Newsgroups: comp.lang.smalltalk
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!oitnews.harvard.edu!news.sesqui.net!imci2!imci3!imci4!newsfeed.internetmci.com!info.ucla.edu!library.ucla.edu!ihnp4.ucsd.edu!munnari.OZ.AU!news.mel.connect.com.au!harbinger.cc.monash.edu.au!news.cs.su.oz.au!tmx!news.tmx.com.au!news
From: Philip Haynes <p.haynes@oose.com.au>
Subject: Re: Are there C++ vs. ST figures?
Message-ID: <3144AEC6.678B@oose.com.au>
Nntp-Posting-Host: 203.5.19.186
Sender: news@online.tmx.com.au (System Administrator)
Mime-Version: 1.0
X-Mailer: Mozilla 2.0 (WinNT; I)
Content-Type: text/plain; charset=iso-8859-1
Organization: Object Oriented Pty Ltd
Date: Mon, 11 Mar 1996 22:52:54 GMT
References: <expertek.315.005A3A52@aone.com>
Content-Transfer-Encoding: 8bit
Lines: 57

Kirk W. Fraser wrote:
> 
> Hi,
> 
> We've seen the figures that a programmer using ST can deliver a system from
> two to five times faster than a C programmer, though the result operates
> slower.  Does anybody have the comparison figures for C++?
> 
> I strongly suspect that ST also blows C++ out of the water but I haven't seen
> it in print.  Solid figures might quench the current C++ hysteria in the job
> market.

I did a study a couple of years back where I wrote a system in C++ &
then rewrote it in Smalltalk. The Smalltalk development was 2x faster.
I wrote this paper up in:  Haynes, P. and T.J. Menzies. "C++ is Better 
than Smalltalk"? in Tools Pacific 1993. Melbourne, Australia: Prentice Hall. 
Whilst hardly being a definitive paper, I think the results it represents
are typical of the productivity difference that can be seen for 1-4 people 
teams.

Since that time, amongst other things, I have been studying the reasons why 
I observed a difference. If the actual structure of the two languages is examined
it turns out that there is not much difference, ie the languages are pretty
much of a muchness, (if used properly). The main productivity difference comes
from the strength of the Smalltalk development environment. 

Thus if Smalltalk is more productivity because of the environment, the next question 
is does the environment scale? Well, currently I have been doing the estimates for
government contracts & the conclusion that I have to reach is, currently no. The reason,
documentation, quality management,  project management, requirements modelling,
system testing activities fall outside the scope of the Smalltalk development environment
and on larger projects chew up a fair wack of the project budget (ie around 60%). So
the net result is Smalltalk makes the design, code, unit test, integration parts of
the project go 100% faster, but you only save 15% off the total project budget.

This said, I ran into some guys at OOPSLA last year & their project experience
was different, in that they reckon Smalltalk is 2x as productive as measured 
by their fixed price quotes.

However, figures of 5-10x more productivity in Smalltalk, in my experience
is marketing hype.


I hope this helps
Regards
Phil Haynes


---------------------------------------------------------
        Philip Haynes           | 
        p.haynes@oose.com.au    | "You cant control what you 
 ,-_|\  Object Oriented Pty Ltd |  cant measure. If you think 
/     \ P.O. Box 528,           |  the cost of measurement is  
\_,-._/ North Sydney            |  high, consider the cost of   
     v  NSW, Australia, 2059    |  being out of control."
        +61-2-9957-1092 (ph)    | 	- Tom Demarco
        +61-2-9956 5089 (fax)   |
