Newsgroups: comp.lang.smalltalk
Path: cantaloupe.srv.cs.cmu.edu!das-news.harvard.edu!news2.near.net!MathWorks.Com!europa.eng.gtefsd.com!howland.reston.ans.net!swrinde!pipex!sunic!news.funet.fi!news.lut.fi!news
From: jahonen@sisu.lut.fi (Jarmo Ahonen)
Subject: Re: Smalltalk V/Win32 application size
Sender: news@lut.fi (Usenet News)
Message-ID: <CwqIHM.I23@lut.fi>
Date: Mon, 26 Sep 1994 11:27:22 GMT
Reply-To: jahonen@sisu.lut.fi
References: <CwDzIx.EHJ@cunews.carleton.ca> <35m248$r0c@news.cs.tu-berlin.de><patrick_d_logan.99.000A4B82@ccm.jf.intel.com> <56@manifold.win.net> <3664vq$mtr@news.cs.tu-berlin.de>
Nntp-Posting-Host: peruna.cc.lut.fi
Organization: Lappeenranta University of Technology
X-Newsreader: IBM NewsReader/2 v1.00
Lines: 58

In <3664vq$mtr@news.cs.tu-berlin.de>, muhr@cs.tu-berlin.de (Thomas Muhr) writes:
>Thomas Murphy (tmurphy@manifold.win.net) wrote:
>: >
>: In pure performance, the 32bit product is quite a bit faster than
>: the 16bit product.  The key being that it is much more memory
>: hungry so you can get to swapping faster and once you begin the
>: disk thrash....well you know the rest.  
>
>: There are some problems (more than some) with win32s.  There are
>: also problems with the windows dll format.  We are well aware of
>: these issues and hard at work on solving them for the next release.
>
>: tom
>Thanks, Tom, for finally admitting the poor quality of VWIN32. I have never
>got an answer (only additional questions) to the problems I described in 
>great detail (MDI, ERROR 87, etc). Digitalk support did try to help, but what
>can they do? This reminds me of my days (as a student) at DEC. Theere were
>two groups (at least) the developers and sales. Sales pushed the developers 
>very hard to release a product prematurely because they already promised
>their customers fantastic things. It was just a shift from beta-testing in-
>house to beta-testing/quality assurance at the customers site.
>I think that DT has made a mistake not to wait until they achieved a certain
>quality level for the VWIN32 product. I also think, that they should have
>improved the existing versions, including the V/286 version (new DOS extender).
>Now ST is back to where it started: a prototyping language that does not
>meet the demands of a professional application. Although I hate Basic, it is
>hard to see that modern dialects surpass the more sophisticated languages
>even in domains that they claim to be specific: prototyping capability,
>module reuse (Visual Basic has myriads of reusable components for professional
>developers, what can ST offer?), event driven programming, support of host
>specific and host independent features like OLE2. The use of the latter
>might adds a level of object-orientedness to applicatiuons that even 
>Smalltalk cannot offer.
>
>I would go so far to say, that I would switch to VB if someone would offer
>a manageable way to transfer 1.5 MB of ST source code. The expected benefits 
>would include efficiency, portability, use of modern concepts and maybe even
>better support for prototyping. 
>

That is mainly true. I use OS/2 versions of ST/V (VOS2 2.0 now),
and although VOS2 is a usable tool for runtime-application delivery
also, its integration with OS/2 does not exist.

There is no direct support to REXX and named pipes and many other OS/2 
features I find important. :-(

And I wouldn't be surprised if the next release would be so different that
porting my existing applications will require very intensive rewrite. :-(

	

Jarmo Ahonen
Lappeenranta University of Technology




