Newsgroups: comp.lang.smalltalk,comp.lang.object,comp.lang.c++
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!satisfied.elf.com!news.mathworks.com!uunet!boole!jka
From: jka@boole.com (John Ahlstrom)
Subject: Re: C++ and Smalltalk Productivity
Message-ID: <1995Jan13.184124.10801@boole.com>
Keywords: Smalltalk, C++, Programmer Productivity
Organization: Boole & Babbage, Inc.
References: <1995Jan11.214042.15872@boole.com> <1995Jan12.185928.12748@molienergy.bc.ca>
Date: Fri, 13 Jan 1995 18:41:24 GMT
Lines: 56
Xref: glinda.oz.cs.cmu.edu comp.lang.smalltalk:19515 comp.lang.c++:107370

The original description is far from complete.  That was all I got from the
original describer.  I have no information whether the Smalltalk DESIGN was
used for the C++ program.  Knowing the original describer, I would be astonished if it were 
but that is different from having information that it was or was not used.  

Can you contribute any good examples for arguing either side of this ?

In <1995Jan12.185928.12748@molienergy.bc.ca> janr@molienergy.bc.ca (Jan Reimers) writes:

>In article <1995Jan11.214042.15872@boole.com> jka@boole.com (John Ahlstrom) writes:
>>I am looking for documented or anecdotal descriptions of actual experience
>>that allows comparison of actual C++ and Smalltalk programmer productivity.
>>
>>Here is one such description:
>>
>> "About a year ago a large North American corporation suspended
>> all Smalltalk development, began all new development in C++
>> and began converting existing Smalltalk to C++.
>>
>> In one conversion from Smalltalk to C++ an 8 person group
>> including many of the 6 original Smalltalkers produced in 12 months
>> about 80% of the functionality that a 6 person group
>> produced in Smalltalk in 3 months.  Roughly 6*3 : 8*12 or
>> 18:96 or 5+:1 productivity edge for Smalltalk."
>>
>> They have abandoned their plans to convert old applications
>>from Smalltalk to C++ but are doing all new development in C++.
>>
>Any c++ guru will tell you that smalltalk DESIGNS do not translate well into
>c++.  This dosen't say the one is better than the other, just that design
>stratigies are very diffierent.  It seems that whoever decided that translating
>from smalltalk to c++ was a good idea was very unknowledgable.   I also suspect
>that the final c++ code will be be riddled with dynamic casts and such.

>This is therefore a very poor example to use for arguing that smalltalk
>better than c++.



>################################

>        Jan N. Reimers 
>      Research Scientist
>    Moli Energy (1990) Ltd.
>     Tel:   (604)-465-7911
>     FAX:   (604)-465-4398
>  email: janr@molienergy.bc.ca

>################################
>-- 
>janr@molienergy.bc.ca
-- 
John Ahlstrom			I can neither confirm nor deny 
Boole & Babbage			that these opinions are or are not
3131 Zanker Rd			held by anyone else.
San Jose CA 95134
