Changing the repr() of a class?

[Alex Martelli]

Steve Holden wrote:
        ...
> I'm sure Alex's book is going to be an excellent contribution to the
> language and its growth. While I was disappointed that it had to line up
> behind "The Python Cookbook", it will probably be better for the wait. (I

We'll see -- I _am_ ensuring it covers 2.2, of course.

> can imagine Alex gnashing his teeth whenever new features requiring
> changes are introduced, though. When you are writing, sometimes you just
> want to shout "Stop!" at the developers ;-)

Actually, as a writer, it's to my net potential financial advantage if the
technology I'm covering changes (as long as the rate of change does not
reduce the acceptance of the technology in the market) -- this gives
me the "upgrade" after-market.  There are technologies which I
needed to track for which I've bought two, three, up to four such
"upgrade" books in a few years.  For the TCP/IP fundamentals, I've
made do with one book throughout my career -- one day or another,
I'll have to buy another for IPv6, but it won't be an "upgrade" of that
dog-eared book which served me so well.

My problems with the prospective of continuing, frequent changes to
the language come from other issues -- basically that innocuous
looking little parenthesis in the previous paragraph.  A rate of change
perceived as "too fast" may well inhibit the language's uptake.  Quite
apart from book authors' financial prospects, I feel it would be a pity
for the world at large if the benefits that Python can bring were to be
less widespread because of such effects.


Alex