[Smartscribo] Fwd: Re: Smart-M3 development within FRUCT

[dkorzun]

-------- Original Message --------
Subject: Re: Smart-M3 development within FRUCT
Date: Mon, 26 Mar 2012 20:24:02 +0200
From: adelia <adelia at arces.unibo.it>
To: dkorzun <dkorzun at cs.karelia.ru>

Hi Dmitri,

sorry for my long absence, and thanks for the mail, I didn't remember
about the action point. I'll answer in this mail beside one point for
which
I need to check:

First of all we have recently found a little bug in the Smart m3 SIB
hosted in source Forge. It is a line that should be cmmented, is commented
in our version, but not on source forge and implies that protections
doesn't work properly in certain case. I will iform you of the file and
line to modify.

We have worked on a new SIB based on the Redland RDF-Store and we found it
very fast, in particular the resercher who worked at this

substituted the piglet module with freeware software by obtaing

- Faster insert and delete about (2ms)
- SPARQL support
- potential but still not obtained support for RDF-XML syntax in triples
specification.

- WILBUR support Lost
- If information persistency is needed performances are still good, but
not anymore near the 2ms

For this reason we are also working at improving the libraries for C# Java
and also an internal version of python in order to support the SPARQL
query
specification and the response parsing.

Since we worked a lot on this new SIB (of which an half is still our old
SIB) we have made some modification that perhaps is not compatible with
the
old piglet based version. However we have put the latest Piglet based
version on the SOFIA web site

our work has been directed on trying to solve the following main
challenges

1) The dimension of the SIB constantly grows because internal sql lite
tables maintain information that is still not up to date, note: by simply
deleting the old information the SIB becames unstable and is very
dangerous
to leave white lines in the SQL tables, is necessary to put at least
something inside. This work of garbage collectionha been performed in a
good way also if not tested under hard conditions and should be part of
the
SIB release on the SOFIA web site. 

2) Subscription stability: subscriptions make SIB unstable for various
reasons, on the redland SIB currently the life of the SIB process is very
long also with frequent and eherogeneous usage. 

3)I am working on a Java SIB in which is possible to choose different
levels of semantics and is possible to have reasoning support by Pellet,
but is still work in progress.

smart m3 Piglet last http://www.sofia-project.eu/node/364 from what i know
it contains an extension to support the inclusion of plugins as external
.so files(coded by a different group in UNIBO) and the garbage collector
of
which I have spoken previously(coded by a resercher in my team).
last release of redland SIB. http://www.sofia-project.eu/node/363

Perhaps the files can be reached only by personnel with a SOFIA account, I
think you have no problem in finding someone to download it. I currently
have problems in sending code by e-mail, this is the reason why i give you
only the link.

I don't see Luca for a long time, usually he was him to keep us in
contact. We are always working with smart spaces, and we have no problems
in sharing results, instead we have problems in beginning new tasks
because
we are involved in many European projects (3). In the current situation i
hope our results can be useful to FRUCT also if not directly originated by
FRUCT AP because it could be very difficult for us to perform new
activities and at the same time we want to remain connected with the smart
space working group, also if we are not able to prioritize it at moment

Hope I answered to all,

Best regards

Alfredo




On Mon, 26 Mar 2012 19:41:23 +0400, dkorzun <dkorzun at cs.karelia.ru> wrote:
> Hi Alfredo,
> 
> Our AP list from FRUCT SS WG telco still includes
> the following action point for us:
> ===
> Dmitry will discuss with Alfredo by e-mail 
> related to the Smart-M3 source code modifications
> ===
> 
> I sent some emails in January 2012 on this topic.
> Could you briefly summarize what modifications
> in Smart-M3 your group has made. Is it possible 
> to make them available (include to the main 
> Smart-M3 code stream)? Do you have certain plans
> to participate in the Smart-M3 development?
> 
> Thank you,
> Dmitry