I (don' ) will kill my Python

[rturpin at my-deja.com]

In article <mailman.979970182.28582.python-list at python.org>,
  "Tim Peters" <tim.one at home.com> wrote:
> Python was never designed for standalone operation.
> That's life.  What they *should* say is:  "This is my
> program!  I have created it by myself! Look! And it's
> only 507 bytes!  Now you go to http://www.python.org/2.0/
> and click on 'Windows installer'.  Then we can share tiny
> Python programs via email forever after."

This makes good sense. For amateurs exchanging their
experiments.

> Businesspeople have other concerns ..

Distribution is much more an issue for commercial
enterprises, especially those that want to support an
application on thousands (millions?) of clients.  Once
you start thinking about this, you quickly believe
that the sensible thing is to create a new Python
distribution on each platform, regardless of whether
one already exists. (a) You want to know that the
platform was installed correctly. If you just check
for the presence of Python, it might be a partial or
corrupt distribution someone ignorantly copied from a
friend of a friend. (b) You want to make sure that the
distribution is the version against which you have
regression tested the code. Despite the strong efforts
to keep Python backward compatible, anyone who has
been involved in QA of commercial products gets very
paranoid about shipping what is tested. Nothing more.
Nothing less. Nothing different. (c) You want to guard
against someone else on the platform purposely or
accidently changing any of the libraries, before or
after your product is installed.

> .. but they can afford to hire Gordon.

Is he job hunting?

Russell




Sent via Deja.com
http://www.deja.com/