Python Is Really Middleware

Tim Daneliuk tundra at
Wed Aug 1 01:20:02 CEST 2001

Terry Reedy wrote:
> As a Smithian hippie, I think the better adage is "Entertainment
> information want to be free", which is to say that people like to
> share non-critical information as part of their social interaction and
> play.  

That "entertainment information" is still somebody's property and 
deserves the same protections as my house, car, or other material
possessions.  The fact that the medium is malleable doesn't change
the importance of property rights - unless, of course, the property
owner chooses to make it freely available voluntarily.

> On the other hand, several people have noted that Python being
> free is a barrier to commercial acceptance.  As a free user, I would

I see no reason to believe this.  Perl, Java, Linux, FreeBSD all
have significant commercial followings and all are freely available.

I think the more germane issue is "how much wood is there behind
the technology arrow?"  That is, to what extent to commercial
enterprises believe that a new (to them) technology has viability,
a large user community, and thus, support.  The best form of
Python evangelism I know, it to solve real commercial problems
with it better/faster than the other choices.  Running code always
wins an argument in front of the CIO.

> like to see more commercial exploitation without enclosure.  (Software
> and information in general does *NOT* suffer from the tragedy of the
> commons.)  One selling point 'should' be that the 100,000 (or

It depends on what you mean by "enclosure".  Per the point above,
organizations, like people, need to enforce their property rights.
If a particular technology makes this impossible or very hard,
that technology will be met by huge opposition.  I once was in
the position of having to authorize buying Sun's 'C' compiler
rather than using the GPLed GCC for exactly this reason.  Even though
I, and all the engineers who worked for me, preferred GCC, I could
not in good conscience expose my company to the possibilty  of
being forced to reveal the inner workings of a product that
was our key competetive distinction in the market we served.
This sort of thing is not unusual.

This, BTW, is why the GPL is under attack and people are looking
for alternatives.  People are looking for a balance
between preserving the open source intentions of the original
authors and protecting their own intellectual property.  Personally,
I like the BSD license, as do a great many other people.

> > I, for one, would like to see a date picked for a permanent
> moratorium
> > on the language proper,
> All the changes in the last year have been discussed for years.  So I
> expect/hope that in another year we will see something of a plateau
> again with Python2final.

I understand this.  You'll notice that I was careful to comment
on the "instinct" of the Python community towards more features,
but I was not being critical at all.  There are still legitimate
things to be done and I'm not carping about them.  I just worry
that Python is going to go down the same ratholes I've seen before.

> I think it worth noting that 2.2 without 'import future' is arguably
> more compatible with 2.1 than most versions of, for instance, 'C' have
> been with each other.  ['Version of  C' = defacto language of
> particular compiler on  particular platform with particular settings.]
> In this sense alone, you are correct: Python is middleware that has
> already successfully hidden some annoying variations in the underlying
> platforms it runs on.

Quite true and well put.

Tim Daneliuk
tundra at

More information about the Python-list mailing list