Feedback on Open Data Management for Public Automated Translation Services

Hi all,

I forwarded the link to the document (https://www.w3.org/International/its/wiki/Open_Data_Management_for_Public_Automated_Translation_Services) to a couple of the core Linport people since Linport would fit into this picture. I got the following back from Jörg Schütz, which he asked me to forward on.

Hi Arle,

Thanks for providing this link. I was already thinking about how this initiative could contribute to the overall "Linport and related" discussions, and what kind of exchange could and should be possible.

My main concern with the W3C initiative, and the requirements document in particular, is that it is very much machine translation focused (even with a strong bias to SMT) which is certainly only one particular aspect within our current API (here, I use API is a generic metaphor for service interfaces) considerations.

In my view, "Public Automated Translation Services" can be more than just SMT and their associated data management (languages, lifecycles, processes, etc.) facilities. Since I'm not sure if this focus is because of the relation to CEF, but looking only at an MT infrastructure when talking about automated translation services would be to shortsighted (here I use "automated" in terms of processes). Think about, for example, terminology support to guide a particular translation request, or the assessment of a translation source content or a translation result for pre- and post-editing purposes, and you are confronted with data and interoperability challenges.

So, in priciple, the initiative might consider an overall framework for translations, and this then would directly lead to the Linport discussions. Obviously, all aspects mentioned in the document so far can be extended to a general framework for public translation services.

I would have liked very much to share my view in today's W3C online meeting but unfortunately 14:00 CEST is not a good time for me during an ordinary workday... Nevertheless, you may forward these lines to the W3C group working on the requirements document, and I would be happy to further explain my view.

Thanks again, and all the best,

Jörg

I think Jörg’s comments point out that we should be clearer (both internally and publicly) about the particular motivation for this document and consider whether we want to talk in terms of a broader infrastructure/ecosystem than just SMT. Since the CEF focus is MT-centric, the present document reflects that, and maybe we want to keep that particular focus, but we might also profitably discuss whether it could be broadened a bit as well.

Best,

-Arle