[python-advocacy] A Few Questions for the Python Community re Survey Response

[Carl Karsten]

I am not answering questions, just discussing how they should be answered :)

The list of questions bothers me.  They seem pretty subjective and not relevant 
to some fuzzy unstated goal.  I know 'we' are not in control of this, so 
bitching won't help fix it, but it may help come up with useful answers.

Useful = someone reading results,  what will help them pick python if python is 
really what they should use?

Jeff Rush wrote:
> The official Forrester Research survey form on dynamic languages has arrived
> and as advocacy coordinator, I'm starting to fill it in.  Here are some
> questions on which I could use some input.  Please reply via private email, to
> avoid cluttering the lists and cross-posting issues, and I'll collate all
> responses.
> 
> *Note* that we want the survey to address cPython, IronPython and Jython, so
> add caveats as necessary to your answers.  And yes, some of the questions are
> vague and difficult to answer precisely - let's do our best and remember the
> objective is to provide practical answers that would be useful to a
> manager/developer choosing a dynamic language, not to debate terms in an
> academic sense.
> 
> 1) Can a developer extend the language by adding new keywords?
> 
> My answer to this is no, short of recompiling the underlying parser -- with
> the new AST engine of 2.5, am I wrong?  Can Jython or IronPython do it?
> 

Technically, 'yes'.  we have the source, and we can propose additions.   "no" 
would pretty much mean the language is closed and frozen.  OTOH, is 'yes' a good 
thing?  I think not.  keywords are not scoped to a namespace.  python's limited 
keyword set is powerful enough to do what we need to do.  So if you assume there 
is an implied "as easy as adding functions" I would agree with 'no.'

That was not an answer because it is conditional  :)

> 
> 2) How large is the on disk runtime that has to be installed to run
>    programs written in this language?
> 
> I don't run either Jython or IronPython -- can anyone who does provide an
> answer?  And yes, it is a bit unclear as in whether it includes all of Java
> and .NET or just the delta above those.  Go with the delta, it's smaller.

For .net - "has to be installed" is 0, given that "it comes with vista"

Same with the mac.

> 
> 
> 3) What applications are available to developers using this language
>    that support application lifecycle for requirements, design,
>    development, build, and test?

Typical problem with English: to qualify, does an app need any or all of the 
features?  If all, does such a thing exist for any language?  and in the context 
of this survey, it seems it would have to be exclusive to one language.

I just pinged someone who worked for Motorola and now works at a 'big company' 
where there are about 1000 programmers (mostly java, some python, perl, etc.) 
who has never heard of anything that covered all that.  this came close:
http://www.smartbearsoftware.com

So if any, write small, cuz it is gong to be a huge list.  be sure to include 
anything that comes close: trac,  subversion, mambo/joomla + plugins? this is dumb.

Maybe just listing off the popular IDEs?

> 
> 
> 4) How many new major and minor features were added in the last release?
> 
> Let's assume they mean the last major release.

damm English again.  do they want one or 2 numbers?  who decides major vs minor?

Are small or large numbers favorable?

> 
> 
> 5) How many committers to the core does the project have relative to the
>    scope of the project?
> 
> Hmm, how to measure the "scope of the project"?  Lines of code in cPython,
> IronPython and Jython?  A rough guess of the class/module count?
> 

Are small or large numbers favorable?

> 
> 6) How many bug reports have there been for the latest release?
> 

ditto.

> 
> 7) How many email postings have there been (including developer and
>    user lists) in the past six months?

1.3 gazillion.

> 
> 
> 8) How much communication occurs within the community? 

lots.

 > Besides email
>    postings, what other forms of communication occur within the
>    community? What forms of collaboration are used?

All of them.

> 
> 
> 9) How complete is the language and framework support for Web service
>    standards?
> 
>    Does the language support Core Web Services?
>    Does the language supports Web Service Security standards?
>    Does the language supports Web Services Management standards?
>    Does the language provides Registry and Metadata support?
>    Does the language support Web service Process and Delivery
>    Control standards?
>    Does the language support Web services Transactions and
>    Packaging standards?
> 
> Not being a web services kind of developer, I need some help here.

I know just enough to say: Someone is buz-word happy - Web Services are not that 
complicated.  you can 'do' all that stuff using python, even if you had to roll 
your own, and you don't.

http://www.diveintopython.org/http_web_services/index.html Copyright © 2000 
(what year are we in now?)


> 
> 10) How many companies offer end user support for the product? Is there
>     one company primarily identified with the project that provides
>     end customer support?
> 
> Definitely not a single company, like Sun for Java, at least for cPython and
> Jython.  Is Microsoft the primary company providing end-user support for
> IronPython though?  Many of these companies may be members of the PSF, so
> speak up.

Another dumb question.  given that no language (or product) can have some 
restriction of who can charge for help with it, there is no way to track how 
many companies are making a buck off of it.  the most limited I can imagine 
would be something like C#: MS, divisions of MS, MS certified professionals (but 
not MS at all) and people who are just good at it and offer training/support.

They might have meant: how many companies are at the tho of the tree, but they 
didn't say that, and who cares?

Maybe there are languages where the number is under 100, like some sf.net 
project that claims to be a language, or maybe BrainFk that isn't really a used 
language, so no one would need to be supported.

So back to: what would be a favorable answer?  is 1,000 enough?  It isn't like 
we need to offer support.

> 
> 
> 11) How many systems integrators use the language in application development?
> 
> Considering Python is a secret weapon for many companies, this will be hard.
> But can anyone provide some anecdotal information?

considering there is no way to know the magnitude for any language, I am 
starting to think you should just forget this whole thing in protest.

wtf is a "systems integrator" and an "application" anyway?  to me it means "how 
many people use python?"

I vote for 1,000,000. +/- 1,000%

> 
> 
> Thank you very much for your time and effort,

Sorry to waste your time with my ranting :)

Carl K