Fast CGI Vs Java Application Servers

Ian Bicking ianb at colorstudy.com
Sun Jun 29 15:08:28 EDT 2003


On Sun, 2003-06-29 at 08:05, Kevin Kenny wrote:
> Charles Handy wrote:
> > 
> > How does FastCGI compare against java Apps running in java app servers
> > like TomCat, Sun One, WebLogic and WebSphere?  Is there a business case
> > for switching from CGI to Java?  Performance? Productivity? Features?
> > Object orientation? Code Reuse?
> > 
> > Any opinions?
> 
> When anyone talks about scrapping working code, I refer them to
> http://www.joelonsoftware.com/articles/fog0000000069.html
> Generally speaking, there's no good reason to ditch stuff that
> works.

Sure there is -- you are doing something considerably different than
what you were doing before, and you are deciding whether to adapt what
you have or reimplement it in the context of your larger application
(though reimplementation and adaptation can be largely the same thing,
depending on what changes you are making).

I frequently see people who think they have code that's valuable, when
it isn't -- because maybe it was written with a difficult decision
process, or by someone who wasn't familiar with the domain, so it was a
struggle to write the first time, even though it could be reimplemented
in a matter of days.

I also see people who think one application is a close enough match to
another application that a different customer requires, that it makes
sense to just adapt the application.  

Sure, Netscape is an example of an application that probably shouldn't
have been rewritten.  But there is a qualitative difference between a
huge program like Netscape, and some random web app.  Those are
radically different environments, and the amount of investment is
probably off by two orders of magnitude.

There's always a breaking point when reimplementation makes sense.  If
not we'd all be programming FORTRAN.  (Though I would agree that you
should seek to reimplement as small a piece as possible at any one
point)

  Ian







More information about the Python-list mailing list