Article: SW = ( SM + SC )

It turns out that my latest SW "article" is also short enough to put
in a message, so while I've published it to the Web at:-

     http://infomesh.net/swnotes/swnotions/

I've also published it here, for the lazy amongst us :-)

o o o

As the SW heads off into the unknown, there certain processes and
theories starting to take place that integrate the pragmatic and
abstract approaches to the Semantic Web. This is an introduction to
some examples of this in the SW community.

Semantic Memories, Semantic Clouds
It is clear that there are going to be "two" parts to the SW. A
personal space, and a global space. There may also be
semi-personal/global spaces that only a certain number of people have
control over.

The interactions between these spaces is something that hasn't been
investigated all that much yet, which is sad because that's probably
the most important part of infrastructure as a whole.

Examples of Semantic Memories:-

   Annotations Servers
   Confidential HTTP server space
   Personal databases

Each of these has one thing in common: access by a limited number of
people. The global space will consist of both "free for all" code
areas, but also of other peoples Semantic Memories (read only, this
time). That's quite interesting because it means that people will be
able to swap SMs at will. For example, the annotations held on my
annotations server may be viewable to anyone, but only I have write
permission to it.

Format Wars
There aren't going to be any format wars: everyone will choose what
format they want to work in. However, the thing that all of these
formats will have in common is "triples" - each SW format will consist
of a means of representing triples. Therefore, it should be possible
to convert from any SW format into flat triples, and from those
triples into any format of your choice.

For example, say your information is stored as Notation3. There should
be Triples => N3 converters, and N3 => Triples converters. You use an
N3 => Triples convertor, and then someone else uses a triples to RDF
convertor to get it into the format that they require. That is an
important step to the SW, and I expect that in the near future there
will be "best format convertor" wars. This is a good thing, because it
will stoke on the SW.

Conversion... Semantic Magic
Equivalence and transformation are the most "magical" properties of
the Semantic Web. The power to say that one URI is equivalent to
another within a certain procedural context is very great indeed.

Equivalence leads to transformations. When you say that two things are
equivalent, say "p1" in schema1, and p2 in schema2, then you can
convert from one to t'other, and vice versa.

There are always debates as to where this information should be
stored... well it's either going to be in the schemas themselves, or
third party indexing databases. The SWAG Database is a good example of
this.

From Conversions to Logic
I actually don't expect the SW to go in the directions that everyobody
seems to think. Logic rules are too complicated for people to type
into search engines - e.g. looking for "the brightest star" on an SW
engine isn't going to yield the result that you want. In short, that
is not how the SW is going to work... it is not a good representation
of what it should be able to do, and it clouds the entire SW issue
when people make assertions of that nature.

Small Conversions, Large Conversions
What we are going to see is local rules, as well as rules between
large systems.

Large Scale
Rules between large systems (remote systems) have often been
discussed. For example, you might have two multinational corporations
each creating their own invoice langauges. I use the invoice langauge
of company A, and my friend uses those of company B. Using Semantic
Assertions that we can find on the SW, we manage to convert from one
to the other.

Small Scale
Logic on a small scale is much less widely discussed. For example,
let's consider two friends creating address books. If they agree to
the terms that they use, they increase the power of their address
books from the very beginning, because you don't have to use logic to
convert from one to the other. Local logic, therefore, will be of the
nature of making assertions about what you want to make assertions
about. In other words, the people creating address books might want to
use logic to state that if one of them was vising the other on a
certain day, then the other was visited on that day. There are
thousands of things that "sharing" data in a local sense can do, and
that is a very powerful demonstration of the SW.

Scaling Up
The next thing that might happen is small groups interacting with
other small groups. For example, the terms that my friends and I agree
to use might not be the same as another group of people uses. However,
using "large logic", we can convert from one to the other, and still
preserve the "small logic" assertions that go on inside each group.
Think of this an intra and inter SW conversions.

This is all just a part of the myriad SW; the ways in which it is
evolving are very fascinating indeed. I think that a lot of the
primary information contained on the SW in the early days will be
about the SW itself, much like the WWW contained a lot of information
about the WWW in the early days.

Intercreativity
I also hope that the SW will go towards solving the intercreativity
problems that the WWW has suffered in the past ten years or so.
Because it will be so easy to create and share information on the SW,
I think that it will go some way towards solving the problem.

Overview
Overall then, I think that we must not lose sight of some of the
essential properties of the SW that are being waved under our noses.
While we must remain pragmatic, we must not lose sight of why we are
building the SW.
o o o


Thanks to Seth and Joshua for the inspiration.

--
Kindest Regards,
Sean B. Palmer
@prefix : <http://webns.net/roughterms/> .
[ :name "Sean B. Palmer" ] :hasHomepage <http://infomesh.net/sbp/> .

Received on Thursday, 22 February 2001 11:34:27 UTC