Leo and Simon, (01)
Thank you for confirming every criticism I have ever made about the
so-called Semantic Web. As I have always said, there are many very
intelligent people who have worked on aspects of the SW, and I cannot
understand why they have not removed this "train wreck" from the tracks. (02)
Leo
> I used to consider blank nodes as just a kind of Skolemization (and
> onecan give a blank node a name easily enough)... (03)
A Skolem constant is one special case. But an existential quantifier
is more general than a Skolem constant. For example, you might say
"There exists an x such that P(x)." If there is exactly one such x,
then the Skolem constant names it. But if there are two or more
x's that could satisfy P(x), then only one of them is chosen. (04)
The choice of just one of the options for the Skolem constant is
acceptable for some purposes, such as a resolution theorem prover.
But there are other purposes for which it's necessary to consider
all the possible values of x. (05)
Leo
> but I've seen many different interpretations of blank nodes. (06)
Yes, indeed. And most of them are special cases of what can be done
with an existentially quantified variable in logic. But others are
examples of what is best done by replacing the backwards E with
the Greek letter lambda to define a lambda abstraction. (07)
If the SW had started with semantics instead of syntax, these
issues would have been obvious a dozen years ago. (08)
Leo
> Pat Hayes would be the one to ask about this. (09)
Pat has more sympathy for the SW than I do. (010)
SS
> Here's a good recent paper:
>
> Mallea, Alejandro; Marcelo Arenas; Aidan Hogan; and Axel Polleres.
> 2011. On Blank Nodes. ISWC 2011.
> http://db.ing.puc.cl/amallea/papers/mahp-iswc2011.pdf. (011)
They are more polite than I am. But they make the same kinds
of observations that Pat and I have made. (012)
SS
> I think Pat has formally defined blank nodes as a kind of train wreck... (013)
Yes. Following is the conclusion of the paper you cited: (014)
Mallea et al.
> Finally, we note that no alternative stands out as “the one solution
> to all issues with blank nodes”. Discussion is still open and proposals
> are welcome, but as the amount of published data grows rapidly, a
> consensus is very much needed. However, in the absence of an undisputed
> solution, the community may need to take an alternative which might not
> be the most beneficial, but the least damaging for current and future users. (015)
The short translation to everyday English: "a train wreck". (016)
As the "least damaging" alternative, I recommend the following method
for clearing the tracks of this train wreck (and many related wrecks): (017)
1. Start with a formally defined version of logic that includes
at least classical first-order logic and model theory. (018)
2. Define each aspect of the SW notations and the operations on them
by a translation to that logic and formal operations on that logic. (019)
3. For any aspect or use that cannot be precisely and unambiguously
translated to the chosen logic, declare it to be *DEPRECATED* . (020)
This exercise should be carried out for *every* RFC that specifies
any aspect of any notation or tool related to the SW. (021)
John (022)
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2012/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012
Community Portal: http://ontolog.cim3.net/wiki/ (023)
|