[python-advocacy] Proposal for Monthly podcast series

[Paul Boddie]

On Sunday 17 June 2007 18:39, Roy Smith wrote:
> >> * Django (Turbgears and Pylons are both great projects and *equally*
> >> deserving of coverage - but would this be too much of a web-app focus)
> >> Oh, and there is always Zope. ;-)
> >
> > How about a 2nd 'track' that is just web technologies?
>
> If there is one thing Python is getting killed on, it's web apps.
> Specifically, Ruby on Rails.  We need to work on our image in this area,
> so I'm not worried if we focus too much on that.

This is an issue of perception, however - that the Rails people are good at 
making lots of noise, while things like Plone seem to be one of the first 
choices in the content management space, even though you probably aren't 
hearing about it. I don't think it helps that despite YouTube apparently 
using a lot of Python, for example, nobody wants to talk about it, and 
there's a continual uncertainty as to whether they really do use Python or 
not. Clearly Python delivers, but the really high profile adopters aren't 
contributing to the advocacy effort.

I think that this has always been a problem in Python advocacy: the whole 
"Python is a secret weapon" attitude which means that as long as you're 
working in an environment where Python is established, then you'll be fine; 
otherwise you'll have a hard time pointing to success stories (with all the 
accompanying "buzz") that your boss will find relevant enough. And when the 
success stories just tell you that "Python is great for our internal 
proprietary systems" then it doesn't stand up very well to something directly 
relevant to such systems that you can actually download - the success story 
could be talking about a million monkeys doing the work for all the benefit 
the reader derives from it.

(It sounds like I'm criticising the success stories on python.org here, but I 
actually think that they're an underrated and underpromoted resource. 
Nevertheless, just as a number of scientific papers I've read recently have 
been somewhat short on the details, there's nothing like an actual recipe for 
success rather than a description of such success, even though most companies 
aren't likely to openly share the former.)

> I've been pushing in my company for us to develop a Python scripting
> interface for our product.  We already have both a home-grown scripting
> language, and Perl.  There is wide-spread agreement that neither is
> satisfactory.  I've been pushing Python, but it's been an uphill battle
> against the Ruby mind-share.  Here's a quote from a recent discussion:
>
> > Python seems to be one of those specialty languages that builds up a
> > following but never makes it into the mainstream.  That doesn't make
> > it bad - Lisp, APL, SNOBOL, and Icon are other examples that come to
> > mind.  In fact, these language often are very well engineered and
> > their followers appreciate that quality.
> >
> > Ruby seems to be gaining rapidly increasing "mind share" these days.
> > Partly this is through the wide use of Ruby On Rails as a Web
> > development environment.
>
> Notice the point being made.  It's not about whether Python or Ruby would
> be a better language to write the product in, but which would be perceived
> by our customers as being more desirable.  And that perception is being
> driven by Rails.  That's what we need to be fighting.

Once upon a time, Python advocacy was all about convincing the out-of-touch 
manager about including Python in the set of technologies mandated for any 
given project. Then someone pointed out that the real driving force was peer 
promotion. Either way, if someone is talking from an academic viewpoint about 
a technology then that viewpoint is likely to be driven by perception rather 
than hard fact. Is your company's product a Web application? If not, why 
would Ruby on Rails be in any way relevant apart from the presence in the job 
market of an unmeasured number of developers having seen Ruby before? In any 
case, would a Ruby scripting interface as opposed to a Python one really 
bring in more developers?

How do you counter such perceptions? You need to show that Python is almost 
ubiquitous, that Python is directly relevant to the domain, and that Python 
is a favoured choice when making scripting interfaces to applications. To 
avoid the "million monkeys" phenomenon, you also need to be able to point to 
concrete work that shows that you aren't going to be reinventing someone 
else's wheel just because they enthused about how round it was.

I'd like to restate how important resources such as python.org should be in 
delivering materials suitable for convincing both managers and peers about 
the merits of Python and the high likelihood that someone has at least 
touched upon any given area of endeavour using Python previously. The 
python.org Wiki has gone a long way in the direction of providing such 
materials, and Jeff's initiatives are complementary to the reference 
information available on the Wiki. Specifically relevant to the above case, 
you'll find information on the Wiki about bindings generators and 
technologies, along with a white paper on the topic (eventually).

It's just unfortunate that a lot of useful material is often buried in places 
that are not so easy to find, and that the wider Python community doesn't 
feature as much in many of the lists of projects that the "inner community" 
regards as the most important. Returning to those lists: stuff like 
setuptools is extremely high profile already; stuff like Freevo (and related 
projects) are often ignored. I know what I'd rather be hearing about.

Paul

P.S. I don't think advocacy-related issues are on the schedule for EuroPython, 
but if there's interest, I can imagine setting up some kind of Open Space for 
interested parties. Speak up on the EuroPython mailing list or on this list 
if this sounds like a good idea!