W3C home > Mailing lists > Public > www-rdf-interest@w3.org > May 2002

Re: The Standards Manifesto

From: Aaron Swartz <me@aaronsw.com>
Date: Wed, 22 May 2002 16:19:16 -0500
Cc: "Simon St.Laurent" <simonstl@simonstl.com>, <www-talk@w3.org>, <www-rdf-interest@w3.org>
To: "Tim Berners-Lee" <timbl@w3.org>
Message-Id: <928362F6-6DC9-11D6-A14A-0003936780B2@aaronsw.com>
On Wednesday, May 22, 2002, at 03:34  PM, Tim Berners-Lee wrote:

> Dear Aaron,

Many thanks for your response.

>> The W3C has been taken over by corporations with only selfish interests
>> at heart.
> 1. Corporations are designed by law to be selfish.  Those which are not 
> (except non-profits) betray their shareholders unless they maximize 
> returns, and they can be sued for it.

Indeed, and this is why I fear anything they get involved with.

>> The Web services people swallow resources for an goal antithetical to 
>> Web Architecture.
> What is that to you?  Why do you need those resources?

Perhaps I don't need those resources but it is disappointing for me, as 
a believer in Web Architecture, to see good people spend their time on 
things like that. But true, these things are perhaps tangential to my 
point...I included them because they motivated my action.

>> The RDF people are afraid to do anything worthwhile with the power of 
>> their technology and instead worry for no good reason about 
>> backwards-compatibility.
> Now there, counting myself as an RDF person, I take objection.
> While having to (agonizingly) chose which of the low hanging fruit to
> go for, I feel that I have hacked a path through a lot of it.

Ah, this is true. There are a group of people who are working on 
interesting things in a productive way. It is these people that I hope 
to bring together.

> Working groups are not designed to inovate.  They are designed to solve 
> the problem that, while several people know that it is straightforward 
> to do something, and have proved it, one eneds a common standard fro 
> interoperability.

Ah, this is what I thought. But unfortunately I see the W3C moving more 
and more in the direction of inventing new things rather than 
standardizing old ones. And even when they have an old spec to start 
with (as with SOAP and WebOnt) they can't seem to resist changing it... 
often a lot.

> Innovation, on the other hand, is best done in much smaller groups.
> A thriving supportive comunity of interested people helps innovation, 
> to my
> mind.
>
> So you have to decide what it is that you want to do.

I think I want to do something in between. The question becomes how. As 
you've pointed out in your talks on fractal society, there's a balance 
to be struck. You and I and Sandro and HP can create new specs all day, 
but at some point we need to move to the next level beyond that. We need 
the small community involved here to agree on something.

RDF Query languages are a good example. I overheard several people at 
WWW2002 who felt that it was time to standardize RDF query languages, 
but perhaps some concern about doing it at W3C. A lot of individuals 
have written their own query system, and they understand the subject 
well. The kind of thing I want to see is for those people to come 
together and work something out.

Then, perhaps, they can take it to the W3C. But in our rather small 
community I don't think there's a need.

>> Things I'd like to see developed in this model:
>>   * RDF-Model (a clean version of N-Triples)
>
> Yes, I'd likethis too.   Adobe's XML has "RDF-7" -- remember  the tlk
> at www2002?  That sounds like a good move.

Yeah, for sure.

> You refer above to people being too careful about backward 
> companability.
> Ignoring back-compatability is fine when you are talking about research.
> However, when you are talking about a declared consesnus then it
> can very very unfair.  Adobe, to take an arbitrary example,  have put a
> lot of investment into RDF.  You can't pull out the basis for a lot of
> software unless you check that everyone is happy about the idea.

Yep, which is one of the things that makes me so upset. If we had given 
up on backwards compatibility after M&S came out, or a year after, or 
more, or even when RDF Core was chartered, and tried again we could have 
come up with something clean and Adobe could have adopted it instead of 
RDF/XML. But we didn't.

> I have suggested that the RDF specs be factored so there is a name
> for the basic model without bags and reification.  After all, most 
> things
> like RDFS and DAML+OIL are built on that without the bags.

Good idea.

>>   * RDF-Logic (a FOL system with URIs)
> Contentious.  RDF can hardly be first order,as it allows making 
> statements
> about documents and hence about statements.  Classical logic doesn't 
> work on
> the web.  Logicians can't seem to figure out what they would call what 
> you
> want.  in other words, here isa case you can call research or spend a 
> *very*
> long time in a working group.

I suppose this is more to the research end of the spectrum. But it's 
useful, I think.

>>   * RDF-Query (a standardized API and syntax for managing RDF)
> This looks more like a case for your crack team of 6. After all, there 
> are around 7-8 SQL-like RDF query languages out there. Just a question 
> of designing one.  How many people would you expect would want to have 
> a say in the result? How would you  -- fairly -- chose 6 people to 
> carry on the task of doing the design, in your model?

I'd pick the six people who have designed RDF query languages.

>>   * RDF-Sites (an outline for how RDF fits with HTTP).
> Not sure what you are getting at here.

Basically I want a document that I can point to that says:

Put up RDF versions of all your HTML, made available by Accept: headers 
on abstract URIs and replacing .html with .rdf where applicable.

with more explanations and examples.

> el of HTTP?
> HTTP 1.1 is not simple. Making a formla model of it would be a very good
> idea.

That would be good too.

>> While such groups may use W3C (or any other group's) resources, like
>> web-space and telecon-bridges, it must be clear that they are not bound
>> by W3C process and are not responsible to the membership.
>
> Earth to Aaron... ;-)
>
> Aaron, you're not diplomatic or practical here.
> W3C and openprojects.net and so on certianly exist to promote things 
> like
> this, and the www-* mailing lists and archives and so on are there to 
> help
> people get on and do good things.  But to blithely announce that you are
> going to take the resources and run, and explictly not care about the

Oops, maybe I miswrote. What I meant was that I'd be happy if W3C 
donated Web space for the project, but that I didn't want the project to 
be bound by The Process. Of course, the W3C need not donate Web space or 
anything, I'm not even sure why I mentioned it now...

> feelings or desires of those who forked out either cash or their own 
> time or both,
> would make people think you felt you had the one truth when it came to
> development ideas and so on.

I, of course, don't feel that way at all.

> Anyone has have a choice - to be a small group and not claim consensus, 
> or
> to make a larger group. There is a sliding scale.

Can the W3C claim consensus? Among whom? It's all relative, a sliding 
scale as you say.

> If you want to start a new initiative you have to be clear whether it 
> will be open to all and building consensus, or moving fast in a smaller 
> group. You can't claim to do both.

That's why I'm trying to strike a balance. A group that's open to all 
and trying to get their want included, and a core team that's moving 
fast and may or may not listen. Although it's not intuitive, I think 
that the larger the group, the more people that get snubbed. Anyway, I 
look forward to trying the experiment and I'm still looking for 
collaborators.

> In the mean time, if there is some way in which the support from W3C 
> for the semantic web development could be improved, then let us know, 
> but explian how it is fair, and fast, and high quality -- and funded. 
> Or which of those goals it drops.

Well, I'd like to see W3C SW development improved in many ways, but 
that's another discussion. Making things fair, fast, high-quality and 
funded is a difficult if not impossible goal. w3c-semweb-dev isn't 
exactly fair, operating in memberspace with a membership that it's not 
at all clear how to join. I'm not sure how to decide whether it's fast, 
since I'm not sure what it's supposed to be doing. And it has a more 
hackish style than "high-quality", although I certainly do not fault it 
for that. It is funded, though. ;-)

--
Aaron Swartz [http://www.aaronsw.com]
Received on Wednesday, 22 May 2002 17:19:18 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:54 GMT