Kingsley and Paul, (01)
The major goal of Ontolog Forum is to develop logic-based methods for
representing and using formal ontologies. But those of us who have
been working on the boundaries of R & D have been frustrated by the
difficulty of moving R to D (AKA technology transfer). (02)
The AI field pioneered technology that proved to be valuable for
many kinds of problems. A lot of AI technology has been adopted by
the mainstream, but it's no longer considered AI. Some examples: (03)
1. John McCarthy worked with the Algol 58 project, for which he
suggested two successful features of LISP: recursive functions
and the if-then-else statement (which is syntactic sugar for
McCarthy's conditional statement in LISP). Garbage collection is
another innovation in LISP that Java brought into mainstream IT.
Except for syntax, Python is closer to LISP than it is to Java. (04)
2. Ted Codd adopted AI methods for using first-order logic and
adapted them to relational databases. SQL was a watered-down
version of Codd's methods, but the WHERE-clause still has the
expressive power of FOL. Codd and other IBM researchers
developed relational databases, but IBM was bogged down in
internal politics while Oracle got a big contract from the CIA
to implement RDBs. (05)
3. All versions of rule-based systems use some ideas pioneered by
the work in AI from the 1960s and '70s. Prolog, for example,
is used by Experian to evaluate everybody's credit worthiness.
They use Prolog so heavily that they bought the Prologia company
so that they could have complete control over the features that
are included in their proprietary version of Prolog. (06)
But note that the commercial success stories are produced by people
who understand the requirements of mainstream IT. They're the ones
who take ideas from AI and implement them in successful products. (07)
KI
> we should also acknowledge when the W3C takes positive steps to address
> said concerns. In the case of Schema.org, the collaboration is genuine,
> as you can even see from Guha's own acknowledgement of the work done
>
> 1. http://lists.w3.org/Archives/Public/public-vocabs/2012Nov/0017.html
> 2. http://lists.w3.org/Archives/Public/public-lod/2012Nov/0011.html (08)
Thanks for the pointers to those notes by Guha and Dan Brickley. That
collaboration is very important. But the subset adopted for Schema.org
will become the de facto standard. (09)
PT
> I'm a (sometime) programmer, and I'm enthusiastic about SW specs
> and prospects. Tool support is spotty, but some are excellent. (010)
You could say the same about LISP. The reason why tool support
for the SW is spotty is that mainstream IT ignored the SW --
mostly because the SW ignored the requirements of mainstream IT. (011)
I don't blame Tim Berners-Lee, whose proposal in Feb 2000 was good.
Tim knew how to design a successful system. Unfortunately, he let
a bunch of academics turn it into an AI research project. (012)
PT
> Am I missing something about schema.org that inspires your repeated
> panegyrics? (013)
I'm not praising Schema.org. I'm just noting that Google, Microsoft,
and Yahoo made the obvious decision. (014)
PT
> I work in the aerospace and defense industry. How does schema.org help
> me? How does their "ontologies for dummies" approach point the way to
> improving my industry's IT effectiveness? (015)
Which of your company's applications use the current SW tools?
What ontologies have they developed? Are they mission-critical
applications? Or just interesting demos and proposals? (I'm not
asking for confidential info, just for a brief summary.) (016)
PT
> If you want schema.org to take the future away from W3C Semantic Web,
> they will need to make a better case for it. Their stated goals
> seem much more modest. (017)
Google made the decision. They have a lot of data about what people
actually use. They didn't take the future away from the SW. They
observed that the SW boxed themselves into a tiny niche. (018)
PT
> I searched all the schema.org documentation pages for "json", and
> found no matches. (019)
Of course not. Google adopted JSON as the foundation for Google Apps.
But Schema.org also includes Microsoft (Bing), and they're not going
to let Schema.org endorse Google's strategic directions. (020)
PT
> But if [JSON] is the centerpiece of some big new approach to semantics
> on the web, [Schema.org] is not advertising it as such. (021)
No, but Google is. JSON is the native notation for data in JavaScript,
which every webmaster uses. Instead of the untyped triples of RDF,
JSON supports typed N-tuples, in which the types can be any tags
from the Schema.org ontology (which now includes GoodRelations).
Following is an example: (022)
{T1:V1, T2:V2, ..., Tn:Vn} (023)
You could map this to RDF, but why bother? You would get a huge
expansion of triples. It is much easier to map them to an RDB or
a spreadsheet in which the Schema.org tags are the column headings. (024)
If you want an implementation, see http://code.google.com/p/jsonengine/ (025)
The jsonengine is a NoSQL database that uses JSON notation for the
data and the queries. It's much, much simpler and more readable
than SPARQL. But you can also use SQL for complex queries. (026)
Bottom line: The current SW tools and notations will survive
for a while, but Schema.org is where mainstream IT is heading. (027)
John (028)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (029)
|