Re: Normative vs Informative

On 11/29/11 3:22 AM, Mo McRoberts wrote:
> Last e-mail on this, because there’s a limit to how willing I am to try to constructively engage with ad-hominem-filled threads.
>
> On 29 Nov 2011, at 01:01, Kingsley Idehen wrote:
>
>>>> If WebID was about Henry, I would have kissed it goodbye this weekend.
>>>> And certainly after this post. Unfortunately, as I've stated repeatedly,
>>>> it isn't about Henry or I. It's about solving a serious problem by
>>>> leveraging the in-built architecture of the World Wide Web where RDF is
>>>> an option.
>>> Enough with the “leveraging architecture” waffle.
>>>
>> So to you the following == waffle:
>>
>> 1. identity data object with de-referencable URIs
>> 2. provide useful information in the form of object descriptions via EAV/SPO based graphs
>> 3. reference other things via URIs.
>>
>> Ah! It stops being waffle if I say:
>> 1. identity data object with de-referencable URIs
>> 2. provide useful information in the form of object descriptions via EAV/SPO based graphs
>> 3. Use standards: RDF and SPARQL
>> 4. reference other things via URIs.
> ?????
>
> For the record, I can't stand SPARQL.

For the record I like SPARQL very much. Ditto RDF. I've been able to 
solve serious problems using both, for many years. My liking or 
disliking them has no bearing on what I am trying to get across re. 
"escape velocity" for WebID bootstrap.

>
> No. What was waffle was that paragraph quoted above. It stops being waffle when you start talking about specific technical points, which is perversely what you did when you said “So to you the following == waffle”. Can you not see the difference between the rambling paragraph of ad-hominems and the list of technical facets? Really?
>
>> RDF is the standard for what exactly? Please really think before responding since RDF is not what you might hope it is should you follow its on specs and abstract syntax. As I've told you already, RDF passes (just about) as an option for implementing Linked Data graphs, so conflating RDF and Linked Data is flawed. They are not the same thing. You can make Linked Data if you use RDF in a specific way. The key to Linked Data lies in the behavior of de-referencable URIs. That gem comes from the architecture of the world wide web itself!
> Where in the e-mail that you're replying to did I mention RDF _at all_?

>
>>> PLEASE. Focus on the _technical issues_.
>>>
>> I guess explaining how hyperlinks deliver de-reference and address-of operations at InterWeb scale == waffle, right? Ditto when you intermingle that with a Linked Data structure (which is a skill moderate programmers posses outside the Web realm).
> You're not *explaining* anything.
>
>>>> I am moving on, you'll come to understand my concerns in due course,
>>>> that I am 100% certain about.
>>> Be 100% certain all you like, but I think you’ll find you don’t speak for my interests.
>>>
>> Clearly, and you weren't my target. Remember, you sought to have WebID tightly scoped to RDF/XML.
> That’s… a remarkable mischaracterisation.

In an earlier post you suggested having WebID more specifically scoped 
to RDF/XML. Basically, your comments about RDFa the terminated in 
comments by Manu.

>> You even prefer RDF/XML over HTML based object descriptors.
> Guilty. I recognise that WWW architecture has included multiple content types for almost as long as the WWW has existed, and<link>, Accept:, Vary:, and Link: for longer than most people who use the Web today have been, and that my _preference_ is not to stuff graphs into HTML when you can stuff them into documents whose format is designed for that purpose, often with a great deal less brittleness. My personal preference for PUBLICATION has little bearing upon ensuring that consumers can work sanely — the means by which I wish to publish a WebID profile is already part of the spec; in that sense I’ve “won” (although I haven’t, because I didn’t take part in any decision to do that), so it doesn’t matter to me as a publisher whether other serialisations are there are not. That’s the extent to which my preferences apply in this regard.

Yes, so you are set. What about others? Developers who are coming at 
this whole thing from a totally different angle? All I want is for this 
endeavor to be more accommodating to other developer profiles. Right now 
we have a solution that is basically scoped to XML, bottom line.

>
> As an implementor, and as somebody trying to ensure that WebID is supported by some others, I’m acutely conscious of the need for the spec to be concrete and detail a uniform approach — whether it specifies RDF, non-RDF, CSV, EDIFACT, or whatever.

The WebID narrative needs to be comprised of:

1. conceptual spec -- this explains WebID conceptually without and being 
clouded by implementation details re. data representation formats and 
syntaxes

2. technical spec -- if such only exist for RDF and XHTML+RDFa that's 
fine as long as other specs for other approaches aren't rejected as a 
defensive jerk reaction.
>
>> You just cannot be my target. That said, you aren't representative re. WebID boostrap profile. The target profile I have in mind is the Web 2.0 developer.
> What does that even mean?

I means there are developers that could actually take a look at WebID if 
it didn't come across as being yet another attempt at forcing the RDF 
issue. Many of these developers are already comfortable with:

1. HTML+Microdata
2. JSON -- with JSON-LD emerging a natural representation for those who 
are only going to work with JSON based syntax.

Thus, instead of an XML only syntax (which is what we have now re. 
RDF/XML and XHTML+RDFa) you end up with syntaxes that work for those 
that already work with HTML+Microdata plus those already working with JSON.

>
>>> Nobody asked you to “move on” — in fact, nobody asked you to do anything but drop the rhetoric.
>>>
>>>> BTW -- can I have some links to WebID compliant stuff you've implemented?
>>> If I had links to share I would’ve already — what I have isn’t public (because it’s tied into other applications), and certainly isn’t stable (the latter in part because it’s been waiting on solid specifications). Does that somehow make the points I’m making irrelevant?
>>>
>> By your own standards, somewhat .
> You really haven’t grasped what the issue I have here is, have you? it's nothing — NOTHING — to do with whether you’ve implemented things or not, and everything to do with recounting — at some not inconsiderable length — about how RDF isn't the only option, about how Linked Data and RDF aren't the same thing, about how you're trying to “solve a serious problem” to ensure WebID can reach “InterWeb scales”. Just _do it_.

It's been done. Done a very long time ago btw. So we continue to speak 
past ourselves.

> Provide the patches to the spec, figure out the stuff which needs to be solved to make it work.

Again, its done.

Negotiating data representation is an inbuilt feature of HTTP.

I am exiting at this point as I find myself looping back to issues that 
where covered long before this IG was created.


[SNIP]


-- 

Regards,

Kingsley Idehen	
Founder&  CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen