Relationship between OWL and DAML+OIL

Everyone,

One goal of the release of OWL is to make it easy for people to understand
it and respond to it. A substantial portion of people who will be reading
these documents will already be familiar with DAML+OIL. There should be NO
NEED for these people to be forced to read descriptions that are essentially
identical to DAML+OIL.  It is a slow painful process figuring out what is
different from DAML+OIL. Frankly, I cannot afford the time to pour over  the
full documentation in any detail, even just the feature summary.   What I
would like is to be able to spend no more than an hour, and possible as few
as 15 minutes, to get a grasp of the essence of OWL.  The best way to do
this is to provide a summary of changes from DAML+OIL to OWL.   Such a
document could dramatically improve the number of people who take the time
to respond to the proposal. I would expect that tool developers would find
this particularly useful also.  

Last night, I asked Ian Horrocks what were the key differences, what I
learned was:

*	OWL and DAML+OIL to a first approximation are the same. There are
some cosmetic changes, and there are some minor technical differences.  For
example: 
o	Improved names that say exactly what they mean, instead of being
semantically opaque geek-speak.: e.g. allValuesFrom and someValuesFrom
rather than toClass and has-class.  
o	no qualified cardinalities-there was no business case defending
their need, very hard to understand, and difficult to implement.

I recommend that someone put together such a document that would probably
only be a few pages long and would be the first place to look hat would probably
only be a few pages long and would be the first place to look for someone
wanting a quick understanding of what is new about OWL.  A longer and more
useful document might be constructed which would contain sufficient detail
so that a tool developer who had already developed full support for DAML+OIL
could use that as a changes specification and checklist for updating the
implementation. They would need only read this document, and any relevant
portions of the full reference documents.  Ideally, they should not even
have to look at the other portions of the documentation.

Perhaps there already is an intention to produce this? 

Mike Uschold

Received on Tuesday, 10 September 2002 17:00:11 UTC