On 12/5/11 7:56 PM, Duane Nickull wrote:
> On 11-12-05 3:05 PM, "John F. Sowa"<sowa@xxxxxxxxxxx> wrote:
>
>> But they yielded to political pressure to use XML, which is a fine
>> But when you try to mix the notation for annotating documents.>graph
>> representation with document annotation, pointer following,
>> and a host of other issues, the graph is lost in the mush. When
>> you encode that mush in a bloated syntax, you get a disaster.
>
> DN: I have worked with XML for a lot of years and find that it too is
> misunderstood. I usually agree with you John but I want to note that to
> me, XML is an "expression" of data. (01)
It can be, but in the context of RDF/XML we are talking about an XML
based markup language based on a directed graph model. As you can see,
my last sentence is easily misunderstood. (02)
RDF/XML is markup for creating actual resources that bear eav/spo
triples. These triples represent relations in their most basic form.
Using RDF/XML or other markup languages you can achieve the same goal.
The same applies when increasing the semantic fidelity of these
relations via RDF Schema and OWL. (03)
> This can be metadata, abstract
> models, UML or just about any data imaginable. XML also has nothing to do
> with semantics which confuses a lot of people when they see RDF/XML. (04)
Correct it has nothing to do with the semantics, and there is confusing
literature out there that says things like:
RDF/XML is a directed graph while XML is a tree. Total confusion and
totally inaccurate comparison since (as mentioned above) RDF/XML is
simply XML based markup for a specific purpose. (05)
> Yes -
> XML is verbose/terse and that is part of why so many developers and
> architects like me love XML. (06)
Yes, but for bootstrap at InterWeb scale, it hasn't worked. Today's
developer aren't ready to absorb the cost of parser development
especially in the Open Source era where purchasing commercial quality
grade libraries isn't the preferred options. People really like the
"Free" part of the "Freedom" aspect of Open Source. Basically, they want
the "Free Beer" and that's where XML parser problems start. (07)
> We don't recommend using XML as an
> interchange format all the time (bad). Certainly when we build mobile
> apps for clients we don't recommend using XML-SOAP based web services
> unless no alternative is available.
>
> I think that your assertion that Tim did it for political pressure is also
> somewhat off though. I cannot speak for him but I have worked in the W3C
> before and they encourage re-use of existing standards. (08)
Tim isn't the culprit re. XML. The issue is that the W3C was/is very
strong on XML, that's a fact. In short, Tim had to go create N3 and
Turtle en route to clarifying his thoughts re. Linked Data. He couldn't
pull it off with RDF/XML without syntax oriented inertia. As of today,
you have the following syntaxes for expressed structured data using
directed graphs under the RDF banner: RDF/XML, RDFa, Turtle, N3,
N-Triples, TriX, TriG, NQuads, JSON-LD, RDF/JSON, and others I am sure
I've left out. Look at the footer of:
http://dbpedia.org/resource/Linked_Data (we've even added CXML,
OData/Atom, OData/JSON, CSV to the mix). (09)
> The W3C has
> standardized a number of metadata formats and other data expressions in
> XML. XSLT, SML Schema etc all come from the same basic model (XML
> Infoset). What is notable about RDF is that XML is not tightly coupled to
> RDF. (010)
Yes, but that is still only known to a few. Of course, tides have
started to change since the emergence of DBpedia and the Linked Open
Data project. (011)
> In fact, the XML expression of XML is a separate piece of work and
> only one of the several recommendations in the RDF suite.
>
> http://www.w3.org/TR/2004/REC-rdf-syntax-grammar-20040210/ (012)
Yes, but as stated earlier, this is esoteric knowledge, unfortunately.
>
> If you go to the formal authoritative SME on the suite of RDF
> recommendations, the syntax (XML expression) is clearly decoupled from the
> Semantics of RDF:
>
> http://www.w3.org/standards/techs/rdf#w3c_all
>
> One thing I am curious to know about is whether anyone else has
> implemented RDF in a non-XML interchange format? (013)
Please seem my comments above plus examples. In a way, posing the
question reinforces the problem i.e., you can implement directed graph
based structured data representation using a variety of syntaxes. Put
differently, the basic triple (3-tuple) doesn't belong to RDF solely.
Enhancing the basic triple with URIs is certainly an RDF innovation but
that still doesn't make it uniquely RDF per se., especially when dealing
with folks that are already familiar with the entity-attribute-value model. (014)
> If so, what has been the
> value? I have seen one example of it being rendered in a 2D drawing in
> the RDF Primer itself
> (http://www.w3.org/TR/2004/REC-rdf-primer-20040210/#figure1) but this
> would obviously not be a useable candidate for machine to machine transfer
> of knowledge.
>
> Open Questions:
>
> Is there anything that precludes someone from expressing RDF is a format
> such as JSON? (015)
Of course not, see my example link above. (016)
> If so, has anyone doe such? (017)
See earlier comment.
> What were the reasons for not using XML. (018)
As Guha and Dan said: it was the fashionable thing at the time, hence
RDF/XML. (019)
> Specific numbers on things like
> parse size, readability, document size etc. would be good to know. (020)
That really isn't the killer problem for RDF/XML. The killer problem
is/was visibility of the basic triple pattern. Then after that
normalization challenges arising from blank nodes. (021)
>
> Duane
>
> ________________________________________
>
> Überity.com
> Principal Data Architect&
> Adobe LiveCycle ES Consultant Services
> http://www.uberity.com
> Blog | http://technoracle.blogspot.com
> Twitter | @duanechaos
>
>
>
>
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/
> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
>
> (022)
-- (023)
Regards, (024)
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 (025)
smime.p7s
Description: S/MIME Cryptographic Signature
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (01)
|