Re: Mad idea: Programming language based on RDF

Hi Andreas and Joshua,
thanks for the pointers.
I was aware of both your works (I presented the first time my work at the
same workshop in which Steffen Stadtmüller was introducing Linked Data-Fu
and I certainly cited Ripple).
My dissertation has been actually 2.5 years ago and I moved to something
else, but I am still interested in the topic (thanks for the good luck
anyway, it is definitely always useful...).

And I think the broader problem of delivering usable tools for programming
the semantic web is still open and relevant.
The ongoing work on Linked Data-Fu is very interesting and it goes in the
direction of having a working platform to execute rule-based RDF programs.
At the same time, my opinion is that designing a usable declarative RDF
language is an effort orthogonal to a system like Linked Data-Fu, as
further aspects as modularity and scope of data have to be considered.

Kind regards,
Miguel

On Thu, 9 Nov 2017 at 15:58 Andreas Harth <andreas@harth.org> wrote:

> Hi Miguel,
>
> similar to your approach, we use SPARQL and recursion via rules expressed
> in Notation3 syntax.
>
> But instead of a dataflow approach we use a sense-act cycle.
> For the sense part, users can specify condition-request rules (with HTTP
> GET request templates in the rule head) to fetch data.
> For the act part, rule conditions are evaluated over the entire integrated
> data to trigger HTTP PUT requests for updating resource state.
> The sense-act cycle runs in a loop, which allows for dealing with evolving
> data.
>
> We have an implementation with our system called Linked Data-Fu [1].
>
> Cheers,
> Andreas.
>
> [1] https://linked-data-fu.github.io/
>
> On 11/09/17 15:21, Miguel wrote:
> > I've been working on a RDF-based programming language during my PhD.
> >
> > My approach [1,2,3] was a dataflow language (having both an RDF-based
> and a visual representation), in which the operators are defined through
> SPARQL queries.
> > By permitting recursion (cyclic pipelines) you can have turing
> completeness.
> > By having as inputs both static RDF graphs, temporal RDF graphs, and RDF
> streams, you can have responsive applications, in a flavour similar to
> continuous queries on temporal databases.
> > Furthermore, by employing a notion of (local, in my case) state you can
> represent interactive applications.
> >
> > In that work I used an RDF-based representation of SPARQL, namely
> SPIN-SPARQL [4], that is based on the abstract syntax tree.
> > Later I decided to explore an RDF representation of SPARQL algebra, more
> useful for query rewriting [5].
> >
> > In the related work of [2,3] you can find references to previous related
> work, including some other RDF-based programming languages.
> >
> > Best regards,
> > Miguel
> >
> > [1] http://swows.org/
> > [2] https://doi.org/10.1016/j.jvlc.2014.10.027
> > [3] https://doi.org/10.1007/978-3-319-05693-7_7
> > [4] http://spinrdf.org/sp.html
> > [5] http://ceur-ws.org/Vol-1644/paper12.pdf
> >
> >
> > On Thu, 9 Nov 2017 at 11:03 Sebastian Samaruga <ssamarug@gmail.com
> <mailto:ssamarug@gmail.com>> wrote:
> >
> >     It's not a programming language but the documentation (early drafts)
> for building a framework which relies on RDF for modelling system behavior.
> >
> >
> https://github.com/CognescentBI/BISemantics/blob/master/Document.pdf?raw=true
> >
> >     https://github.com/CognescentBI/BISemantics
> >
> >     Best,
> >     Sebastián Samaruga
> >     ---
> >
> http://exampledotorg.blogspot.com.ar/2017/09/hi-everyone-im-sebastian-software.html
> >
> >     On Nov 9, 2017 7:01 AM, "Martynas Jusevičius" <
> martynas@atomgraph.com <mailto:martynas@atomgraph.com>> wrote:
> >
> >         Might have been done already: https://github.com/trith/trith
> >
> >         On Thu, Nov 9, 2017 at 9:32 AM, Peter Brooks <
> peter.h.m.brooks@gmail.com <mailto:peter.h.m.brooks@gmail.com>> wrote:
> >
> >             dot -> svg does the trick, going from graph to executable
> code in one step.
> >
> >             http://www.graphviz.org/doc/info/output.html
> >
> >             On 9 November 2017 at 04:50, Victor Porton <porton@narod.ru
> <mailto:porton@narod.ru>> wrote:
> >              > Just a few seconds ago I had a mad idea:
> >              >
> >              > Make a programming language based on RDF rather than on
> plain text.
> >              >
> >              > Well, this would require many (...) lists to specify the
> order of
> >              > execution.
> >              >
> >              > What do you think?
> >              >
> >
> >
> >
> >             --
> >             Peter Brooks
> >
> >
> >             Mobile: +27 82 717 6404 <+27%2082%20717%206404>
> <tel:%2B27%2082%20717%206404>
> >             Direct: +27 21 447 9752 <+27%2021%20447%209752>
> <tel:%2B27%2021%20447%209752>
> >             Skype:  Fustbariclation
> >             Twitter: Fustbariclation
> >             Google+: Fustbariclation
> >             Author Page: amazon.com/author/peter_brooks <
> http://amazon.com/author/peter_brooks>
> >
> >
>
>
>

Received on Thursday, 9 November 2017 17:22:32 UTC