Path: utzoo!news-server.csri.toronto.edu!cs.utexas.edu!yale!cmcl2!uupsi!sunic!enea!sommar
From: sommar@enea.se (Erland Sommarskog)
Newsgroups: comp.lang.misc
Subject: Re: Runtime Polymorphism -- To Have and Have Not (Par
Message-ID: <2772@enea.se>
Date: 9 Mar 91 17:15:45 GMT
References: <559@coatimundi.cs.arizona.edu> <2741@enea.se> <46626@nigel.ee.udel.edu>
Organization: Enea Data AB, Sweden
Lines: 36

Also sprach Darren New (new@ee.udel.edu):
>It depends on whether you are building a prototype to
>be thrown away or whether you start out with the
>entire 500.000 line program and try to get it right
>on the first try.  It seems reasonable to me that you
>may want a "quicky" language to prototype a medium-size
>(~10K-30K lines) application which can then be rewritten
>into the final application in a different language.
>
>Remember: always plan on writing it at least twice.
>You will write it twice anyway, so you might as well
>plan on it.

Good point. However, there is a trap hiding here. Once
you have your prototype, be sure that some managment
person will give you a very blank stare when you tell
him that you will throw the prototype away and start
anew. I am not agreeing with him, I just think that
if you are not sure that you don't have management
behind you on such a point, abadon the idea.

As for writing it twice, you do not necessarily make
it more organized. I worked for two and half years
with a 500.000 lines information system. We did an
extensive revision of the system once we had in-
stalled it at the first customer. Not a complete
rewrite, this was impossible since we constantly
added new functions, but we took care of database-
access and other general functions, extracted them
and renamed them. But a lot of the things we did, 
couldn't have been written that way originally when 
the project first went into coding (which was eight
months before I came in), since several tools, both 
local and vendor-supplied came along the way.
-- 
Erland Sommarskog - ENEA Data, Stockholm - sommar@enea.se