W3C home > Mailing lists > Public > public-sweo-ig@w3.org > November 2006

Re: Collection of Comments re. RDF/XML

From: Kingsley Idehen <kidehen@openlinksw.com>
Date: Thu, 30 Nov 2006 15:46:26 -0500
Message-ID: <456F4322.70705@openlinksw.com>
CC: public-sweo-ig@w3.org, Ivan Herman <ivan@w3.org>

Ivan Herman wrote:
> Kingsley Idehen wrote:
>   
>>>>     
>>>>         
>>> About this layercake: there has been a long discussion on the layercake
>>> diagram on the Semantic Web Coordination group some months ago. It was
>>> initiated by the fact that the way the Rules work is in the stack is not
>>> really in line with their lines. After discussions, here are two
>>> versions that we ended up with:
>>>
>>> http://www.w3.org/2006/07/layerCake-2.png
>>> http://www.w3.org/2006/07/layerCake-4.png
>>>
>>> I think the second is closer to the rules people's work and mindset
>>>
>>>  
>>>
>>>       
>>>> I will provide a URI for an enhancement to the current "Stack Cake" that
>>>> attempts to diffuse the aforementioned RDF/XML  issue.
>>>>
>>>>     
>>>>         
>>> As said: this work has already been done, and I think it reflects those
>>> issues as well...
>>>
>>> Ivan
>>>   
>>>       
>> Ivan,
>>
>> Certainly an improvement over the current Cake, but the issue of XML
>> remains. XML should be stacked vertically in a rectangle labeled:
>> "Serialization Format" or "Interchange",  that also contains N3, Turtle,
>> others. This would be in pillar like fashion just  like "crypto". In
>> addition, there should be a bottom layer rectangle for "Data", "Data
>> Sources", or "Data Providers".  That's all I believe needs to be added
>> to the layer cake for now with the goal of separating the RDF Model from
>> the Serialization Formats.
>>
>>     
>
> Kingsley,
>
> that was discussed, as you can imagine, when we ended up with picture.
> While everybody knows the problems around RDF/XML, and that it is a
> serialization only, there was some resistence on putting it completely
> aside. The fact is that RDF/XML has a special role in that it is (for
> the moment) the *only* standardized serialization, which also means that
> it is the only RDF serialization format that ensures interoperability.
> Also, all serialization formats use, in some way or other, some sort of
> namespace mechanism which, though it may not be *exactly* the XML
> namespace specification, it is nevertheless very close to it. In this
> respect, giving it a special position to the XML box is justified.
>   
Ivan,

Understood, but in this matter lies the critical conundrum :-) The 
official Serialization format is a "wedge issue" when it comes down to 
broader comprehension and acceptance of the overall vision -- across 
certain communities (as per my prior commentary). Thus, we have to bring 
N3 and Turtle to the fore during the early stages of the eduction and 
general messaging process, especially as both provide a simple route to 
understanding the elements of a RDF Statement. In addition, Turtle is 
the pattern used in SPARQL, which is another critical piece of the 
"adoption puzzle".

I think the "Stack Cake" revision steers us towards this goal (assuming 
there aren't any additional calls for clarity etc) :-)

Kingsley

> The issue of data sources is a bit different. It may work, but I am a
> bit worried to get the diagram a bit too complex. But let us see.
>
> Cheers
>
> Ivan
>
>   
>> I will send my rendition attempt of the latest cake diagram that
>> reflects my comments above.
>>
>> Kingsley
>>
>>     
Received on Thursday, 30 November 2006 20:46:37 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:17:34 GMT