On Thu, May 29, 2014 at 2:15 PM, Benson Margulies <bimargul...@gmail.com> wrote:
> >> I hate to rain on your parade but any restriction on how the result is > >> used that goes further than the (very few) restrictions in the ALv2 > >> means that the result it can't be licensed under the ALv2 and that will > >> cause problems. > > Let me clarify the restriction to assist in deliberation. > > Step 1: Cordova pushes a lot of text through Bing translate and published > it. > Step 2: Some third party scrapes the web site in both languages, > aligns the results, and adds it to an MT corpus. > Step 3: Microsoft is unhappy. > > HOWEVER, > > I find it hard to believe that MS imposes this constraint in this > instance. It's one thing to tell a direct user of the API that they > can't go training models. This would seem to require any web site that > used Bing translate to have a TOS forbidding training MT engines on > the multiple languages. I'd clarify if MS is really asking for that. Also, if you choose to switch to a different translator at some point, it is likely that they will use the previous translations as the base for a translation memory even if humans are doing the translation. That counts as the project using the text to train a translation engine.