XML to XML transformation with Feature Differentiation (ditaval)

[Anonymous1]

Hello,

we have a large documentation project that includes different products and therefor a lot of feature differentiation.

Now we want to release the smallest product out of this project and need to translate it first. The problem is that we don't want to translate every topic of the whole project but only the topics and elements that apply to the smallest product.

Is it possible to have an XML (out DITA topics) to XML (also DITA topics) transformation that outputs only a certain feature differentiation? So only our small product. We could then import this smaller batch of XML topics into our translation memory system (memoQ).

I couldn't find any XML to XML transformation in Oxygen for my use case. Does anybody have a recommendation on how to achieve a filtered DITA XML output?

[Radu]

Hi,

There is a DITA OT plugin called "normalize":

https://github.com/dita-ot/org.dita.normalize

we ship it with the latest Oxygen 19.0 but you can also install and use it separately. But it will not generate the content exactly as it was before, the DOCTYPE declaration will be missing + various other changes as well.

Regards,
Radu

[Anonymous1]

Thank you for the suggestion. This is not exactly what we where hoping for. We will try to solve this differently.

[Aliahc]

Forgive me if I'm being naive; but isn't what is being described by B-E-N a case of applying conditional profiling?

For example, whether at topic level or content-level, you apply metadata (e.g. product attribute) to be either larger product, smaller product or [larger product smaller product] such that it is an OR action for the content to apply to both products should allow filtering in publishing.

I am currently doing that for different documents using different depths of the same content; i.e. a user guide, a tutorial, a quickstart reference - all having overlapping content, but do have slight differences that can be caught by such profiling.

Is this not essentially what is being described?

[Radu]

Hi,

From what I understand he wants to apply profiling but he wants to obtain DITA content in the output folder in order to give that content to the translators.

Regards,
Radu

[Anonymous1]

Radu is right.

We'd like to have the same topics but only with the content of a certain feature differentiation. DITA to DITA so to say. Or XML to XML.

The XHTML output, for example, resolves all variables and related links. This is not what we want. We want to stick remain an XML source to give to the translators and afterwards use those filtered, translated XML files to build our documents (PDF and WebHelp).

[annetteb]

Hello B-E-N,
Did you solve your problem? And if so, what did you do?
We have the same issue and still haven't found a good solution.
Best regards,

[Radu]

Hi,

One possible solution might be to first copy the entire DITA project to a separate folder, then open the DITA Map in the DITA Maps Manager, right click the root and choose "Refactoring->XMl Refactoring->Delete element" and somehow come up with an XPath which removes all elements which have a certain combination of profiling attribute values. After this, right click again the DITA Map and choose to export it to a zip file.

Regards,
Radu

[annetteb]

Hi Radu,
Thanks for the suggestion but this solution cannot be used for "mass production". In our team, the people who prepare the package for localization do not know Xpath at all (and I'm not an Xpath expert either), so it would be hard to define the right set of expressions for each ditaval.
Best regards,

[Radu]

Hi Annette,

I think that you could define a set of XPaths and then create some internal documentation to be used with a precise set of steps which could be followed through even by people who do not know XPath.
Of course like every workaround it depends if you are more willing to do what you currently do than implement it.
At some point probably a future version of Oxygen will have built-in support for such filters and when it does we'll update this forum thread.

Regards,
Radu

[xephon]

Hi B-E-N,

you should give Maxprograms Fluenta DITA Translation Manager a try. It creates an XLIFF file for you, which you can send to the translator. You pass a DITA Map and a DITAVAL file to Fluenta, so it performs some kind of publishing for you. When you receive the translated XLIFF file, Fluenta extracts the file for you. Fluenta has a small Translation Memory included, so it remembers translated segments/sentences.

Greetings,
Stefan