Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!dali.cs.montana.edu!caen!math.lsa.umich.edu!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!att!cbnewsk!pegasus!hansen
From: hansen@pegasus.att.com (Tony L. Hansen)
Newsgroups: comp.object
Subject: Re: A Rewrite Is *NOT* An Experiment
Summary: how much language and how much environment?
Message-ID: <1991May7.140550.15767@cbnewsk.att.com>
Date: 7 May 91 14:05:50 GMT
References: <44.UUL1.3#913@acw.UUCP> <jls.671420246@rutabaga> <4405.281bed6a@iccgcc.decnet.ab.com>
Sender: hansen@cbnewsk.att.com (tony.l.hansen)
Organization: AT&T Bell Laboratories
Lines: 35

< From: klimas@iccgcc.decnet.ab.com
< One can raise the argument that when rewritting a program in an OOP
< language from a non OOP language, it is easier because one is
< benefitting from the accumulated design knowledge in the existing
< code.

< This was an interesting argument that was difficult to disprove until I
< had a real life experience relayed to me by a software tool maker.  To
< meet some tight deadlines for an upcomming trade show, they did a show
< demo in Smalltalk. Afterwards they had to reimplement the same program
< in Objective-C which was a little harder and took longer even though
< the developer's were well versed in the language.  Finally, the same
< tool was rewritten for C++ which was much harder to do and took
< significantly longer.  The tool developer has his greatest market with
< C++ users so he's reluctant to tell his customers they are not
< necessarily using the best technology.

< This experience seems to disprove some of the ideas that it is easier
< to redo software the second time around.

I can't help but wonder how much of the problems doing the reimplementation
were due to lacks in the language vs. a lack of an equivalent library and/or
environment. For example, if the Smalltalk library provided a particular
flavor of collection class which was crucial to the implementation, and a
comparable collection class weren't already available in Objective-C and/or
C++, such a collection class would probably have to be coded from scratch in
order to continue with the port. Should that be considered a problem with the
technology or a problem with the maturity of the technology? In other words,
the technology may be sound but just not have sufficient support.

We need more data.

					Tony Hansen
			    hansen@pegasus.att.com, tony@attmail.com
				att!pegasus!hansen, attmail!tony