RE: [bug] Provenance

Ok, make sense then.

Thanks.

-ys

 

From: Leroy Finn [mailto:finnle@tcd.ie] 
Sent: Wednesday, January 09, 2013 7:14 AM
To: Yves Savourel
Cc: Multilingual Web LT-TESTS Public
Subject: Re: [bug] Provenance

 

Thanks Yves. I haven't re ran the output in a while and Pablo was updating the provenance input files over Christmas so that the input files were more in line with LQI . I will have to correct a few things in the parser for provenance which I will try to get around to this week. I should have updated the output sooner though. Thanks for the help.

 

Thanks,

Leroy

 

On 9 January 2013 12:59, Yves Savourel <ysavourel@enlaso.com> wrote:

Hi Leroy, all,

We've finally decided to implement Provenance.
And when running the test suite we've found quite a few issues with both the input and output files.

They are listed here:
https://github.com/finnle/ITS-2.0-Testsuite/commit/af878d03ca7b4f4fd0d62094801f2bb13d1f60d8
https://github.com/finnle/ITS-2.0-Testsuite/commit/f8442cf8dbd5c8be7ec0e0e67e5f073783d350c0

Most files were incorrect.

Some of the problems were just related to output conventions, for example: Spaces vs tabs to separate the fields; fields ordered by name then by index instead of by index then by name; index on provenanceRecordsRef; etc. I went by what was decided in the LQI output since it makes sense to have all standoff output following the same conventions.

But other things were a bit stranger: provenanceRecordsRefPointer="#pr1" instead of provenanceRecordsRefPointer="@xml:id"; many invalid XMLs (extra closing tags); output data completely different from the input data, etc. We should have caught those issues long ago. Didn't anyone implemented Provenance yet?

Anyway, hopefully several of you can double-check my corrections and make sure we have proper input/output now.

Thanks,
-yves