RDF 2 Wishlist

[Sandro Hawke]

So, what should W3C standardize next in the area of RDF, if anything?
OWL 2 added a bunch of stuff to OWL that users wanted and implementors
were willing to tackle.  Are there things like that around RDF?

My own answer is in a recent blog post:
     http://decentralyze.com/2009/10/30/rdf-2-wishlist/

What's yours?

Two quick caveats:

   * W3C takes backward compatibility very seriously.  If you're
     proposing something that doesn't have a solid migration story,
     please call it something else, something that doesn't look like
     it's taking over from RDF.  Serious proposals should allow
     existing data-consumer and data-producer systems to keep working,
     with only gentle pressure for upgrading as people want to
     interoperate with the new features.

   * While public input (like this) is welcome, and good for laying
     out the options, to actually have a seat at the table in deciding
     what W3C does next, an organization has to join W3C and help pay
     the bills.  See http://www.w3.org/Consortium/membership for
     details.  Argue facts and designs here, but priorities there.

Thanks.  

    -- Sandro (W3C staff contact for RIF, OWL, SPARQL, eGov)

[Danny Ayers]

2009/11/1 Sandro Hawke <sandro@...>:
> So, what should W3C standardize next in the area of RDF, if anything?

Turtle syntax.

> OWL 2 added a bunch of stuff to OWL that users wanted and implementors
> were willing to tackle.  Are there things like that around RDF?
>
> My own answer is in a recent blog post:
>     http://decentralyze.com/2009/10/30/rdf-2-wishlist/
>
> What's yours?
>
> Two quick caveats:
>
>   * W3C takes backward compatibility very seriously.  If you're
>     proposing something that doesn't have a solid migration story,
>     please call it something else, something that doesn't look like
>     it's taking over from RDF.  Serious proposals should allow
>     existing data-consumer and data-producer systems to keep working,
>     with only gentle pressure for upgrading as people want to
>     interoperate with the new features.
>
>   * While public input (like this) is welcome, and good for laying
>     out the options, to actually have a seat at the table in deciding
>     what W3C does next, an organization has to join W3C and help pay
>     the bills.  See http://www.w3.org/Consortium/membership for
>     details.  Argue facts and designs here, but priorities there.
>
> Thanks.
>
>    -- Sandro (W3C staff contact for RIF, OWL, SPARQL, eGov)
>
>

--
http://danny.ayers.name

[adriandwalker]

Hi Sandro,

Now is probably a good time for the next versions of the RDF and SPARQL specs to recognize that most of the world's economy runs on SQL-style closed world negation in its database applications.

Adding closed world negation (and also closed world aggregation) to RDF/SPARQL semantics could be done in an upward compatible manner.  This would in particular reduce the need for ugly workarounds like "filter" -- the many examples of which on the discussion lists do at least illustrate that CW negation is needed.

Including the ability to "close" is really just commonsense -- how can you add up a list of numbers if you cannot find out what the list consists of?

The following is an exercise with negation and aggregation showing how to do the above, that folks can view, run and change using browsers:

   www.reengineeringllc.com/demo_agents/RDFQueryLangComparison1.agent

W3C could also usefully consider adopting the rule language used therein and in [1] as a standard for both syntax and (model theoretic, human comprehensible) semantics.  Is anyone interested in supporting that consideration with a financial contribution to W3C?  If so, we would be glad to help with the technical work.

Just my 3 cents.

                                    -- Adrian

[1] Internet Business Logic
A Wiki and SOA Endpoint for Executable Open Vocabulary English over SQL and RDF
Online at www.reengineeringllc.com    Shared use is free

Adrian Walker
Reengineering

On Sun, Nov 1, 2009 at 1:51 PM, Sandro Hawke <sandro@...> wrote:

So, what should W3C standardize next in the area of RDF, if anything?
OWL 2 added a bunch of stuff to OWL that users wanted and implementors
were willing to tackle.  Are there things like that around RDF?

My own answer is in a recent blog post:
    http://decentralyze.com/2009/10/30/rdf-2-wishlist/

What's yours?

Two quick caveats:

  * W3C takes backward compatibility very seriously.  If you're
    proposing something that doesn't have a solid migration story,
    please call it something else, something that doesn't look like
    it's taking over from RDF.  Serious proposals should allow
    existing data-consumer and data-producer systems to keep working,
    with only gentle pressure for upgrading as people want to
    interoperate with the new features.

  * While public input (like this) is welcome, and good for laying
    out the options, to actually have a seat at the table in deciding
    what W3C does next, an organization has to join W3C and help pay
    the bills.  See http://www.w3.org/Consortium/membership for
    details.  Argue facts and designs here, but priorities there.

Thanks.

   -- Sandro (W3C staff contact for RIF, OWL, SPARQL, eGov)

[Sandro Hawke]

> 2009/11/1 Sandro Hawke <sandro@...>:
> > So, what should W3C standardize next in the area of RDF, if anything?
>
> Turtle syntax.

Yeah...  Any insights into how to handle the costs of having multiple
syntaxes?  Should the expectation be that all RDF consuming software
will handling exactly three syntaxes (RDF/XML, RDFa, and Turtle)?  I
guess many systems already do, and compared to the other two, parsing
Turtle is trivial.

    -- Sandro

[Pius A. Uzamere II]

I think RDF-JSON would be a good thing to get into a W3C recommendation:

http://n2.talis.com/wiki/RDF_JSON_Specification

Not only is JSON nicer to deal with than XML, it's also a much lower impedance mismatch with many of the "schema-less" document-oriented stores that are becoming popular and tend to serialize documents in JSON or JSON-like formats (e.g. CouchDB and MongoDB).

-Pius

On Sun, Nov 1, 2009 at 2:31 PM, Sandro Hawke <sandro@...> wrote:

> 2009/11/1 Sandro Hawke <sandro@...>:
> > So, what should W3C standardize next in the area of RDF, if anything?
>
> Turtle syntax.

Yeah...  Any insights into how to handle the costs of having multiple
syntaxes?  Should the expectation be that all RDF consuming software
will handling exactly three syntaxes (RDF/XML, RDFa, and Turtle)?  I
guess many systems already do, and compared to the other two, parsing
Turtle is trivial.

   -- Sandro

[Danny Ayers]

2009/11/1 Sandro Hawke <sandro@...>:

>> 2009/11/1 Sandro Hawke <sandro@...>:
>> > So, what should W3C standardize next in the area of RDF, if anything?
>>
>> Turtle syntax.
>
> Yeah...  Any insights into how to handle the costs of having multiple
> syntaxes?  Should the expectation be that all RDF consuming software
> will handling exactly three syntaxes (RDF/XML, RDFa, and Turtle)?  I
> guess many systems already do, and compared to the other two, parsing
> Turtle is trivial.

The costs of having multiple syntaxes? How much has RDF/XML cost?

No need for anyone to change their parser design, just make a simple
syntax the first thing people see when they look for RDF.

It's a bit late in the day I know, but I wish I'd have bugged the
SPARQL folks about matching the syntax.

Ch
--
http://danny.ayers.name

[Danny Ayers]

2009/11/1 Pius Uzamere <pius@...>:
> I think RDF-JSON would be a good thing to get into a W3C recommendation:
> http://n2.talis.com/wiki/RDF_JSON_Specification
> Not only is JSON nicer to deal with than XML, it's also a much lower
> impedance mismatch with many of the "schema-less" document-oriented stores
> that are becoming popular and tend to serialize documents in JSON or
> JSON-like formats (e.g. CouchDB and MongoDB).
> -Pius

I can't advocate that particular syntax through conflicts of interest,
but it's a good point. RDF-JSON would be a good idea, the Semantic Web
should be as available on the Web site developer's (& consumers) plate
just like everything else.

Alignment with HTML5 maybe?



--
http://danny.ayers.name

[thomas-3]

Am 01.11.2009 um 18:51 schrieb Sandro Hawke:

> So, what should W3C standardize next in the area of RDF, if anything?
> OWL 2 added a bunch of stuff to OWL that users wanted and  
> implementors were willing to tackle.  Are there things like that  
> around RDF?

A notion of context or scope (or at least named graphs).

I found these slides very convincing:
http://www.slideshare.net/PatHayes/blogic-iswc-2009-invited-talk

Would probably also solve Adrians problem.

Ciao
Thomas

>
> My own answer is in a recent blog post:
>     http://decentralyze.com/2009/10/30/rdf-2-wishlist/
>
> What's yours?
>
> Two quick caveats:
>
>   * W3C takes backward compatibility very seriously.  If you're
>     proposing something that doesn't have a solid migration story,
>     please call it something else, something that doesn't look like
>     it's taking over from RDF.  Serious proposals should allow
>     existing data-consumer and data-producer systems to keep working,
>     with only gentle pressure for upgrading as people want to
>     interoperate with the new features.
>
>   * While public input (like this) is welcome, and good for laying
>     out the options, to actually have a seat at the table in deciding
>     what W3C does next, an organization has to join W3C and help pay
>     the bills.  See http://www.w3.org/Consortium/membership for
>     details.  Argue facts and designs here, but priorities there.
>
> Thanks.
>
>    -- Sandro (W3C staff contact for RIF, OWL, SPARQL, eGov)
>

[Simon Reinhardt]

Hi

Sandro Hawke wrote:

>> 2009/11/1 Sandro Hawke <sandro@...>:
>> > So, what should W3C standardize next in the area of RDF, if anything?
>>
>> Turtle syntax.
>
> Yeah...  Any insights into how to handle the costs of having multiple
> syntaxes?  Should the expectation be that all RDF consuming software
> will handling exactly three syntaxes (RDF/XML, RDFa, and Turtle)?  I
> guess many systems already do, and compared to the other two, parsing
> Turtle is trivial.

If anyone was concerned about the costs of multiple syntaxes then we wouldn't have 3 native OWL 2 syntaxes (plus all RDF forms of it), 2 RIF syntaxes, 2 SPARQL query results formats and possibly multiple presentations of the to-be-defined RDF2RDB mapping language [1].

Turtle is out there and to my knowledge every important RDF library supports it - and OWL API does as well. I support having it as a recommendation - not only to give it the status it deserves but also to finally sort out the media type problems around Turtle and N3. :-)
And picking up work on the Turtle version of the Primer [2] again would be nice as well.

Because of GRDDL you could say that the number of RDF-interpretable formats officially supported by W3C is endless. I'm not sure which formats a conforming RDF tool should be required to parse but I think RDF/XML and Turtle should both be on the list.

Regards,
  Simon


[1] http://www.w3.org/2009/08/rdb2rdf-charter#scope
[2] http://www.w3.org/2007/02/turtle/primer/

[PeterAnsell]

2009/11/2 Simon Reinhardt <simon.reinhardt@...>:
> Hi
>
> Sandro Hawke wrote:
>>>
>>> 2009/11/1 Sandro Hawke <sandro@...>:
>>> > So, what should W3C standardize next in the area of RDF, if anything?
>>>
>>> Turtle syntax.

> Turtle is out there and to my knowledge every important RDF library supports
> it - and OWL API does as well. I support having it as a recommendation - not
> only to give it the status it deserves but also to finally sort out the
> media type problems around Turtle and N3. :-)

To really sort out the issues between Turtle and N3 it would be nice
to also push N3 through the process if it is developed properly at
this stage.

What is the chief reference for N3 right now by the way?

Cheers,

Peter

[Simon Reinhardt]

Peter Ansell wrote:

> 2009/11/2 Simon Reinhardt <simon.reinhardt@...>:
>> Turtle is out there and to my knowledge every important RDF library supports
>> it - and OWL API does as well. I support having it as a recommendation - not
>> only to give it the status it deserves but also to finally sort out the
>> media type problems around Turtle and N3. :-)
>
> To really sort out the issues between Turtle and N3 it would be nice
> to also push N3 through the process if it is developed properly at
> this stage.
>
> What is the chief reference for N3 right now by the way?

That would be <http://www.w3.org/TeamSubmission/n3/> I think.
And for Turtle it's <http://www.w3.org/TeamSubmission/turtle/>.

[Sandro Hawke]

> Hi
>
> Sandro Hawke wrote:
> >> 2009/11/1 Sandro Hawke <sandro@...>:
> >> > So, what should W3C standardize next in the area of RDF, if anything?
> >>
> >> Turtle syntax.
> >
> > Yeah...  Any insights into how to handle the costs of having multiple
> > syntaxes?  Should the expectation be that all RDF consuming software
> > will handling exactly three syntaxes (RDF/XML, RDFa, and Turtle)?  I
> > guess many systems already do, and compared to the other two, parsing
> > Turtle is trivial.
>
> If anyone was concerned about the costs of multiple syntaxes then we wouldn't
> have 3 native OWL 2 syntaxes (plus all RDF forms of it),

Exactly one syntax for OWL is required (the RDF/XML based one).  All the
rest are optional.  I wouldn't publish them on the open web, unless I
was content-negotiating with RDF/XML as well.  But tool makers want to
use them inside systems, and in books and such.

> 2 RIF syntaxes,

The presentation syntaxes are just for people reading the spec and test
cases.  As with OWL, there is exactly one canonical/required syntax (the
XML one).

So, yes, we still have the social cost of multiple syntaxes, but at
least systems gathering W3C-standard data off the open web don't have to
understand a zillion syntaxes.

> 2 SPARQL query results formats and possibly multiple presentations of
> the to-be- defined RDF2RDB mapping language [1].

I don't think those features increase the cost of implementing data
consumers.

> Turtle is out there and to my knowledge every important RDF library supports
> it - and OWL API does as well. I support having it as a recommendation - not
> only to give it the status it deserves but also to finally sort out the media
>  type problems around Turtle and N3. :-)
> And picking up work on the Turtle version of the Primer [2] again would be ni
> ce as well.

Yep, that sounds good to me as well.

> Because of GRDDL you could say that the number of RDF-interpretable formats o
> fficially supported by W3C is endless. I'm not sure which formats a conformin
> g RDF tool should be required to parse but I think RDF/XML and Turtle should
> both be on the list.

True.   GRDDL does make things interesting.

Thanks for your input....

   -- Sandro

[Danny Ayers]

sorry Sandro - it just occurred to me that the thing syntax-wise is
needed is really Turtle + named graphs - well below cwm stuff, but
maybe get the recipe for formulae from there


2009/11/1 Sandro Hawke <sandro@...>:

>> Hi
>>
>> Sandro Hawke wrote:
>> >> 2009/11/1 Sandro Hawke <sandro@...>:
>> >> > So, what should W3C standardize next in the area of RDF, if anything?
>> >>
>> >> Turtle syntax.
>> >
>> > Yeah...  Any insights into how to handle the costs of having multiple
>> > syntaxes?  Should the expectation be that all RDF consuming software
>> > will handling exactly three syntaxes (RDF/XML, RDFa, and Turtle)?  I
>> > guess many systems already do, and compared to the other two, parsing
>> > Turtle is trivial.
>>
>> If anyone was concerned about the costs of multiple syntaxes then we wouldn't
>> have 3 native OWL 2 syntaxes (plus all RDF forms of it),
>
> Exactly one syntax for OWL is required (the RDF/XML based one).  All the
> rest are optional.  I wouldn't publish them on the open web, unless I
> was content-negotiating with RDF/XML as well.  But tool makers want to
> use them inside systems, and in books and such.
>
>> 2 RIF syntaxes,
>
> The presentation syntaxes are just for people reading the spec and test
> cases.  As with OWL, there is exactly one canonical/required syntax (the
> XML one).
>
> So, yes, we still have the social cost of multiple syntaxes, but at
> least systems gathering W3C-standard data off the open web don't have to
> understand a zillion syntaxes.
>
>> 2 SPARQL query results formats and possibly multiple presentations of
>> the to-be- defined RDF2RDB mapping language [1].
>
> I don't think those features increase the cost of implementing data
> consumers.
>
>> Turtle is out there and to my knowledge every important RDF library supports
>> it - and OWL API does as well. I support having it as a recommendation - not
>> only to give it the status it deserves but also to finally sort out the media
>>  type problems around Turtle and N3. :-)
>> And picking up work on the Turtle version of the Primer [2] again would be ni
>> ce as well.
>
> Yep, that sounds good to me as well.
>
>> Because of GRDDL you could say that the number of RDF-interpretable formats o
>> fficially supported by W3C is endless. I'm not sure which formats a conformin
>> g RDF tool should be required to parse but I think RDF/XML and Turtle should
>> both be on the list.
>
> True.   GRDDL does make things interesting.
>
> Thanks for your input....
>
>   -- Sandro
>

--
http://danny.ayers.name

[Danny Ayers]

nonono

SPARQL include stuff - that's how you get graphs.

2009/11/1 Danny Ayers <danny.ayers@...>:

> sorry Sandro - it just occurred to me that the thing syntax-wise is
> needed is really Turtle + named graphs - well below cwm stuff, but
> maybe get the recipe for formulae from there
>
>
> 2009/11/1 Sandro Hawke <sandro@...>:
>>> Hi
>>>
>>> Sandro Hawke wrote:
>>> >> 2009/11/1 Sandro Hawke <sandro@...>:
>>> >> > So, what should W3C standardize next in the area of RDF, if anything?
>>> >>
>>> >> Turtle syntax.
>>> >
>>> > Yeah...  Any insights into how to handle the costs of having multiple
>>> > syntaxes?  Should the expectation be that all RDF consuming software
>>> > will handling exactly three syntaxes (RDF/XML, RDFa, and Turtle)?  I
>>> > guess many systems already do, and compared to the other two, parsing
>>> > Turtle is trivial.
>>>
>>> If anyone was concerned about the costs of multiple syntaxes then we wouldn't
>>> have 3 native OWL 2 syntaxes (plus all RDF forms of it),
>>
>> Exactly one syntax for OWL is required (the RDF/XML based one).  All the
>> rest are optional.  I wouldn't publish them on the open web, unless I
>> was content-negotiating with RDF/XML as well.  But tool makers want to
>> use them inside systems, and in books and such.
>>
>>> 2 RIF syntaxes,
>>
>> The presentation syntaxes are just for people reading the spec and test
>> cases.  As with OWL, there is exactly one canonical/required syntax (the
>> XML one).
>>
>> So, yes, we still have the social cost of multiple syntaxes, but at
>> least systems gathering W3C-standard data off the open web don't have to
>> understand a zillion syntaxes.
>>
>>> 2 SPARQL query results formats and possibly multiple presentations of
>>> the to-be- defined RDF2RDB mapping language [1].
>>
>> I don't think those features increase the cost of implementing data
>> consumers.
>>
>>> Turtle is out there and to my knowledge every important RDF library supports
>>> it - and OWL API does as well. I support having it as a recommendation - not
>>> only to give it the status it deserves but also to finally sort out the media
>>>  type problems around Turtle and N3. :-)
>>> And picking up work on the Turtle version of the Primer [2] again would be ni
>>> ce as well.
>>
>> Yep, that sounds good to me as well.
>>
>>> Because of GRDDL you could say that the number of RDF-interpretable formats o
>>> fficially supported by W3C is endless. I'm not sure which formats a conformin
>>> g RDF tool should be required to parse but I think RDF/XML and Turtle should
>>> both be on the list.
>>
>> True.   GRDDL does make things interesting.
>>
>> Thanks for your input....
>>
>>   -- Sandro
>>
>
>
>
> --
> http://danny.ayers.name
>

--
http://danny.ayers.name

[Dan Brickley-2]

On Sun, Nov 1, 2009 at 9:35 PM, Peter Ansell <ansell.peter@...> wrote:

> 2009/11/2 Simon Reinhardt <simon.reinhardt@...>:
>> Hi
>>
>> Sandro Hawke wrote:
>>>>
>>>> 2009/11/1 Sandro Hawke <sandro@...>:
>>>> > So, what should W3C standardize next in the area of RDF, if anything?
>>>>
>>>> Turtle syntax.
>
>> Turtle is out there and to my knowledge every important RDF library supports
>> it - and OWL API does as well. I support having it as a recommendation - not
>> only to give it the status it deserves but also to finally sort out the
>> media type problems around Turtle and N3. :-)
>
> To really sort out the issues between Turtle and N3 it would be nice
> to also push N3 through the process if it is developed properly at
> this stage.

Don't forget Turtle/N3 and SPARQL patterns, or N3 'and/or/vs' SPARQL
Named Graphs...

It'd be good to see the named graph concepts brought into the RDF/OWL
mainstream, rather than left in the query language. So Pat Hayes'
recent paper was encouraging there...

Dan

[Lee Feigenbaum-2]

Danny Ayers wrote:
> sorry Sandro - it just occurred to me that the thing syntax-wise is
> needed is really Turtle + named graphs - well below cwm stuff, but
> maybe get the recipe for formulae from there

Yes, a la TriG (http://www4.wiwiss.fu-berlin.de/bizer/TriG/), which
would be my number 1 choice if W3C were to bless a specification of
another RDF(-related) syntax.

But really, I'm not so sure the lack of blessing on these various
alternative syntaxes has been too big an interoperability headache....
has it?

Lee

> 2009/11/1 Sandro Hawke <sandro@...>:
>>> Hi
>>>
>>> Sandro Hawke wrote:
>>>>> 2009/11/1 Sandro Hawke <sandro@...>:
>>>>>> So, what should W3C standardize next in the area of RDF, if anything?
>>>>> Turtle syntax.
>>>> Yeah...  Any insights into how to handle the costs of having multiple
>>>> syntaxes?  Should the expectation be that all RDF consuming software
>>>> will handling exactly three syntaxes (RDF/XML, RDFa, and Turtle)?  I
>>>> guess many systems already do, and compared to the other two, parsing
>>>> Turtle is trivial.
>>> If anyone was concerned about the costs of multiple syntaxes then we wouldn't
>>> have 3 native OWL 2 syntaxes (plus all RDF forms of it),
>> Exactly one syntax for OWL is required (the RDF/XML based one).  All the
>> rest are optional.  I wouldn't publish them on the open web, unless I
>> was content-negotiating with RDF/XML as well.  But tool makers want to
>> use them inside systems, and in books and such.
>>
>>> 2 RIF syntaxes,
>> The presentation syntaxes are just for people reading the spec and test
>> cases.  As with OWL, there is exactly one canonical/required syntax (the
>> XML one).
>>
>> So, yes, we still have the social cost of multiple syntaxes, but at
>> least systems gathering W3C-standard data off the open web don't have to
>> understand a zillion syntaxes.
>>
>>> 2 SPARQL query results formats and possibly multiple presentations of
>>> the to-be- defined RDF2RDB mapping language [1].
>> I don't think those features increase the cost of implementing data
>> consumers.
>>
>>> Turtle is out there and to my knowledge every important RDF library supports
>>> it - and OWL API does as well. I support having it as a recommendation - not
>>> only to give it the status it deserves but also to finally sort out the media
>>>  type problems around Turtle and N3. :-)
>>> And picking up work on the Turtle version of the Primer [2] again would be ni
>>> ce as well.
>> Yep, that sounds good to me as well.
>>
>>> Because of GRDDL you could say that the number of RDF-interpretable formats o
>>> fficially supported by W3C is endless. I'm not sure which formats a conformin
>>> g RDF tool should be required to parse but I think RDF/XML and Turtle should
>>> both be on the list.
>> True.   GRDDL does make things interesting.
>>
>> Thanks for your input....
>>
>>   -- Sandro
>>
>
>
>

[Danny Ayers]

2009/11/2 Lee Feigenbaum <lee@...>:

> But really, I'm not so sure the lack of blessing on these various
> alternative syntaxes has been too big an interoperability headache.... has
> it?

In the wild the most complaints I've heard, and the biggest
misunderstandings, have been through RDF/XML syntax. It works, and it
can be understood (and danbri has a nice history doc somewhere) but
it's horrible compared to most serialisations of stuff.

I'm not a huge fan of e.g. JSON (because namespaces/URIs aren't built
in) but the rest of the world is.

The tasty bits of RDF aren't immediately accessible through an XML parser.

(you know damn well I've been an RDF/XML advocate for years, position
shifted m'fraid)


Danny.

--
http://danny.ayers.name

[Paul Gearon]

On Sun, Nov 1, 2009 at 9:24 PM, Lee Feigenbaum <lee@...> wrote:
> Danny Ayers wrote:
>>
>> sorry Sandro - it just occurred to me that the thing syntax-wise is
>> needed is really Turtle + named graphs - well below cwm stuff, but
>> maybe get the recipe for formulae from there
>
> Yes, a la TriG (http://www4.wiwiss.fu-berlin.de/bizer/TriG/), which would be
> my number 1 choice if W3C were to bless a specification of another
> RDF(-related) syntax.

Named graphs would have to be the biggest on my list. After all,
practically everyone works with named graphs, and they're an important
part of SPARQL, but they're not mentioned in RDF anywhere. Pat's
presentation also showed some interesting possibilities if you accept
assertions (and negative assertions) about named graphs.

Then if you have named graphs, then a document format that allows
multiple named graphs would be really useful as well (e.g. TriG)

While less obvious, one thing I'd love to see is literal subjects.
It's my understanding that the main reason they were disallowed was
because of the inability to express them in RDF/XML, but with N3
(along with all these endorsements of Turtle), then I don't see that
as an issue. Personally, I hate having to create extra nodes that
represent literal values, just so I can relate literals together (e.g.
think about saying that '3'^^xsd:integer is the square root of
'9'^^xsd:integer. There are more interesting relationships between
numbers that are worth storing, but this one illustrates the point)

I'd also love to see blank predicates. I know that not everyone likes
blank nodes, but they are very useful in some circumstances, and have
just as much meaning for a predicate as they do for a subject.

Regards,
Paul Gearon

[Karl Dubost-3]

Le 1 nov. 2009 à 12:51, Sandro Hawke a écrit :
> So, what should W3C standardize next in the area of RDF, if anything?
> OWL 2 added a bunch of stuff to OWL that users wanted and implementors
> were willing to tackle.  Are there things like that around RDF?


not really RDF2 but super needed.

* Turtle (for describing syntaxes)
* RDF Cookbook
   be on http://stackoverflow.com/
   or on http://code.activestate.com/recipes/

[Lee Feigenbaum-2]

Danny Ayers wrote:

> 2009/11/2 Lee Feigenbaum <lee@...>:
>
>> But really, I'm not so sure the lack of blessing on these various
>> alternative syntaxes has been too big an interoperability headache.... has
>> it?
>
> In the wild the most complaints I've heard, and the biggest
> misunderstandings, have been through RDF/XML syntax. It works, and it
> can be understood (and danbri has a nice history doc somewhere) but
> it's horrible compared to most serialisations of stuff.

Right, which is why those alternative syntaxes exist and have many
interoperable implementations.

Are there a lot of people/systems out there that refuse to do anything
but RDF/XML because it's the only serialization that is a W3C
Recommendation?

Lee

> I'm not a huge fan of e.g. JSON (because namespaces/URIs aren't built
> in) but the rest of the world is.
>
> The tasty bits of RDF aren't immediately accessible through an XML parser.
>
> (you know damn well I've been an RDF/XML advocate for years, position
> shifted m'fraid)
>
>
> Danny.
>