OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: [xml-dev] RDDL (was RE: [xml-dev] Negotiate Out The Noise)

[ Lists Home | Date Index | Thread Index ]

Paul T wrote:
> ----- Original Message -----
> From: "Jonathan Borden" <jborden@mediaone.net>
> > RDDL can be used as a way to create distributed java applications (via
> > RDDLClassLoader).
> Which makes it a perfect framework to dustribute viruses,
> so certification is unavoidable.

Actually email has shown itself to be the perfect way to distribute viruses
yet at least I have made the choice that using email provides benefits that
outway the risk of viruses.

As everyone knows the java sandbox provide a measure of protection against
viruses, certainly those that exploit buffer overflows and the ilk.

This issue has nothing to do with RDDL _the language_ per se, rather how
RDDL, or any distributed system, might be used.

> As any person who've
> written java applets would know, caching is also unavoidable,
> because one can not afford making a request over the internet
> every time some tag get's processed. And to keep the cache up
> to date one should provide a mechanism for distribution, similiar
> to those used by DNS / FidoNet. And so on.

This is surely true. The common way one directs such activities is during a
URL resolution process. Caching I assume is implemented by the HTTP software
(at least on my system it works this way). This is the _exact_ reason why
resolvable URIs _should_ be used for namespace names, so that one can handle
such details outside the RDDL document (for the specific namespace) and the
RDDL specification itself. Otherwise one wastes his or her time replicating
caching and resolution activities. I am more than happy to let the Web
system handle these details and live with the consequences. Your mileage
apparently does vary.

> I've written about this stuff already. It may be not bad idea to
> re-read my first letters to this thread.

If your prior letters were either understanable or convincing then things
would be different. I find that:

70% of the time I have no idea what you are saying
15% of the time I disagree with you
15% of the time I agree with you

Much of the problem may be related to my assumption that English is not your
first, and perhaps not your primary language. You should realize that some
of what you say seems offensive on first reading. I am giving you the
benefit of the doubt by asking you to explain yourself.

I realize that you must be frustrated with your difficulty in communicating,
and it is not appropriate for this forum to degenerate into a lesson in
English grammar or semantics. Please accept that I am honestly trying to
understand what you are saying, and so you may need to explain yourself in a
different way. My suggestion that you provide code was not an attempt to be
derogatory, rather to allow you to express yourself in a more mathematical,
or formal, manner. My assumption is that you would be at less of a
disadvantage (math is math). As well, my hope is that I, would not be
offended by your code.

> All the ways of possible
> abuse of namespace URL listed by you and not listed by you -
> I thought of them already,  and I'm almost sure that some
> other people, like Clark Evans, Don Park and I think Gavin  -
> they also did.

I consider this _use_ not abuse of namespaces. I have stated this position
many times: The _entire_ reason to create unique names via the URI mechanism
is precisely so that all of the systems and software associated with
resolving URIs may be leveraged. Of course I belief my definition of a
namespace is more useful than a purely syntactic one [1]. This is not in the
RDDL spec because it is the view that I personally hold, and the view that I
am personally promoting. Tim Bray and I are listed as _editors_ NOT
_authors_ of the RDDL specification http://www.rddl.org/ as this document is
supposed to speak in some collective fashion for the XML-DEV list (at least
the activity of the list at _that_ point in time). It is not indended to
reflect the views of each member of the list, but rather some general
consensus of the list as a whole.

I such cases, much is left out because what is put in reflects the extent of
consensus. So be it.

> I also know that this abuse is not ideal thing, because it should
> not be talking about the namespaces at all, but I can live
> with that. It is not important.
> Any  kind of functionality you may think of, you can provide
> *not* abusing the namespace URL. However, I understand
> that because the loophole for abuse is huge, there is a strong
> desire to abuse it before some big company does.

Company or individual specification. FWIW RDDL is not intended to solve the
problems that Topic Maps nor RDF are directed at, nor extended XLinks for
that matter, rather to provide an acceptable, flexible, simple and mutually
agreed on format to serve as a document associated with a namespace URI.
Rather than 'solving' a particular problem, it is intended to _prevent_
problems from occuring (if possible), for example: what will happen if
people start randomly placing text, XML Schemas, RELAXNG, RDF Schemas, WOL
etc etc at namespace URIs? In such cases, my view is that the _general
potential usefullness_ of the namespace URI will vanish or at least be
diminished. Namespaces would only be able to play with 'like' namespaces.
Architectures would get locked into place etc.

> Not that I have anything against 'let's be first to abuse
> namespaces URL' ( and you *already* started the abuse
> with hidden schema extension ) ,

For the record _I_ have not created a hidden schema extension, that was
created, I believe on an experimental basis, by Henry Thompson, as a part of
a software program he developed. What could possibly be wrong with that? (to
be literal: this is not actually a question, rather a construct where I am
making a statement "This is not wrong" under the guise of asking a question.
Imagine my raising both hands in the air and quickly giving them a shake as
I am saying this.)

> As I've said many times ( and right from the beginning ), there are
> just a few *particular* problems with RDDL that are:
> 1. The Process RDDL is developeing. "It is just a documentation,
> with a few URLs, no it has some proprietary hooks into schema
> parsers already, no - we don't actually know, but our document
> is styled with W3C stylesheet and we got some people who
> may like to promote this RDDL word with the articles that
> contradict our own words for the reasons that nobody would
> understand".

 Where is this quote from? _Certainly_ I have never said that RDDL is _just_
documentation (i.e. intended for human consumption) Again, when it is said
that RDDL is _acceptable_ for documentation, this is not making the
assertion that RDDL is _only_ acceptable for documentation.

I could make a more mathematical statement if my English is not clear:

Let RDDL denote the set of documents that validate against the RDDL DTD
Let Documentation denote the set of documents that are acceptable as human

RDDL < Documentation does not imply RDDL = Documentation

> 2. No clear understanding how would somebody gonna *use*
> the resulting framework.

RDDL is a document format that is intended to be used by the _Web_
framework, assumpting all the benefits and problems associated with the Web.

> Namespaces was solving a clear problem : "how can
> we differentiate between tag A used by one company and
> tag A used by another company". I'm not saying that Namespaces
> is a right answer to this problem, but at least the problem
> itself was kinda clear.

That is not my view of Namespaces, see [1,2,3]
> What is the *clear* problem that RDDL is gonna solve?
> There is *no* such.

Your view of the world is limited. So be it.

> Can you correct me? Can you? That would be a great day
> for the Web. That would be the day we'l get the second
> TBL who will lead the Web v 2.0 to it's *full* potential,
> because current web is a huge book, and the next step is
> to turn it into huge 'brain'.

I do not claim that RDDL alone will solve the problem of the Semantic Web.
For the record I do believe that what we call 'knowledge' is best encoded,
at the present time, in first order predicate logic (FOPL), perhaps
description logic (DL). I do _not_ believe this is literally how humans
encode 'knowlege' but if we are talking about 'knowledge and semantics on
the Web, then this is the best formal encoding. Work in this area has been
ongoing for many many decades so I will state my belief.

For the record I don't work in Topic Maps because I currently believe that
there is more of a basis in knowledge representation using FOPL/DL. I am
willing to be convinced otherwise, honestly, but until convinced this is my

On the other hand, I think the use of the term "Resource" is _hopeless_ in
the current Web e.g. RFC 2396 vs. RDF terminology. I do think that the
terminology definition in TM is far better than the current RDF (e.g.
subjects, identifiers). On the other hand Pat Hayes, whom I have great
respect for is currently revising the RDF Model Theory. I expect that this
will provide RDF with the mathematical rigor it needs.

Beyond RDF, I have great _hopes_ for the Web Ontology project, which I
expect will actually bring FOPL/DL to the Web. If RDDL in some way allows
XML Schema/WOL (web ontology language) and RELAXNG to better cooperate, then
I will see this as a success.

> Perhaps we would need to wait for years for the
> *problem* to appear, because, maybe, nobody
> needs the huge brain. If somebody would need,
> that somebody would tell us what is the application
> he want's to get produced.

This is a very limited view. When deciding to have a child, do you decide
what job on an assembly line he or she might have? My own view is that the
'Semantic Web' will greatly assist medical reasearch (e.g. help to cure
cancer). Distributed ontologies, pattern recognition etc have _already_ made
significant contributions to the human genome project (see the broad field
of 'Bioinformatics' for many examples [4]).

> Actually, I don't think you have any answer to the
> only question I've asked you several times:
> "how do you gonna use it?"
> If playing fair, one should just be honest and say :
> "I don't know. I just got excited with the fact that
> I can grab the 'meaning' of some tag  from the internet,
> and maybe some day I'l understand what is the purpose
> of the code I'm writing".
> Do you think that if playing fair you would get a single
> bad word from me? No way. The only thing *I* need
> from XML and the only reason why I'm wasting my
> time with XML is that I think that maybe, some day
> it would be possible to create ... RDDL.
> Whatever. The task is really complex and it can not
> be accomplished in the closed, spy vs spy, buzzword
> producing way, which is a firm style of W3C that
> is now copied by smart people for the reasons I will
> *never* understand.
> It not aways been like that!

This is part of the 15% of what you say that I agree with, and precisely why
I get offended when you explicitly state or merely imply that RDDL has been
created under a closed process! That is outrageous. Nothing in RDDL has been
done in any closed fashion in any way whatsoever, do you want to come to my
house and view my private email? How can this be more clear?

> So the first and the last thing I said - the problems I have with
> RDDL are :
> 1. Process.
> 2. I don't understand who and how is gonna use the framework.
> In result of this thread I got no information that would ensure me
> that:
> 1. Process of developing RDDL is open and sound.
> 2. Maintainers of RDDL have *any* clue on what they
> really gonna do and *why*.

Actually the breakdown is more like:

50% I can't understand what you are saying
15% I disagree
15% I agree
20% Blantantly rude

I expect you to understand these small English words and understand they are
simply rude, so what is your excuse?


[1] http://www.openhealth.org/talks/Extreme%20RDDL.ppt
[2] http://www.openhealth.org/talks/What%20is%20a%20namespace.doc
[3] http://www.openhealth.org/talks/rddl.paper.xml
[4] http://www.openhealth.org/talks/XMLBioInformatics.ppt


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS