Re: [office-metadata] RDFa model and xml:id

[Svante Schubert <Svante.Schubert@Sun.COM>]

Elias Torres wrote:
> Svante.Schubert@Sun.COM wrote on 12/13/2006 10:00:08 PM:
>
>   
>>> Unfortunately, I can't say the same about your proposal because you
>>>       
> have
>   
>>> not explicitly stated on the mailing list or wiki.
>>>       
>> Correct. I have not the advantage to skip the design and take an
>> implementation out of the hat.
>>     
>
> I am pretty sure you fully understand that the above paragraph is a gross
> mischaracterisation and directedly aimed as an offensive statement to how I
> design and develop solutions and software in my everyday job. 
Hello Elias,

It was certainly not my intention to offend you and I beg your pardon if 
I should have done by mistake.
I tried to state that you have already two papers about the grammar of 
RDFa, by this you are far ahead in process, it seems even beyond the 
design phase.
In the SC we had not been discussing basic design decisions to come 
closer to a solution, I had the impression that your approach was to set 
RDFa as the the default solution and we might adapt it or take it as it is.


[...]
>> Aside of this RDFa interleaves datatypes with the content, which is
>> quite the opposite of being modular.
>>     
>
> How does interleaving datatypes with content makes it NOT modular? Could
> you show me an example? I think you are doing the same thing that you
> claimed Bruce was doing by not showing you an example.
>
>   
What does modular mean, if not encapsulating functionality, having 
components with little dependencies and high cohesion. Interleaving 
brings dependencies.
In our special case there is a variety of metadata in the world for the 
same semantic. If you exchange / transform such semantic equivalent 
metadata, you have less risks, if you touch only the metadata package 
and not the content.

>> By picking CSS, yes the reason for separating the CSS styles from the
>> content - to be able to handle different things separately - should be
>> the same reason for us to separate Styles and Semantic.
>>     
>
> Are you missing the point that CSS can be embedded within the same file?
> Are you forgetting that styles are stored as attributes in both @class and
> @style within the content of the web page?
>   
Although in the same file, there is modularity, as you don't have to 
change any longer the html:body if you want to change style properties.
>   
>> How would you describe your way to decide if something is a good design
>> or a bad design?
>>     
>
> By testing its ability to address the requirements and use cases. Also by
> testing how easy is to implement, understand and extend.
>   
Correct.

Best regards,
Svante