Newsgroups: comp.object,comp.lang.smalltalk
Path: cantaloupe.srv.cs.cmu.edu!rochester!cornellcs!newsstand.cit.cornell.edu!news.kei.com!simtel!torn!nott!cunews!dbuck
From: dbuck@superior.carleton.ca (Dave Buck)
Subject: Re: Smalltalk-80: The Language
X-Nntp-Posting-Host: superior.carleton.ca
Message-ID: <DIqLyt.74t@cunews.carleton.ca>
Sender: news@cunews.carleton.ca (News Administrator)
Organization: Carleton University, Ottawa, Canada
References: <johnson.817221545@sal> <49cjig$jl1@newsbf02.news.aol.com> <yf3u43qq066.fsf@sabi.demon.co.uk>
Date: Tue, 28 Nov 1995 05:01:41 GMT
Lines: 95
Xref: glinda.oz.cs.cmu.edu comp.object:41490 comp.lang.smalltalk:31134

In article <yf3u43qq066.fsf@sabi.demon.co.uk>,
Piercarlo Grandi <piercarl@sabi.demon.co.uk> wrote:
>>>> On 27 Nov 1995 09:56:16 -0500, bytesmiths@aol.com (Bytesmiths) said:
>Bytesmiths> There he goes again, folks! For some reason, the Smalltalk
>Bytesmiths> archaeologist Piercarlo Grandi is suffering from the
>Bytesmiths> delusion that his favorite tool for pointless argument,
>Bytesmiths> "Smalltalk-80: The Language," is somehow "_the_authoritative
>Bytesmiths> definition of the Smalltalk-80, version 2, language".
>
>This is not a delusion: if you could remember the contents of the book
>in question, this is what the book claims, and I have chosen to take
>that at face value.

The basic Smalltalk syntax has changed little since the publication of
the Blue Book.  The only major changes are:
   - addition of temporaries within blocks
   - addition of double precision float literals
   - support for class instance variables
   - change of assignment symbol to :=
   - underscores are now allowed in identifiers

There may be a few other changes that I can't remember right now.  The
class library, however, has changed dramatically.  The UI classes are
all changed, Forms don't exist any more, the behavior hierarchy has
changed, screens and windows are completely changed, double precision
floats have been added and many more.

>Bytesmiths> This may be true for Smalltalk-80,
>
>So, make up your mind (if any, and not null or void): either I suffer
>the delusion that a book with a title like "Smalltalk-80: The Language"
>by Goldberg and Robson is the definition of Smalltalk-80, or it is not a
>delusion.

The point is that Smalltalk-80 documents an old version of the language.

>Bytesmiths> but it is utterly irrelevant.
>
>I think here we hear the voice of ignorance speaking. Since you have
>been caught out several times as not "remembering" the contents of this
>book, now you claim it is irrelevant. That's a very convenient position.

I don't think anyone else on this forum would ever consider calling
Jan Steinman "ignorant".  Jan is very highly respected in the
Smalltalk community.  If I were you, I wouldn't be so quick to make
judgements and to post inflamatory remarks about people who may have
disagreements with you.

>Bytesmiths> Since then, ParcPlace has done whatever it felt like with
>Bytesmiths> ObjectWorks/VisualWorks, since they "owned" the spec, and
>Bytesmiths> other vendors have stretched things even further.
>
>You are in effect claiming that there are no two truly compatible
>implementations of "Smalltalk", and that anybody choosing one is locked
>into a specific proprietary solution forever. If this is true, and in
>effect in part I tend to agree, it is the single best reason for people
>not to invest in Smalltalk that you could have found.

The same thing happened in the C world.  For the longest time, the
de-facto standard was K&R C.  This was actually a terrible dialect and
most venders ended up adding new extensions to it such as enums.
These additions were recognized and made official in the ANSI standard
and many new features were added.

>Think of it: either people who invest in Smalltalk can *rely* on what is
>written in "Smalltalk-80: The Language" as the _only_ vendor independent
>common base, or they are are the mercy of any vendor's policies and
>continued exitence.

No Smalltalk vender these days supports an underscore as the standard
assignment symbol (although some may have kept it as a carry-over from
the old days).  No Smalltalk vender supports Forms, MVC as documented
in the Blue Book, fonts (as documented), pens, paths, bitblt, and so
on.  With modern Smalltalks, the Blue Book is about half correct and
half incorrect.  As such, it can't be trusted as a reference to the
language.

The standardization effort is an attempt to provide some level of
compatability between the different dialects, but only at a base
level.  Even in the C++ world, you have no standard library to use for
windowing systems.  Borland has OWL, Microsoft has MFC, and other
venders have other packages.  Again, it's no different.  If anything,
there's more commonality in Smalltalk than there is in C++.

David Buck
dbuck@ccs.carleton.ca

_________________________________
| David K. Buck                 |
| dbuck@ccs.carleton.ca         |
| The Object People             |
|_______________________________|
 


