RE: "industry needs"

Dear All,
 
As you probably know, I'm coming from academic background and this time
my job involves representing a large company's (Betfair) interests in
W3C and in the WG. The company has invested money and resources, it
re-defined some of its strategies to be W3C compliant, at least 900 IT
engineers are ready to implement what is defined PROVIDED it is useful.
Our input says that the industry needs to be able to react to complex
events (see, e.g., http:://complexevents.com"). These can be simple
notification rules that send notifications when certian conditions are
reached or they can be something much more involved. Nobody wants here
to generate imperative code from declarative. Going from BPEL, for
example, one could make procedural code more declarative allowing for
derivation rule to used for decision making instead of an elementary
SWITCH. There are approached that do this and I am ready to give a
presentation in Budva explaining the various flavours of such scenarios
that in my view will be able to interoperate with data (SQL,RDF) and
events. My company requires value from this group, it is a business
decision at the level of our CEO to participate and commit and we have
committed big time.
 
Let us all be constructive and listen to each other and foster mutual
understanding and promote actual business needs of the participants. The
group should be fortunate that it can draw on as much company expertise
instead of chasing somewhat unrealistic use cases. We are preparing to
assist you with all necessary resources to make interesting use cases
more concrete and create new ones that we fell we critically depend on.
The healthy completion we have here should result in better sceince more
balanced solutons and excellent standard.
 
>From the sematics point of view, consider action sequences. They may
include derivation rules, updates, and communication acts. The do not
include FOR, WHILE, IF, SWITCH. The nature of updates is limited to that
in Prolog or Transaction logic. The communiciation acts can be captured
by a Petri net formalism among other things. The virtual machine like
this derives, queries and updates data (SQL/RDF/OWL), sends and receives
messages. I'm more than sure that the group is more than capable in
providing an acceptable semantic level such that it will even be
possible to formally model check the behavioural part, and deal with any
issues if they arise.
 
Thanks
 
Alex Kozlenkov
AC-Rep
www.betfair.com
 

________________________________

From: public-rif-wg-request@w3.org on behalf of Gerd Wagner
Sent: Tue 06/06/2006 10:49
To: 'Sandro Hawke'; 'Francois Bry'
Cc: public-rif-wg@w3.org
Subject: RE: "industry needs"




> > It is very puzzling, I feel, to have to argue about this. I thought,
> > industry needs were better accepted in W3C...
>
> In my experience with the W3C, needs of various communities
> are accepted
> in accordance with how those needs are expressed in the Working Group.
> If a community would benefit from having something in a standard, it
> needs to have people in the Working Group who will participate
> effectively on its behalf.  That means, at least, presenting clear use
> cases and requirements, and helping do the work of developing the
> technical specificaition.

This has been done, e.g. by Gary and by members of REWERSE, but
still others (e.g. you) have tried to argue that RIF doesn't
really need ECA, and that it can all be done with "logical" rules.

Isn't that a bit strange, a SemWeb enthusiast community (around
RDF/N3) and an academic community (around OWL) arguing that the
SemWeb/RIF doesn't need to take industrial technologies (such
as PRs and RRs/ECA) as first class citizens, but rather that
they can do it better their own ("logical") way?

Why don't you suggest this to the W3C DOM Working Group, that
they should turn their XML event listeners (which are ECA
rules) into some form of "logical" rules?

-Gerd



________________________________________________________________________
In order to protect our email recipients, Betfair use SkyScan from
MessageLabs to scan all Incoming and Outgoing mail for viruses.

________________________________________________________________________

Received on Tuesday, 6 June 2006 11:24:46 UTC