RE: Ontolingua tests

James Rice <rice@HPP.Stanford.EDU>
Date: Thu, 17 Nov 1994 11:33:05 -0800 (PST)
From: James Rice <rice@HPP.Stanford.EDU>
Sender: rice@HPP.Stanford.EDU
Reply-To: rice@HPP.Stanford.EDU
Subject: RE: Ontolingua tests
To: Mike Uschold <mfu@aiai.edinburgh.ac.uk>
cc: Ontolingua@HPP.Stanford.EDU
In-reply-to: Mike Uschold's message of Thu, 17 Nov 94 13:36:59 GMT: <25114.9411171336@subnode.aiai.ed.ac.uk>
Message-id: <XLView.785103101.5964.rice@hpp-ss10-1>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; CHARSET=US-ASCII

Mike Uschold <mfu@aiai.edinburgh.ac.uk> writes:
>>I am interested in participating in testing the ontolingua editing
>>tools over the web.  PLease send more info.

>>Mike Uschold

Hi,
This is the current state of affairs:

[For those of you who didn't know] We're in the midst of producing
a user interface for Ontolingua which will allow the browsing,
creation and editing of ontologies from anywhere on the net using
your favorite web browser[*].

The goal is that rather than us shipping you software and you
having all of the pain of installing it on some random machine
we can concentrate on the interesting issues, such as the
collaborative development of ontologies.  To this end we're
planning to support remote read/write access to our ontology
library.  I suppose that that means that we hope to be a kind
of clearing house for ontology development.

For those of you already familiar with the way that the released
Ontolingua generates hyperwebs of your ontologies (see
http://ksl-web.stanford.edu if you haven't seen any of this yet)
there will be no major surprises.  Broadly speaking, the user interface
looks just like the existing hypertext display of ontologies, only
whereas previously the hyperweb was created by a batch compilation
of whole directories of ontologies, the user interface does real-time
display of an in-core representation of the ontology.  Whereas
previously you were just presented with a page to look at, you now
have a whole bunch of edit buttons and commands available.

This all happens within the same broad outline of viewing an
ontology that you're used to from the current release.  Thus,
if you chose to edit a slot value on a frame, you end up seeing
the same page displaying a frame as you did before except that
where the slot value was displayed you now get a typein box.
There's more to it than that, but this is the gist of how it
works/looks/feels.

FWIW, the interface also displays inherited slots, values and
facets and deduces things like inverse slots.

In order to do all of this we've had to do a substantial rewrite
of Ontolingua on top of a whole lot of new code.  For example,
because Ontolingua used to be a single-user batch process it
didn't care about poluting the global environment.  We now need
to be able to support multiple simultaneous users and we need
to be able to GC the state that they create without rebooting
the server.

This has all been done and seems to be working reasonably reliably.
At present we're alpha testing within the lab.  We hope soon to
be able get a few alpha testers from around Stanford (it's still
very useful for us to be able to watch over the shoulders of new
users since that gives us ideas about how we may be able to polish
the user interface design).  With any luck we'll be going into
beta mode soon.

The bottom line is that I'm afraid you'll have to wait a bit.
You might go a bit blue if you hold your breath, but you shouldn't
drop dead.





Rice.





[*] Note that the user interface uses some of the newer features
of HTML.  At present, the only web browsers that we are aware of
that support these features sufficiently are Mosaic >=2.5b2 and
Netscape.  However, Netscape has proved to be sufficiently buggy
that it is not currently usable as a interface to Ontolingua.  Note:
this is not a general reflection on Netscape, which for most uses
seems to be very slick and reliable.