W3C home > Mailing lists > Public > public-lod@w3.org > February 2010

Re: Enterprise level RDF Scripting ( was: [foaf-protocols] foaf classes for primary Topic )

From: Story Henry <henry.story@bblfish.net>
Date: Tue, 2 Feb 2010 00:14:12 +0100
Cc: Linked Data community <public-lod@w3.org>, "Leo Soto M." <leo.soto@gmail.com>, Peter Williams <home_pw@msn.com>, Dan Brickley <danbri@danbri.org>, foaf-protocols@lists.foaf-project.org, Kingsley Idehen <kidehen@openlinksw.com>
Message-Id: <97B7D6DD-6A0C-4AEB-B1C3-9AAA64E07BB8@bblfish.net>
To: Aldo Bucchi <aldo.bucchi@gmail.com>

On 1 Feb 2010, at 15:08, Aldo Bucchi wrote:

> Hi,
> On Mon, Feb 1, 2010 at 9:11 AM, Story Henry <henry.story@bblfish.net> wrote:
>> On 31 Jan 2010, at 17:25, Peter Williams wrote:
>>> Let's build that linq2rdfa driver! Itís the killer app for the semweb, in
>>> Microsoft land.
> Cool!
> This topic is coming up again ;)
>> I agreee. From the Java perspective this is very much what I found too.
>> When I first learned RDF I was really intrigued about how it related to Object Oriented Programming, which I was familiar with. So I tried to write a mapper for it to Java, which to my astonishment was really a lot easier than I thought. This really helped me understand how OO programming and the semantic web mesh together. Having these tools for run of the mill programmers is really important as Peter points out to get them to overcome the fear of the new: by bringing the semweb back to something they know. This thought is what led me to develop the Sommer library:
>>   https://sommer.dev.java.net/sommer/index.html
>> Doin this made it clear that really the major difference is that we name things and fields with URIs.
>> The next difference is that of Graphs, which is a little more difficult to merge correctly into OO languages (or for that matter most traditional programming languages).
>> Still before tools such as Hibernate and such for Java came out, people worked with SQL directly. So people went on a long time with this pain point... To start off we need really good examples of the usage of RDF. And I think foaf+ssl is that key driver, because it has the ability to give people access to things they would not otherwise have had access to: eg: going to a party.
> Henry:
> Here's an idea. Instead of trying to bake this cake by ourselves, can
> you tap into Martin Odersky's team and propose some use cases? I could
> help with that if you get his attention.
> He's the guy behind Scala and they are on a rampage these days.

Yes, I have played a little with Scala. It should be easy to fit in. I think some initial work may even have been done by the Elmo folks. http://openrdf.org/ .

> Now. Even w/o tweaking the compiler or the grammar to support URIs,
> the DSL capabilities of Scala make it possible to do beautiful things:
> http://code.google.com/p/scardf/

yes, that is interesting... Thanks for the pointer.

> That project is a good start but there are still some issues to address.
> I have my own experimental framework but, just as you describe, N1
> issue is native type integration and we should explore this at a lower
> level. If we could just get native URIs / XSD datatypes things would
> be awesome.

yes, and one should not get stuck with xsd data types. I think in fact that is a problem with many frameworks is that they are stuck into not allowing datatpes as subjects for example. 

"123"^^xsd:int .

is very similar to 

[ xsd:int "123" ].

which in N3 can be written.

"123"^xsd:int .

> Then, Number 2 is graph traversal which can be solved by having some
> sort of path language baked in, ( which is compiled to a single query
> underneath instead of being evaluated step by step via API, which is
> the case in scardf ) .

sounds interesting.

> Number 3 is datasource management. Tying up a datasource as an
> implicit value or bound to a thread. Nothing new under the Sun here (
> pun intended ) ;)
> This is where things get interesting as the datasource could actually
> be a full blown Linked Data client such as Virtuoso with Spongers
> turned on.

you mean something like tying to a graph or a virtual graph?

> Finally. My initial look at the Scala grammar suggests that we should
> be able to fit in URIs. There is support for XML literals and URIs are
> easily recognizable as they have a strict syntax. Qnames could also
> happen.

Language support would be great. I want to use Scala more myself. Though I have been in the last year moving towards email as my every day tool. Need to get back to programming a little bit... 

I'd be happy to support an effort such as this. I'll see how much time I have though... It depends how much time I end up spending programming again.


> Regards,
> A
>>        Henry
> -- 
> Aldo Bucchi
> skype:aldo.bucchi
> http://www.univrz.com/
> http://aldobucchi.com/
Received on Monday, 1 February 2010 23:14:51 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:16:03 UTC