W3C home > Mailing lists > Public > public-multilingualweb-lt@w3.org > November 2012

Re: [ACTION-256]: Compile and circulate itsTool examples togehter with proposal text

From: Felix Sasaki <fsasaki@w3.org>
Date: Wed, 7 Nov 2012 21:53:24 +0100
Message-ID: <CAL58czrxNN5mRbHKt7n_7NG4B8+53qmHsHJMaG-y7O0xq0EXTg@mail.gmail.com>
To: Yves Savourel <ysavourel@enlaso.com>
Cc: public-multilingualweb-lt@w3.org, Ankit Srivastava <asrivastava@computing.dcu.ie>
2012/11/7 Yves Savourel <ysavourel@enlaso.com>

> Hi Felix, all,
>
> > Defining the parameters (version, value, ...) woudl mean that we
> > do the same as in the "toolInfo" element, but just don't serialize
> > as XML, but as URI parameters. The problem stays the same:
> > no interop between
> > mymt.org/toolinfo?version=456&value=FR-to-EN-General
> > vs.
> > mymt.org/toolinfo?version=x6h3&value=fr-t-EN
>
> Indeed the values are not interoperable (they never have been in Mt
> Confidence).
>
> But do the MT Confidence users wants to be able to access the values even
> if they don't understand them?
> These values are mostly labels anyway. They could be at least displayed by
> all tools in a more consumable manner than a URL.
>
> I'm asking because not defining the parameter' names would be losing some
> level of interoperability compared to the current proposal in the draft.
>


We could also describe this like "it lowers the burden for implementers".
If not every producer of MTConfidence information wants to use these
values, but just an arbitrary string for identifying the tool, having no
need to create URL parameters is a big plus - see the "collation URI"
example. We have a use case of gathering content with metadata for MT
training. I assume with mass data, such a use case will just compare
strings and not parse URIs to detect the values in them.

Looking at
https://docs.google.com/spreadsheet/ccc?key=0AgIk0-aoSKOadG5HQmJDT2EybWVvVC1VbnF5alN2S3c#gid=0
"implementers" means you and Ankit. Ankit, others, any thoughts?

>From the discussion about text analytics annotation, I have the impression
that we couldn't easily come up with field values either, but not sure.

Best,

Felix



>
> Cheers,
> -yves
>
>
>
>


-- 
Felix Sasaki
DFKI / W3C Fellow
Received on Wednesday, 7 November 2012 20:53:49 UTC

This archive was generated by hypermail 2.3.1 : Sunday, 9 June 2013 00:25:03 UTC