Todd, Kingsley, Michael, and Doug, (01)
Please note my observation: (02)
JFS
> This doesn't mean that the SemWeb is dead. But it does mean that
> the 2-page quick reference for the Semantic Media Wiki is a better
> clue to the future of the SemWeb than any of the layer cakes. (03)
I agree that the current SW tools (RDF, RDFS, SPARQL, and OWL) are
widely used. I'm sure that they will continue to be used. But note
the evolution of the layer cakes from Tim B-L's DAML proposal of
Feb 2000 and various intermediate reports, to the final report
in 2006. (See http://www.w3.org/2000/01/sw/DevelopmentProposal ) (04)
Tim's original vision was much broader and put more emphasis on
a general logic foundation (the Semantic Web Logic Language, SWeLL).
It proposed full support for a broad range of languages including
SQL, Prolog, and others. Two keywords that Tim emphasized in the
first report are *heterogeneous* and *diversity*. (05)
Later DAML reports narrowed that vision. The final report eliminated
diversity and claimed that OWL met the goals of SWeLL. I don't blame
them for not implementing Tim's full vision in five years. But I do
blame them for ignoring Tim's much broader vision -- especially the
support for SQL and the emphasis on diversity. (06)
TJS
> I've used SMW and there is an ontology (of sorts) and
> there is a reasoner (TSC), as an external option. (07)
That's good. It shows how the SMW is evolving. But there is still
a long way to go. (08)
Since Google and Microsoft are two of the major supporters of both
WebPlatform.org and Schema.org, I would expect a lot of convergence
with them. But Tim's proposal emphasizes diversity, and we should
expect a much broader range of logics and ontologies to be included. (09)
KI
> see also:
>
> http://bit.ly/UnOpac -- LInkedWiki Demos (basically MediaWiki +
> Semantic MediaWiki + LinkedWiki extensions deployed via Virtuoso...
> enhanced via use of SPARQL instead of SQL (010)
But note that most vendors of triple stores support SQL as an option
for complex queries. I noticed that none of the examples on that page
use the SPARQL operators FILTER, OPT, or UNION. If SPARQL is going
to survive for the long term, it will have to replace them with the
SQL operators: AND, OR, NOT, and EXISTS. (011)
People understand AND, OR, NOT, and EXISTS -- and their combinations.
They can be evaluated in polynomial time in all SQL implementations.
But nobody, not even the SPARQL developers, understand FILTER, OPT,
and UNION. They're still publishing papers to show that combining
them makes the evaluation NP complete. (012)
MB
> But the properties they use always have a subject and one object, they
> do *not* have n-ary relations/properties. This is why they can offer
> to store the data in a triple store and use it for faster queries (013)
They also support files with comma-separated values (CSV) -- the oldest
and moldiest way of storing large volumes of data. Spreadsheets, which
map to and from RDBs, are very widely used. They're not going away. (014)
Furthermore, triple stores are just a particular implementation.
Note Michael Stonebreaker's publications about NOSQL (in the sense
of Not Only SQL). He has an extremely fast implementation that uses
the same optimizations as triple stores, but he extends them to
support relations with any number of arguments. (015)
MB
> N-tuples may be desirable or not but the implementation decisions
> of the Semantic MediaWiki developers do not indicate this, IMO. (016)
That's true of the current version. But when the heavyweights with
multi-billion-dollar revenue streams take charge, the Golden Rule is
enforced: he who has the gold makes the rules. (017)
DF
> What is it a definition of?
>
>> {{#set_internal:Is president of|Has name=James Madison
>> |Has start year=1801|Has end year=1809|Has vice
>> president#list=George Clinton, Elbridge Gerry}}
>
> It seems that this is a definition of either a Presidency or a Role.
> I note that the "Has start year", "Has end year", and "Has vice president"
> predicates can also apply to the Presidency or the Role. (018)
I won't defend their logic or their ontology. It shows what happens
when people ignore logic and just start hacking. The only logic RDF
ever had was the LBase report by Hayes & Guha, which was consistent
with Tim's original SWeLL. But the W3C ignored both LBase and SWeLL. (019)
Observation #1: Google, Microsoft, Apple, Adobe, Facebook... don't use
SW tools. They accept RDF as input, but they use other representations
for their processing. They will drive WebPlatform.org, which will drive
the future development of SMW, which will drive whatever the SW becomes. (020)
Observation #2: Academics who are experts in logic and ontology
have no money. Some of them got funding from DAML and drove it
in directions that do not support mainstream IT. As a result,
companies that profit from mainstream IT dropped their funding. (021)
Observation #3: The big corporations that joined forces to support
WebPlatform.org will cooperate when it's to their advantage, but they
are also competitors. The only way that academics can influence their
direction is to develop methods that (022)
1. can interoperate with the tools they are already using, (023)
2. are demonstrably superior to their current methods, and (024)
3. show promise of improving their bottom line. (025)
Summary: I was enthusiastic about the goals for the Semantic Web
as Tim outlined them in his original talk (1994) and in the DAML
proposal (2000). But Tim's vision was lost. It's time to revive it. (026)
John (027)
_________________________________________________________________
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 (028)
|