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

RE: [all] call for concensus on Translation Provenance Agent (related to ISSUE-22)

From: Yves Savourel <ysavourel@enlaso.com>
Date: Thu, 26 Jul 2012 06:57:05 +0200
To: "'Dave Lewis'" <dave.lewis@cs.tcd.ie>, "'Multilingual Web LT Public List'" <public-multilingualweb-lt@w3.org>
Message-ID: <assp.0554c8c960.assp.0554dceb11.019c01cd6aeb$1b084cc0$5118e640$@com>
Hi Dave, all

Thanks for the updated text Dave.

A few notes:

1) Spelling

The text seems to be in UK English (e.g. organisation vs organization).
I think we use US spelling in the ITS specification.


2) IRI vs URI

In ITS 1.0 we used URI. I can't recall exactly why (maybe IRI was not final yet then?). But we need to be consistent and use one or the other for 2.0.


3) Example text (very minor)

To avoid the wrath of the purists, in the example for local markup its-trans-agent="C3PO" should be its-trans-agent="C-3PO", if it refers to what I think.


4) Global rules

The information we carry with this data category is: a) the direct provAgent or provRevisionAgent values, or b) a reference to a resource where those agents are defined. That is nicely expressed in the local attributes.

my understanding of the global rules in general is that they have two purposes:

- One is to provide a way to map an existing construct in another vocabulary that implements the same data category.

- The other is to allow to annotate the content of a document from a different location, so the annotation doesn't interfere with the content itself (e.g. because the vocabulary does not allow you to have ITS attributes locally).

I think the proposed global rules don't cover the first goal. We can associate a prov(Revision)Agent and a prov(Revision)AgentRef defined in a global rule with selected nodes, but we cannot tell that a given element or attribute of the host vocabulary has existing constructs that implement such information.

In other words: the values of prov(Revision)Agent and prov(Revision)AgentRef can be held only by ITS attributes.

For example how would we indicate that 'agent' and 'revAgent' are the equivalent of its:provAgent and its:provRevisionAgent in this document:

<text>
 <title>Translation Provenance Agent: Local Test in XML</title>
 <body>
  <par agent='C-P3O' revAgent='Luke'>This paragraph was machine translated and then postedited.</par>
  <legalnotice agent='Luke'>This legal text was subject to translation by manual means.</legalnotice>
    </body>
</text>

(DocBook or DITA may have better examples).

I think we'll need four extra attributes in the global rules: provAgentPointer, provAgentRefPointer, provRevionAgentPointer, and provRevisionAgentRefPointer. See the Localization Note data category for an example of similar pattern (http://www.w3.org/International/multilingualweb/lt/drafts/its20/its20.html#locNote-implementation)


Cheers,
-yves
Received on Thursday, 26 July 2012 04:57:31 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:31:47 UTC