Python - open forever ?

[Peter Hansen]

Romans Krjukovs wrote:

> To make this reality we have to include Python in our strategy
> plans for at least 3 nearest years.
> This would be a green light to our developers to start new projects on
> Python.
> 
> At the presentation of Python and its usage posiibilities in our 
> home environment to the members of the project managers board,
> we got such questions: 
> - There is a risk that Python can become closed and unsupported in the
> future.
>   (Remember RedHat ?) 

I remember RedHat.  In fact, I can still download it, and install and
use it.  For free.  And get the source code.  And get support (if I'm
willing to pay, of course, but it would be ludicrous for a company to
insist that free support is a necessity for a given technology).
What is "closed and unsupported" about it?

> - Who can guarantee that Python will be usable and available  to us if
> it is develeped and maintained by the hackers from all over the world
> without any obligations and guarantees ?

Python has been usable and available for the last thirteen
years, but perhaps the last thirteen years have been atypical...

> - How we can minimize such risk ? (Become a member of some club,
>   buy licenses, support etc.)

You can read the Python license and, if you really insist, consult
with your lawyers, whom you pay to give legal and strategic advice of
this nature.  Then you will understand that the nature of the Python
license is such that these silly unreliable hackers, who won't guarantee
anything, cannot prevent you from using Python and taking the source
code and doing what you will with it, for the most part.  But I'm not a
lawyer...

Note that if you have money to spend, you can certainly pay various
companies for Python support.  Of course, that won't provide you
with any guarantees either, but perhaps it will be costly enough to
make your management believe so....

If you don't want to spend the money, it's a little unclear why you
would whine about obligations and guarantees... the economy doesn't
really work that way, even in the Open Source movement.

> Python is very fast in development, stable and fast code, easy to learn,
> but never the less big and business critical project can't be started
> without mentioned risk analyses.

> It would be nice to know what Python society members think about this.

Basically this: if you are going to use Python in "big and business
critical projects", you want stability.  To achieve stability, among
other things you want to chose one version of Python and stick with it
for a very long time, rather than consantly upgrading with each new
release (IMHO).  Given that strategy, you need only assure yourself that
the license on the version of Python that you adopt, and on any third-
party modules you use, allows you to use it in the ways you want to use
it, including getting the source code and even redistributing it if you
wish.

At my previous employer, a 100+ employee telecom in Toronto, we chose
Linux (RedHat, even) and Python as the basis for a range of products.
The extent of the risk assessment required was to verify that the
licenses allowed us to do the development without the risk that someone
(other than SCO, say) would constrain us from continuing to use it.
We didn't expect, or look for, any guarantees that all future versions
would forever be completely free and unconstrained.  Again, pointing
to SCO, there _are_ no guarantees...

-Peter