python-dev summary, Jan. 16-31

[Andrew Dalke]

amk:
>The release of 2.1 offers a second calibration on their effectiveness.
>2.1 is the first Python release to have been carried out using PEPs as
>the mechanism, so there are no sizable changes in 2.1 that don't have
>a corresponding PEP.  Yet many people were *surprised* by some of the
>changes in Python 2.1 such as function attributes and nested scopes,
>even though PEPs were written and discussed, often in lengthy threads
>months ago.

I for one am not on the python-dev mailing list.  It's too
much mail traffic for me to follow.  There were about 1,000
messages in January and while the signal level is high,
the content is usually beyond my ken.

(Following the newsgroup is easier because I can defer reading
the newsgroup then kill threads I don't want to follow.  I
suppose I can do that with email as well, by redirecting
incoming messages to a special mailbox, but that's a hassle
and I would have to see if it's possible and learn how to
do it.  Barrier to entry.)

Speaking of the archive, it's always been hard for me to
remember where the archive is located.  There's nothing on
the python.org page pointing to it.  I know it's at sourceforge,
but python.sourceforge.com doesn't work nor does .org.  I
finally figured out it was python.sourceforge.net but I
don't see a link to the back archive.  (Sorry, but my ability
to remember names isn't perfect - that's one reason I like
Python :)

There are also no links from the python.org page to the PEPs
although python.sourceforge.net points directory to it.
More importantly, the PEPs are rarely announced to c.l.py,
which means I don't know what's going on.  When a PEP *is*
announced on c.l.py there are discussions about it, which
implies that others are like me in being interested in
giving input but don't have the time to follow all of the
technical details of python-dev.

I used DejaNews to search for threads named "PEP".  The ones
which came up were:
  PEP 4
  PEP 201
  PEP 203
  PEP 214
  PEP 222 (because of you)
  PEP 223
  PEP 224

Given that there are some 40 PEPs, this implies that only 20%
of them were announced to c.l.py, which greatly reduces the
dissemination of their ideas to the rest of the Python world.

Missing Finished PEPs are: 207, 208, 217, 221, 240 and 232.

I'm looking at PEP 0001, "PEP Purpose and Guidlines". It says:

] PEP authors are responsible for collecting community feedback on a
] PEP before submitting it for review.  A PEP that has not been
] discussed on python-list and python-dev will not be accepted for
] review.  However, wherever possible, long open-ended discussions
] on public mailing lists should be avoided.  A better strategy is
] to encourage public feedback directly to the PEP author, who
] collects and integrates the comments back into the PEP.

This is obviously not being done.  I searched Deja for "PEP 217",
which I first knew about from your python-dev summary.  There
was only a single reference from Tim Peters about it.  Take
PEP 232, which is the one on function attributes.  You said

> Yet many people were *surprised* by some of the
> changes in Python 2.1 such as function attributes and nested scopes,
> even though PEPs were written and discussed, often in lengthy threads
> months ago.

The first mention on c.l.py on function attributes was from Jan. 24
 http://www.deja.com/[ST_rn=ps]/getdoc.xp?AN=719772745&fmt=text
even though, yes, the PEP was created on Dec. 2.

Jeremy Hylton first mentioned PEP 227 ("Statically Nested
Scopes") on Nov. 19 in the thread "Ruby and Python"
 http://www.deja.com/[ST_rn=ps]/getdoc.xp?AN=695547859&fmt=text
then was mentioned again in passing by Tim on Dec. 20.  The PEP
was created on Nov. 1.

No wonder there have been complaints -  how is anyone not
interested in following python-dev supposed to know what is
going on with Python development?

>To me, this makes it crystal clear that the summaries aren't achieving
>their goal of making the development process more transparent to the
>community.

The problem is the summaries are just that - summaries.  I
*really* enjoy reading them because there are a concise description
of what occured and the only way I have these days to know
what is going on with development.

However, consider this latest summary.  I received it on
the 7th, which is 1 to 3 weeks after the discussions it talked
about.  To contribute to a thread, I would need to catch up
on the mailing list to see what arguments have been presented,
then bring up my questions or objections.  But it would be
about two weeks late, which means people may have already
decided on what to do, and making changes is more likely to
be frowned upon.

(Not saying that that will happen, but that's my perception,
which makes it harder for me to decide to put in the effort
to make a useful comment.)


> Perhaps giving the PEPs higher visibility -- posting
>announcements of status changes and new PEPs to
>comp.lang.python.announce, for example -- would do a better job

Just mentioning them on c.l.py as per PEP 0001 would help!

I *really* miss the pre-python-dev newsgroup days.  Spinning
the core technical discussions off has not only reduced my
awareness of Python's direction, but it has almost eliminated
my education in CS topics like continuations.  (Which you
clarified for me at the Houston conference :)  And there
are far fewer posts from Tim Peters than the schriebfest
days.

So I really place the problem on the main python-dev people
who don't bother to disseminate their ideas to the broader
audience of people who develop a lot of Python code but
aren't language designers.

>I feel there's no point in continuing to write them.

I agree with the others who have replied - I will miss the
summaries greatly.  They were nearly the only way I had to
know what was going on with Python development.

>(Should anyone want to volunteer to continue writing them, please do
>so; contact me if you want copies of the previous summaries for a
>complete archive.)

I can't as I don't have the time to follow python-dev, much less
comment anywhere near as intellegently about it as you have.

                    Andrew
                    dalke at acm.org