Re: D2005 shelved

From: Captain Jake (jake[nospam)
Date: 12/28/04 

Date: Tue, 28 Dec 2004 16:26:44 -0600

Ximian <ximian@aol.com> wrote in message <41d1d34e@newsgroups.borland.com>
> From where I stand the going to scratch is a luxury nowdays,
> but I think that everytime a product goes from scratch it fixes design
> errors and that's how software gets better.

But it is not enough just for it to be better. It has to be better by a wide
enough margin that this improved quality results in more additional revenue
than it costs to do the rewrite.

Here's a fictional example:

  The XYZ Widget Company is making $4 million in annual sales. Their
developers want to use a new development framework, called ShinyMetal, for
now on, and they want to completely rewrite the existing apps using this new
development framework. This development framework is less expensive by $200 a
developer, and the developers estimate that they will be twice as productive
with the new tool. They convince the management at XYZ to let them go ahead
and rewrite all their software in ShinyMetal. The original estimate is that
it will take their team of ten developers 6 months to rewrite the software.
At $80K per developer, this means the rewrite will cost them $400,000 in
software developers' salaries. The team estimates that one additional QA
person should be enough for the rewite, an additional cost of $30K.
  So one year later the rewrite is finally done, at a cost of $860,000. The
better design in the software results in 10% in additional revenue, but bugs
introduced during the redesign require two man-years to fix, at an additional
cost of $160,000.

Total cost of rewrite: $1,020,000.
Total additional revenue pre year: $400,000.

So the rewrite would pay for itself in 2.5 years, except that the developers
refuse to do "maintenance", and so the firm has to hire two additional
programmers to do maintenance, at an annual cost of roughly $100,000. This
means the rewrite supposedly pays for itself in 3.3 years. Only problem there
is that six months after release of the rewrite, marketing comes over and
says the product is obsolete with that feature set. The same developers that
said ShinyMetal was the be-all and end-all of existence just a year and a
half ago now don't want anything to do with it, instead preferring to use the
brand new framework JibberJabber, so the firm fires them all and out-sources
the addition of new features to a firm in Absurdistan.

  That is actually a wildly optimistic scenario. In most cases, there will be
NO additional sales of the rewritten software, at least none that result from
it's having been rewritten. 

-- 
***Free Your Mind***
Posted with JSNewsreader-BETA 0.9.4.362

Relevant Pages

  • Re: Trivial performance questions
    ... > system rewrite because the collected work can no longer perform ... > Python code in the last few years. ... > difference between your group and mine is that many of our developers ... > persuade me. ...
    (comp.lang.python)
  • Re: Why no serious MS Application in .NET yet ??
    ... BG> becomes compelling to rewrite any code base, ... becomes too overloaded and new technologies become old and are going to die, ... But I saw projects where developers tried to refactor to some ...
    (microsoft.public.dotnet.framework)
  • Re: Why no serious MS Application in .NET yet ??
    ... BG> becomes compelling to rewrite any code base, ... becomes too overloaded and new technologies become old and are going to die, ... But I saw projects where developers tried to refactor to some ...
    (microsoft.public.dotnet.framework.performance)
  • Re: Why no serious MS Application in .NET yet ??
    ... BG> becomes compelling to rewrite any code base, ... becomes too overloaded and new technologies become old and are going to die, ... But I saw projects where developers tried to refactor to some ...
    (microsoft.public.dotnet.general)
  • Re: Why no serious MS Application in .NET yet ??
    ... BG> becomes compelling to rewrite any code base, ... becomes too overloaded and new technologies become old and are going to die, ... But I saw projects where developers tried to refactor to some ...
    (microsoft.public.dotnet.framework.clr)