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

Re: Parameters for rules?

From: David Lewis <dave.lewis@cs.tcd.ie>
Date: Sun, 22 Apr 2012 22:07:47 +0100
Message-ID: <4F947323.2020005@cs.tcd.ie>
To: public-multilingualweb-lt@w3.org
Hi Yves,

Interesting idea. Did you have a specific idea in mind of how values 
might be bound to this variable?

Certainly its assigment could itself be a data category, but would we 
want mechanisms that bound the parameter to a value held outside of the 
file, e.g. a HTTP POST parameter passed with an ITS conformance file to 
a web service (this does get us into the broader scope of defining web 
APIs)?

Further thoughts on your specific questions below:

On 20/04/2012 13:02, Yves Savourel wrote:
> Hi all,
>
> One thing I've noticed during the past years working with ITS 1.0 is that, quite often, having a way to use 'variables' in ITS rules would have been a lot more efficient having to create multiple ITS files.
>
> For example, imagine some rules that say:
>
> In file 1:<its:translateRule selector="//*[@stage=LevelA]" translate="no"/>
>
> In file 2:<its:translateRule selector="//*[@stage=LevelB]" translate="no"/>
>
> When the calling tool could instead use a single file with:
>
> <its:translateRule selector="//*[@stage=${LevelValue}]" translate="no"/>
>
> Or something similar.
>
> I imagine that with the new data categories, this kind of construct would be even more handy than in ITS 1.0.
>
> This is not a possible requirement for a new data category, but rather an enhancement of the overall mechanism. So a few questions:
>
> 1) Would having such variable mechanism (with possibly default values in the rule file) be useful?

For this we need some concrete use cases. One generic case is when 
several rules and elements in a document repeat the same attribute 
value, but you would like to change it for different uses of the file, 
e.g. sending different versions of the same file to different MT 
services with different target language tags. But the real test, as with 
everything, is if potential implementer have compelling use cases they 
want to support.

> 2) Is it ok to contemplate changes like this to the basic ITS mechanism, or are they out-of-scope?

It doesn't seem any more at variance with current ITS than some other 
issues we are discussing, such as reference to stand-off meta-data for 
indicating user-defined value sets for variables or for recording 
provenance information in the CMS rather than in the document, or even 
the localsiaiton of deep-web data, e.g. document-level tags (see 
http://www.w3.org/International/multilingualweb/lt/wiki/CMS-level_data_category_scoping). 
So no, I don't such variance from ITS '1.0' as a blocking issue at this 
time.
> 3) If this is something we'd like to include, where do we capture this in the requirements document?

I'll put it in as a non-functional requirements for now.

I think this seems to impact on the design of different data category 
implementations more than data categories themselves.

cheers,
Dave

> Cheers,
> -yves
>
>
>
>
Received on Sunday, 22 April 2012 21:08:12 UTC

This archive was generated by hypermail 2.3.1 : Sunday, 9 June 2013 00:24:55 UTC