Hi Tim,

I got stymied by this problem as well. Not sure what the details of
your project are, but you might find the Google Language API useful.
It has a language detection method that works pretty well and even
returns a confidence value (http://code.google.com/apis/ajaxlanguage/
documentation/#Detect). It's not perfect, but it's suitable for most
purposes.

Cheers,
-- Matt

On May 26, 6:57 pm, Tim Hastings <[email protected]> wrote:
> Thanks for the reply Matt.
>
> Did Summize compute the language from the content? -- bonus clever
> points if so!
> I may have to look into doing that classification myself. Gulp.
>
> Cheers,
>
> Tim
>
> On May 26, 4:35 pm, Matt Sanford <[email protected]> wrote:
>
> > Hi Tim,
>
> >      The language code is only available in search and not in any  
> > other output. You are totally correct that language is a throw back to  
> > the Summize.com days. I don't know of any current plans to implement  
> > language identification in the other systems but I'll look into it.
>
> > Thanks;
> >   – Matt Sanford / @mzsanford
> >       Twitter Dev
>
> > On May 26, 2009, at 2:56 AM, Tim Hastings wrote:
>
> > > Hello,
>
> > > I notice that in the search results that each status update includes:
> > > iso_language_code = en, nl etc...
>
> > > Language is not present in the timeline, data mining feeds, or the new
> > > spritzer streaming feed.
>
> > > The language is very useful for excluded content that a user is not
> > > going to understand, but I do not want to make plans if it is going
> > > away.
>
> > > Is language a throw back from the aquired search platform?
>
> > > Any information about plans for support for iso_language_code in other
> > > API calls would be greatly appreciated.
>
> > > Thanks in advance,
>
> > > Tim Hastings
> > >http://tagwalk.com/

Reply via email to