Re: [ontolog-forum] Last Call: OWL 2 and rdf:text primitive datatype

Pavrithra,

 > You are questioning the basic terminology of Object Oriented
 > techniques of Computer Science , which is in practice from 1989/1990
 > may be even earlier.  It is accepted and is in practice in the
 > industry.  Many of us studied OO as part of graduate program as
 > early as that.

I learned OO techniques from the Simula 67 language, which began
the practice of using the word 'class'.  SmallTalk adopted that
term from Simula.  But Simula also influenced the work on abstract
datatypes, which were also being developed during the 1970s.

 > Many analysis, design, and programming languages use them. OOA,
 > OOD, MDA,  SOA,and all the programming languages that support OO,
 > have them as well..

I realize that.  It's too bad that they didn't adopt the terminology
used for abstract datatypes.

 > The world is pro innovations!

I agree.  But sometime it's desirable to revise the terminology
to reduce the confusion.

For example, when the Vikings invaded England, Old Danish and Old
English were mutually intelligible.  In most cases, the Anglo-Saxon
terms survived, but occasionally, the Danish terms won out.  The
Danish 'egg' replaced the Anglo-Saxon term, which had become a homonym
for 'eye'.  In other cases, both variants survived, but with slightly
different senses.  So English has the Anglo-Saxon derivatives 'shin'
and 'shirt' as well as the Danish derivatives 'skin' and 'skirt'.

In current terminology, we have three terms 'set', 'class', and 'type'.
The words 'set' and 'type' have long histories in both technical and
informal usage, and there is no confusion about their meanings.

The word 'class' has been used in so many conflicting ways in both
technical and informal usage that it causes a great deal of confusion.

Therefore, my recommendation is to prefer the words 'set' and 'type'
for all metalevel usage, both technical and informal.  The word
'class' could still be used for particular languages in which it
is endemic.  But to promote precision, the words 'set' and 'type'
should be preferred for all cross-language discussions.

John Sowa

Received on Sunday, 26 April 2009 19:25:34 UTC