W3C home > Mailing lists > Public > semantic-web@w3.org > February 2014

Re: The upcoming user interface "Dizmo" and the Semantic Web, any Ideas?

From: adasal <adam.saltiel@gmail.com>
Date: Thu, 27 Feb 2014 00:26:28 +0000
Message-ID: <CANJ1O4q0h7=O+0diqmjjFOVj=2TjhAYKLMfwDXPDFzo4n8EipA@mail.gmail.com>
To: Urs Holzer <urs@andonyar.com>
Cc: "semantic-web@w3.org" <semantic-web@w3.org>, Melvin Carvalho <melvincarvalho@gmail.com>
Thanks for your reply.
It seems that droplet were swallowed by Oracle and zaplet still exist but
have become specialised in the Governance, Risk and Compliance (GRC) PaaS
space.
It is instructive as both were component based cloud general purpose
offering at inception. i.e. a solution without a problem that then had to
be marketed into a problem space.

Although I have bought into gizmo - it could provide a solution for me -
there is always the problem of definition: what problem does it address?

PIMO & Co.
(Gunnar Aastrand Grimnes, Remzi Celebi, Leo Sauermann: Using linked open
data to bootstrap corporate knowledge management in the OrganiK
project. I-SEMANTICS
2010<http://www.informatik.uni-trier.de/~ley/db/conf/i-semantics/i-semantics2010.html#GrimnesCS10>


You also might want to look up anything to do with Gnowsis and PIMO. e.g.
http://dev.nepomuk.semanticdesktop.org/wiki/PimoOntology )

Nepomuk is a component of the KDE desktop and entails local indexing plus
an ontology for categorisation of discovered entities.

This area is complex. For instance PIMO has failed as utilised by Gnosis.
But there are plenty of solutions in the corporate knowledge surfacing
arena that succeed. All the many search and index solutions. They have the
flexibility to take a statistical approach not just a categorical approach
and can be made (programmed) to fit around the existing problem, rather
than redefining it.

Ontologies being so high level are, in a way, too close to the knowledge
consumer apart from in very well defined cases in highly circumscribed ways.

It seems to me that Dizmo are wanting to operate at a greater level of
generality, and too much ontology use will just create a mess.

However, foaf + WebID for controlled and incremental access rights and ID
management does seem to me to be a cool use case, one begging to be
implemented.
This is because it is a solution that can be deployed across different
architectures, that is in the cloud server/client or federated or some
mixture of this.

That seems to me to be a choice that people should be enabled to make - by
which I mean an attractive (one that may become popular) implementation
will invite different deployment scenarios so long as it doesn't preclude
them through clumsy design.

But Dizmo have to express features that will be popular, it is sort of do
or die.
The further point here is if the cost of implementation is little a lot of
experimentation can go on. It seems this is Gizmo's hope in that they
express the desire to build a developer community - a very common aim with
such projects.
There is always a balance to be struck between ease of programming and
power.
One that recurs in each aspect of the product, for instance how to change
visual appearance being one aspect and horizontal build out options another.

I hope they find a happy compromise and wish them luck.

Adam


On 26 February 2014 22:37, Urs Holzer <urs@andonyar.com> wrote:

> Carvalho wrote:
> > Great idea.
> >
> > What you can do is allow users with a FOAF to have their own personal
> > data spaces, using the pim ontology
> >
> > http://www.w3.org/ns/pim/space
> >
> > So then they can save a bunch of stuff related to themselves or the
> > various applications they are using.  They can share it with other
> > friends etc.  I like the idea of a collection of RSS feeds to see
> > what your contacts are up to.  I sometimes use the following free
> > software package on my desktop
> >
> > http://tt-rss.org/redmine/projects/tt-rss/wiki
> >
> > But it's not been webized yet ... perhaps a good exercise!
>
> This certainly amounts to interesting ideas and usecases. An RSS reader
> is definitely something they need to put into a Dizmo. Also, I will keep
> the Workspace Ontology in mind, I didn't know about it. Thanks for the
> hint.


> I think I will bring up the idea of shared RSS feeds at some point.
> Additionally, sharing calendars would also make sense. I kind of like
> the idea of attaching a calendar dizmo to a FOAF dizmo in order to see
> the person's public calendar entries.
>
> Thanks for your thoughts.
>
>
Received on Thursday, 27 February 2014 00:26:56 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:42:48 UTC