Nice presentation of Dana Moore on Oscon

[Gerrit Muller]

A colleage attended me on the presentation of Dana Moore on Oscon:
"The Seven Habits of Highly Effective Technology Disruption
Python Survival in a Java & .Net World"
http://www.infoether.com/~rich/pycon2003.pdf

A very nice presentation. I do have one small comment: Python is
characterized as "weakly typed", which should be "dynamic types", or
what people often mean to say "not static typed".

The difference has been discussed many times in this newsgroup. Dynamic
typing requires a different approach than static typing. Programmers
from the static world have to make a mental adjustment to become
effective in this different paradigm.

kind regards, Gerrit Muller

 

[Aahz]

A very nice presentation. I do have one small comment: Python is
characterized as "weakly typed", which should be "dynamic types", or
what people often mean to say "not static typed".

Unfortunately, Bruce Eckel has been perpetuating that terminology. I'm
in the process of writing a rebuttal.

http://www.artima.com/intv/typing.html

 

[Martin Maney]

Unfortunately, Bruce Eckel has been perpetuating that terminology. I'm
in the process of writing a rebuttal.

http://www.artima.com/intv/typing.html

In this it sounds as though Mr. Eckel is coming around to the right view.
In the first part he seems to be using "latent", and only says at one
point "...latent typing, sometimes called weak typing...". And he
repeatedly uses the phrase "strong static type checking" for what is
supposed to be the strength of C++ and Java, and notes that "Strong
static type checking forces the programmer to do a lot of extra work."

So I get the impression that he is making that journey of discovery
that we know so well. Or, anyway, I know it - maybe you didn't go
through the same wringer that made B&D static type checking seem like a
pretty good idea in contrast to what came before? I used to think C++
was the best thing since the symbolic assembler, or maybe object

linking...

I don't know that I care much for "latent typing", although it has a
certain attraction.  If nothing else, it's shorter than "strong dynamic
typing", the first word of which seems necessary these days.  :-(

But whether or not Mr. Eckel is still contributing to it, I can attest
that there is a general confusion about this among programmers, perhaps
especially java programmers.  Just the other day I whacked one upside
the head (virtually; we were chatting on IRC) with the distinction
between "strong" and "static" typing and had the pleasure of watching
his eyes light up.  Well, I imagine they lit up.  But that was only
one, so if you'd get cracking on that paper...  :-)

 

[Aahz]

Unfortunately, Bruce Eckel has been perpetuating that terminology. I'm
in the process of writing a rebuttal.

http://www.artima.com/intv/typing.html

All right, my rebuttal is now at
http://www.artima.com/weblogs/viewpost.jsp?thread=7590

 

[Edward K. Ream]

Many thanks for your rebuttal.

My picture of the situation is this: the Python interpreter is in effect
placing assert statements after (inside) every Python statement. This makes
Python in practice much more safe and convenient than any static language.

No, this doesn't directly relate to types. However, IMO the real issues
aren't about types, they are about safety and convenience. Python crushes
any static language as far as convenience goes. This much is obvious.

Theoretically all languages are unsafe. Python programs can throw uncaught
exceptions just like C++ asserts can fail. In practice, though, the Python
interpreter makes development so much safer that one can just "blast away"
without worrying too much about the kinds of things compilers typically
worry about. And my Python programs have been much more robust than the
equivalent C++ programs, even without pychecker's help.

This is probably all obvious to Pythonists, and incomprehensible to others.
Oh well...

Edward