mltext in Web Quick Start

[aussen2]

In the website I am building on top of Web Quick Start (Alfresco 3.4d) I have two different kinds on multilingual web assets.

a) Assets that require translations of it's metadata as well as its content (e.g. text articles)
b) Assets that have the same content in every language but require translations of it's metadata (e.g. photos)

My question regards the second kind of assets (b). I want to use the Alfresco built-in multilingual capabilities for mltext because otherwise I would have to create different copies of the same content just for the purpose of translating its metadata. This is especially problematic as I want to associate assets of type b with assets of type a (example: a text article contains photos and when I create a translation of that text article I want to use the same photo instead of creating and associating a copy of it).

Currently the WQS webapp does not support the transmission of any locale information when communicating with the Alfresco repository. Is there any way (or any work in this direction) to use the mltext functionality in WQS (via Share UI) and then request the different translations of the metadata from the WQS webapp?

[bremmington]

No, there is nothing in the current development plan that would offer better support for MLText properties. Sorry.

[aussen2]

I was going to ask if it would work to extend the ApplicationDataInterceptor of the WQS web app so that it would change the locale setting based on the path information. But now I found that there already is such a functionality implemented in HEAD. So my question now is: How could I get this locale information included in the WQS API calls to the repository? I would like to use it as an additional request parameter in the WQS module web scripts so that I could modify these to return the content of my ml:text fields for the requested locale instead of the default server locale. Any hint would be appreciated.

[bremmington]

Sorry - missed this question.

Rather than doing what you are suggesting, I would tend to alter the WQS REST API so that it returns all the MLText values for each MLText property on each requested node, and then put the logic for selecting the correct value into the WQS client API layer. This would mean all the asset caching in the client API would continue to work usefully.

To describe exactly what would need to change to make this work is beyond the time I have available right now, but the areas to look at would include:

1. The node service implementation used by the AssetGet webscript (the current one interprets MLText values into single text values)


2. The AssetSerializerXmlImpl and AssetDeserializerXmlImpl classes to marshal/unmarshal MLText objects into XML


3. The AssetImpl class to process getProperty requests for properties that are MLText.

Hope that offers some guidance.