the project, they are in the WebHelp skin. The search interface texts from Zoom actually already have translations in language files that you can select. Open the skin in Help+Manual and go to Configuration > Publishing Options > WebHelp > Full Text Search. In the lower right area there are load and save icons where you can load and save language files. French is definitely available, along with most other western European languages. There are a couple of interface texts that belong to the skin itself rather than Zoom, including "Enter one or more keywords". Depending on the skin those will be in either the text variables in the skin (Configuration > Common Properties > Text Variables) or in the source code of the search template, in the same Full Text Search section referred to above.
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.
Tim (EC Software Documentation & User Support)
Private support:
Please do not email or PM me with private support requests -- post to the forum directly.
-
- Posts: 4
- Joined: Tue Jul 10, 2018 10:26 am
- Location: Russia, Moscow
Re: Questions about TRANSLATION process
Hi Ed,Ed Hawco wrote: - Although I sent a note along with the project saying which build conditions do NOT need to be translated, all text has been translated. This means we're paying for translation of writer's notes, stuff that is intended for English only, etc. Should I take this as an indication that the translator is not very experienced with this kind of project? Or is parsing build conditions within the translation memory software particularly difficult? (This is important because one of the documents to be translated has three versions built in and only one needs to be translated -- roughly half of the document.)
If you are talking about build conditions that are set on the topic level (when you define to which build a topic must be included), there is no way a CAT tool (more generally accepted name for what you called 'translation memory software') can determine what topic belongs to what build. Thus, the only thing your translation agency can do is manually search for XML files that are included in the build you want to translate and ingore the rest.
In my company, before sending files to a translation agency, we use a PowerShell script that parses the XML file that contains the table of contents to automatically remove XML files that are not included in the build we translate from the Topics folder.
If you are talking about build conditions on the text level (conditional text), there is a possibility to teach a CAT tool to ignore text that is not included in the build you are translating: Basically, the translation agency must re-configure settings for the filters that they use to import your XML files to their CAT tool (for example, if the translation agency uses SDL Trados, then they need to adjust the SDL Trados configuration files).
As for the writer comments, there is an element called <draft-comment> that must be ignored when importing your XML files into a CAT tool.
If you cannot do what I have mentioned above in-house, you must explain to your translation agency what builds and conditional text are, and explicitly state in your translation briefings what builds must be translated. That way you are eligible not to pay for content that should not be translated, but nevertheless was translated. Same applies to the writer comments.