[sc34wg3] Re: [tmcl-wg] Type versus Class: settled yet?

Murray Altheim sc34wg3@isotopicmaps.org
Mon, 31 Mar 2003 14:37:26 +0100


[I was away all of last week and only after writing this noted that
Mary gave up fighting with you over this one. I send this not as a
voting SC34 member but to simply note that she was not being "fussy"
alone. I guess I won't be using TMCL, or will redefine its use of
"type" within my own work, if that's possible. I *do* make a strong
distinction between "type" and "class" (I use both), and I understand
them in a way incompatible with the Barcelona decision, in keeping
with the logic I'm using.]

Lars Marius Garshol wrote:
> * Mary Nishikawa
[...]
> | So now we should be using the terms topic type, subtype, and
> | supertype. While working on this draft, I found I really wanted to
> | really use "class"not "type." I don't think they are interchangable
> | in all cases.
> | 
> | Class of things is the set of things. Type of things is not, it is
> | the category that describes the things in the set.
> 
> I can't say that I agree with this. You may be able to find an example
> of someone using these two terms this way, but I don't think that this
> is generally correct.

Do you have any more evidence for this being generally correct than
the statements that Mary and I have made? We're merely bandying about
opinions at this point, it seems to me.

I do think one thing is quite clear, and that some people don't care
that much about this distinction in language, or don't see a distinction,
and others care deeply (and do see the distinction), and want to get it
"right".

> | So we usually say, what type or (kind) of ice cream is your
> | favorite?  We do not say, what class of ice cream is your favorite.
> 
> This is colloquial use of the terms, not technical.

As I mentioned previously, there's been many a long, historical argument
about this issue. We're here repeating many of the discussions I've read
from the 19th century, so I'd hardly call this a colloquialism. If she'd
used DNA being discussed by a group of scientists instead of ice cream,
the example would have still held.

> * Dan Brickley
> |
> | Hey, if you folks haven't picked names for these two relations yet,
> | I hereby propose 'type' and 'subClassOf'; type relates an instance
> | to a class it is a member of, while subClassOf is a relation between
> | classes.  FWIW these are the terms we use in RDF/S and OWL.
> 
> I discussed this very briefly with Dan a couple of days ago:
[...]
> Mar 21 20:13:05 <larsbot>	"we call it 'type' and 'class' 
>                                 interchangeably. could you please be
>                                 as inconsistent as us?" :-)
> Mar 21 20:13:10 <danbri>	ppl were arguing about names
> Mar 21 20:13:13 <danbri>	isa etc
> Mar 21 20:13:21 <danbri>	yeah, fair point
> Mar 21 20:13:27 <danbri>	a things type is a class
> Mar 21 20:13:41 <larsbot>	we decided to just call it "type" and
>                                 leave "class" undefined
> Mar 21 20:13:46 <danbri>	fair enough
> 
> (I'm quoting my IRC client logs, since I'm not online as I write
> this.)
>  
> | It would be good if we came to some agreement on the use and try
> | make sure that these are crystal clear.  So I am asking you to look
> | at these definitions in RDF/S and OWL and comment on it. The
> | definitions are not clearly defined (as far as I can see) but we
> | need to do so ASAP. It very important for the TMCL.
> 
> I think what we should do is go with the Barcelona decision: we call
> it type, and that's that. SAM has a simple definition of that notion
> and I'm not sure we need to do anything more than that.

So your personal preference is "type" though Dan has essentially
confirmed Mary's and my point about something's type being a class
(Mar 21 20:13:27).

How the PSIs' titles and documentation are written is important,
and I believe the current XTM 1.0 set is correct in its use of "class".
As in many fields that rely heavily on taxonomies, they are defined as
class relationships, not type relationships (e.g., zoological
taxonomies). Something can have a type or be a member of a class. It
does not have a class or be a member of a type. (I've never heard that
usage anyway.)

I understand you are comfortable with the Barcelona decision because
you agree with it, but it's hardly "that's that". If that decision
stands, I'll just simply not use those PSIs, since they disagree
quite strongly with the logical terminology I'm using -- that's one
of the beauties of PSIs. But I'll also then not be using TMCL if its
language is based on this (to my understanding) incorrect use of
type and class. So be it.

I think one thing that may have come about in our private discussions
about LTM is that I seem to be using superclass-subclass relations
a lot more than you -- they form the basic framework of most of my
ontologies, so I put more emphasis and importance on them, and their
PSI definitions.

I just wouldn't suggest that the Barcelona decision on "type" is in
some way correct or universal -- it's not.

Murray

......................................................................
Murray Altheim                  <http://kmi.open.ac.uk/people/murray/>
Knowledge Media Institute
The Open University, Milton Keynes, Bucks, MK7 6AA, UK

    Hunt the Boeing! And test your perceptions!
    http://www.asile.org/citoyens/numero13/pentagone/erreurs_en.htm