NetNews Usenet Archive 1993 #3/NN_1993_3.iso/spool/comp/object/5052

Internet Message Format | 1993-01-25 | 3.4 KB

Path: sparky!uunet!pipex!unipalm!uknet!gdt!aber!fronta.aber.ac.uk!pcg
From: [e-mail] (Piercarlo Grandi)
Newsgroups: comp.object
Subject: Re: FAQ  Part 1 (of 2) [ a bit of polemic ]
Message-ID: <[e-mail]>
Date: 25 Jan 93 21:32:31 GMT
References: <[e-mail]> <[e-mail]>
    <[e-mail]> <[e-mail]>
Sender: [e-mail] (USENET news service)
Reply-To: [e-mail] (Piercarlo Grandi)
Organization: Prifysgol Cymru, Aberystwyth
Lines: 59
In-Reply-To: [e-mail]'s message of 21 Jan 93 02: 44:52 GMT
Nntp-Posting-Host: decb.aber.ac.uk

>>> On 21 Jan 93 02:44:52 GMT, [e-mail] (Bob Hathaway) said:

Hathaway> "Simply put, an object is a variable"

Hathaway> Actually, I changed the above to "instance" instead of
Hathaway> "variable" and all should be well.

On this I would like to speculate. Your original text contained too many
times the word "variable" out of place to make me think it was just a
typo. Many times I have had the impression that your rather terrible
terminology covers a fairly intuitive (but not very deep) understanding
of this field.

In the particular case my impression is that you meant 'an object is a
variable' as a deformation of the concept, that indeed appears in the
literature, that an object is a *mutable* value/data structure, which is
quite a different thing, but one that it is rather likely you might have
been confused by.

The "OO programming is discrete simulation" school of thoughts maintains
that objects represent the state, associated with state access/change
functions, of "real world" entities. In other words, that an object is
the state vector of a (finite) state machine (the state machine
definition is the class -- in programming terms this translates to the
procedure/procedures instance view of classes/objects).

According to this view, an object is inherently mutable, and has been
described as such in the literature, because an immutable state vector
is quite inconceivable in a simulation context.

So I speculate that this might have been what you had in mind when you
wrote the above definition.

(let me register that even if I think that the definition of object as a
"mutable data structure" is defensible on the grounds above, and can be
found in the literature, I do not like it; the view that objects are the
state vector of some state machine, precisely because this makes the
notion of "frozen" state, i.e. a constant/immutable object, rather
unnatural, when instead it is quite useful).

Hathaway> Also, I seem to recall you trying to give an example where a
Hathaway> hierarchical class structure was not necessary and failed
Hathaway> miserably

What I remember is being challenged to give examples where a
hierarchical class structure is not *sufficient* and me pointing out
that I need not to, as the limitations of hierarchical data
classification schemes have been well documented in database textbooks
for over a dozen years (the network and relational data models, and the
need to have non hierarchical data dictionaries too, have been around
for a dozen years).

Hathaway> Anyway, my main point was to ignore these superflous comments
Hathaway> and get on with things, not to heat them up, so no more.

This is against all netiquette as clearly illustrated by Emily Netnews.
--
Piercarlo Grandi <[e-mail]> c/o Dept of CS
University of Wales, Penglais, Aberystwyth SY23 3BZ, UK