W3C home > Mailing lists > Public > public-owl-wg@w3.org > November 2007

UFDTF - who are we writing for

From: Jim Hendler <hendler@cs.rpi.edu>
Date: Mon, 26 Nov 2007 12:00:52 -0500
To: "OWL Working Group WG <public-owl-wg@w3.org>" <public-owl-wg@w3.org>
Message-Id: <A80EEDE0-78C6-4B64-AE53-C5FECF636142@cs.rpi.edu>
I have taken a long time to answer this question, because I had to  
count to 10, and then count to 10 again -- I've now reached somewhere  
around 10^^6 and am calm enough for an answer - in fact, several:

ANSWER 1: Someone like me wants a reference manual
  OK, Bijan argued that since users will mainly use OWL through  
tools, they didn't really need to know most of this.  Ok, maybe, but  
suppose I got to a SWOOP or Protege for OWL1.1 and create a new  
property - I get a menu asking me what kind of property it is.  One  
of my choices is
         owl11:irreflexive property.
hmm, I vaguely remember that reflexive and irreflexive were things I  
learned about in college math over 30 years ago, so I had better  
check the documentation to see if this is what I want.  The overview  
gives me something like an example
which confuses me a bit because the term my editor shows me is  
"irreflexiveProperty" but even assuming these are the same, this one  
example doesn't tell me anything. (and the example for Reflexive is  
"knows" which happens not to be a reflexive property in many systems,  
but I digress)
  In the structural document all I can discover is that it is  
syntactically legal in some syntax I don't use (remember, I'm doing  
this through an editor as Bijan argued, so I don't see any syntax  
except what is displayed by the editor).  So that doesn't help.   
Gritting my teeth, I go to the model-theoretic semantic (most users  
by now would have simply decided not to use it, but I'm a bit  
stubborn).  I find there (in a table, which means my browsers search  
didn't find it so I had to scroll through the whole document) that
     x ∈ ΔI implies ( x , x ) is not in RIpo  (sorry for the  
syntax, but that's how it comes out when I cut and paste to mail it  
to someone who might be able to answer the question).
    Well maybe from that I could figure out the answer if I was  
slightly less stupid, but I'm still confused, I check the RDF  
mapping, but it also only gives me syntax.
   So now I go out on the Web and find some definition of irreflexive  
and I finally get it -- time wasted because there was no manual - on  
the order of 30-45 min.  And that's for one of the easy one (try  
ObjectExistsSelf where { x | ( x, x ) ∈ RIpo } is the only  
   There's a lot of functions in this thing named after math I'm not  
familiar with and defined only in a logic I'm not familiar with.  My  
conclusion - to hell with OWL, make up my own stuff in RDFS.

ANSWER 2: My students - a guide
  Ok, I give a lecture in class, I give some examples, I assign my  
students to go write an ontology.  Oops. every one of them has the  
problem above in spades - not only don't they understand this stuff,  
but they don't have my experience with OWL.  So they puzzle out what  
they think things mean, but they want to try it -- boy, would be a  
lot easier if they had some examples they could cut and paste into  
something (validator, etc.) so they could not only figure out the  
meaning, but also puzzle out the syntax.

ANSWER 3: People considering use of OWL for a project - better overview
  Dean Allemang and I are in the finishing stages of a book about how  
to use OWL 1.0 -- we were motivated because in teaching professionals  
interested in using the Semantic Web (usually the decision makers for  
their companies or people who report to the decision makers) we found  
that they had read the overview, decided they liked the idea of OWL  
and wanted to take a course before they dove in.  (They then wanted  
something that emphasized the stuff we had in the course, so we wrote  
the book).  We've been teaching the course for several years now, we  
are finally getting the book done, but again, these users have  
usually read the overview - so the book alone wouldn't have been  
enough for them - they needed something to whet their appetites.
  In general these people are business people working for companies  
or for the government.  They are well-educated professionals (college  
grads in general, usually some sort of professional masters,  
occasional PhD in engineering or some non-CS field) with a high level  
of domain knowledge, and usually some programming background, but  
they've rarely taken a logic course.
  These users have a lot of demands on their time, without a simple  
thing to look at and decide they want to learn more, they usually end  
up doing something else. A number of them have  Shelley Powers' RDF  
book and are pretty much convinced by RDF/RDFS but are a bit  
skeptical of OWL (and I fear will be much more so of OWL 1.1 because  
the more complex a language, the higher the learning curve, and many  
of these folks are learning-curve adverse because of the time it takes)

So there's my top 3.  I've left out Web Application developers  
because in my experience they will mainly ignore all our documents  
and just use the pieces of the RDF (OWL Full) vocabulary that make  
sense to them, and ignore the rest - however, if there is a  
reference, they do tend to make sure that what they think is being  
defined is right - i.e. they would like to make sure the operational  
semantics of what they develop matches as close as possible to the  
reference manual.
p.s. I note that the OWL Reference and Overview are the two highest  
ranked in Google and interestingly the S&AS document, despite being  
normative, is lower in Google rank than any of the others except for  
Use cases, with which it is tied...  I know Google rank is variable  
and also not a definitive argument for anything, but it does indicate  
more people linking to the user facing documents over the developer- 
oriented one.  Doesn't deny the importance of that document, just  
stresses the importance of the others to uptake...

"If we knew what we were doing, it wouldn't be called research, would  
it?." - Albert Einstein

Prof James Hendler				http://www.cs.rpi.edu/~hendler
Tetherless World Constellation Chair
Computer Science Dept
Rensselaer Polytechnic Institute, Troy NY 12180
Received on Monday, 26 November 2007 17:01:30 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:42:00 UTC