rdfapi-php-interest

[Rdfapi-php-interest] Re: SPARQL in RAP

[Benjamin N. <bn...@ap...>]

On 28.04.2005 10:44:08, Chris Bizer wrote:
>
>Hi Benjamin,
>
>Hannes pointed me at your SPARQL implementation yesterday. 
>We are currently looking for a way to get SPARQL into RAP and are =
wondering
>if we could use your parser.
yes, sure. although I'd probably suggest waiting for the next release.
the current one has not been tested in any way other than with some
hand-written test queries. I have already extracted a list of queries
from the sparql WD but haven't found the time yet to test them with
the parser. the embedded turtle parser needs some more work, too.
Although including my stuff into RAP and getting some bug reports
could be helpful to me as well.

>Do you mind if we include your code into RAP? Would be great!
I'd be honoured :)
Another interesting thing could be to try porting my SPARQL-to-SQL
rewriter to the RAP db model. I received some helping comments and
bug reports from Richard Cyganiak, so more work is needed there, too,
but a next version of the rewriter could perhaps be usable already.

>What are your future plans for your code? Do you plan to support all =
SPARQL
>features or only a subset? What is the timeline of your work?  
ouch, timeline questions ;)
The parser currently doesn't parse constraints. That'd be the next
thing to do, especially as I want to use at least PrimaryExpressions
in the rewriter (such as FILTER ?date > 2005-03). I'm aiming at being
able to *parse* all of the SPARQL grammar features, but I'm actually
not sure if I want to try to add support for custom filters in the
rewriter. I'll probably focus on a subset that can be translated to
SQL.
Re timeline: Apart from obvious bug-fixes I'll probably not be
able to continue working on the parser before delivering an initial
version of the overdue semanticweb.org site. It may be that I have
to improve the parser *for* the portal (e.g. to add support for simple
constraints), but otherwise, I fear it won't happen before mid-may..


>I'm looking forward to (hopefully) meeting you in Crete. 
yeah, that would be great. I'm not sure if I can afford the eswc
this year, but I'm still hoping to at least get there for the
scripting workshop...

cheers,
benjamin

--
Benjamin Nowack

Kruppstr. 100
45145 Essen, Germany
http://www.bnode.org/

>
>Chris
>
>
>-- 
>Chris Bizer
>Freie Universit=E4t Berlin
>Phone: +49 30 838 54057
>Mail: ch...@bi...
>Web: www.bizer.de
>

[Rdfapi-php-interest] SPARQL in RAP

[Chris B. <bi...@ze...>]

Hi Benjamin,

>>Hannes pointed me at your SPARQL implementation yesterday.=20
>>We are currently looking for a way to get SPARQL into RAP and are
>>wondering if we could use your parser.
>yes, sure. although I'd probably suggest waiting for the next release.

Great, thanks a lot.

Next release also sounds good, because we first want to get Named Graphs
into RAP and build SPARQL on top afterwards.

>the current one has not been tested in any way other than with some
>hand-written test queries. I have already extracted a list of queries
>from the sparql WD but haven't found the time yet to test them with
>the parser.=20

Did you see the DAWG tests?
http://www.w3.org/2001/sw/DataAccess/tests/

>the embedded turtle parser needs some more work, too.

Is the parser an adoption of existing N3 code or a new implementation?
Can it also be used separately from SPARQL?

>Although including my stuff into RAP and getting some bug reports
>could be helpful to me as well.

I can guarantee lots of bug reports from the RAP users. Even more than =
you
probably want ;-)

>>Do you mind if we include your code into RAP? Would be great!
>I'd be honoured :)
>Another interesting thing could be to try porting my SPARQL-to-SQL
>rewriter to the RAP db model. I received some helping comments and
>bug reports from Richard Cyganiak, so more work is needed there, too,
>but a next version of the rewriter could perhaps be usable already.
>

Yes, this would be interesting. I'm also having the loose plan to look =
for a
student who might port D2RQ to PHP
(http://www.wiwiss.fu-berlin.de/suhl/bizer/d2rq/index.htm). This would =
be
great because then you could query all relational databases of CMS, =
blogs,
wikis or whatever using SPARQL. But this is more a mid-term plan ...

>>What are your future plans for your code? Do you plan to support all
>SPARQL
>>features or only a subset? What is the timeline of your work? =20
>ouch, timeline questions ;)
>The parser currently doesn't parse constraints. That'd be the next
>thing to do, especially as I want to use at least PrimaryExpressions
>in the rewriter (such as FILTER ?date > 2005-03). I'm aiming at being
>able to *parse* all of the SPARQL grammar features, but I'm actually
>not sure if I want to try to add support for custom filters in the
>rewriter. I'll probably focus on a subset that can be translated to
>SQL.
>Re timeline: Apart from obvious bug-fixes I'll probably not be
>able to continue working on the parser before delivering an initial
>version of the overdue semanticweb.org site. It may be that I have
>to improve the parser *for* the portal (e.g. to add support for simple
>constraints), but otherwise, I fear it won't happen before mid-may..
>

Fits perfectly with our timeline. I guess we won't start with SPARQL QL
before June. See:
http://sourceforge.net/mailarchive/message.php?msg_id=3D11593860

>
>>I'm looking forward to (hopefully) meeting you in Crete.=20
>yeah, that would be great. I'm not sure if I can afford the eswc
>this year, but I'm still hoping to at least get there for the
>scripting workshop...
>

Which would of course be even more important than the conference ;-)

Cheers,

Chris

>cheers,
>benjamin

--
Benjamin Nowack

Kruppstr. 100
45145 Essen, Germany
http://www.bnode.org/

>
>Chris
>
>
>--=20
>Chris Bizer
>Freie Universit=E4t Berlin
>Phone: +49 30 838 54057
>Mail: ch...@bi...
>Web: www.bizer.de
>

Re: [Rdfapi-php-interest] SPARQL in RAP

[Benjamin N. <bn...@ap...>]

On 28.04.2005 14:56:14, Chris Bizer wrote:

>Did you see the DAWG tests?
>http://www.w3.org/2001/sw/DataAccess/tests/
ah, great. I heard there were some, but didn't expect
so many already. thanx for the pointer.

>>the embedded turtle parser needs some more work, too.
>
>Is the parser an adoption of existing N3 code or a new implementation?
>Can it also be used separately from SPARQL?
it's a new implementation, integrated in the sparql parser.
It was written at an "I'm sick of parsing" stage and is currently
limited to what I thought would be neccessary for queries. *could*
be possible to separate it from the sparql stuff, though.

>I can guarantee lots of bug reports from the RAP users. Even more than =
you
>probably want ;-)
heh

>I'm also having the loose plan to look for a student who might port
>D2RQ to PHP
>(http://www.wiwiss.fu-berlin.de/suhl/bizer/d2rq/index.htm). This would be
>great because then you could query all relational databases of CMS, =
blogs,
>wikis or whatever using SPARQL. But this is more a mid-term plan ...
yeah, that'd definitely be amazing.


>I guess we won't start with SPARQL QL
>before June. See:
>http://sourceforge.net/mailarchive/message.php?msg_id=3D11593860
cool.

benjamin

--
Benjamin Nowack

Kruppstr. 100
45145 Essen, Germany
http://www.bnode.org/


>Cheers,
>
>Chris
>
>
>-- 
>Chris Bizer
>Freie Universit=E4t Berlin
>Phone: +49 30 838 54057
>Mail: ch...@bi...
>Web: www.bizer.de
>

[Rdfapi-php-interest] Editor support

[Jonathan H. <hen...@ya...>]

The inference engines of RAP do basic inference with a standard RDFS -  
but, for example, SWOOP OWL and Protege RDFS uses a different namespace. 
So when using inference RAP only supports the RDFS and OWL built in. I 
am getting at a more general question here:

RAP has done a lot of work in standards compliance and keeping pace with 
other initiatives like Jenna. But perhaps part of making RAP a most 
useful semantic web infrastructure tool would be to "integrate" with the 
output of  most popular editors. I think that RAP at its core must be an 
angel of standards compliance, so maybe these integration tools aren't 
meant to be part of RAP? But the question is, where does not supporting 
alternative namespaces leave some one like me? (A single developer 
trying to piece together a whole development/deployment environment) 
Would you tell me to switch to POWL?