Re: [ISSUE-42] tool reference

2012/9/26 "Prof. Dr. Jörg Schütz" <joerg@bioloom.de>

> Nice idea! In addition, what about considering a global repository with
> discoverable tool descriptions through a unique id (aka URI/IRI)?
>

Good point. The proposal discussed in Prague is based on Yves' proposal
below and would accomodate tool identification via a URI - so you could
point to such a registry.

Felix


>
> Cheers -- Jörg
>
>
> On Sep 26, 2012 at 07:04 (CEST), Yves Savourel wrote:
>
>> Hi all,
>>
>> One thought for the issue of referencing tools :
>>
>> We could have a lone its:toolRefs attribute that belong to no data
>> category and can be placed anywhere.
>>
>> It would take one or more composite values. Each one made of:
>> - a data category identifier
>> - the id of a tool description item in the external tools list.
>>
>> e.g: toolRef='datacatId1/toolId1 datacatId2/toolId2'
>>
>> The behavior would be that tool-id for a given data category is
>> overridden only if the attribute specifies a new tool-id for that data
>> category.
>>
>> For example: if we have a first instance its:toolRefs='mtConfidence/t-**1'
>> in the document
>> And then one of its children elements has its:toolRefs='disamiguation/t-*
>> *2'
>> The current tool id for mtConfidence would still be t-1
>>
>> This would allow to use toolRefs on several overlapping data categories,
>> without losing information.
>> It's basically like if each data category had its own toolRef attribute.
>>
>> Just a thought.
>> -yves
>>
>>
>


-- 
Felix Sasaki
DFKI / W3C Fellow

Received on Friday, 28 September 2012 08:50:06 UTC