jos van de ven

==
4ePost: 1,534 bytes in mail. Elapsed time to buffer: .000078 sec.

albert van der horst

I would except C# to be somewhere on the list? At least it is
better known than idl, I reckon?

By the way, interesting read, whatever they say about it.


--

--
Albert van der Horst, UTRECHT,THE NETHERLANDS
Economic growth -- like all pyramid schemes -- ultimately falters.
albert@spenarnc.xs4all.nl http://home.hccnet.nl/a.w.m.van.der.horst

anton@mips.complang.tuwien.ac.at (anton

Albert van der Horst <albert@spenarnc.xs4all.nl> writes:


There is no comp.lang.* group on C#. There may be C# groups in other
hierarchies (msn.*? alt.*?), but I'm not going to chase after them
(the numbers for different hierarchies may be less comparable than
those from the Big 8, because auf different cultures, differences in
spamcanceling etc.).


If that's the IDL that has something to do with CORBA, I guess that
the rank does not reflect its popularity, either (there maybe more
discussion about it in a group on CORBA).

- anton
--
M. Anton Ertl http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
New standard: http://www.complang.tuwien.ac.at/for...forth200x.html

anton@mips.complang.tuwien.ac.at (anton

Marc Olschok <invalid@nowhere.com> writes:


True. But does it matter? What would it tell us that the 172
postings per day in comp.lang.python come from 100 people? Or 1000?
Also, my experience with Usenet groups is that all of them have some
heavy posters, and many light posters (maybe a Pareto distribution);

IIRC Netscan <http://netscan.research.microsoft.com> did have such
data, but it seems not to work right now.


True. But does it matter? What would it tell us? One obvious
interpretation of a short thread is "quickly dispatched homework
question", one for a long thread is "flame war". Constructive
discussions can be either short or long.


Right. That matters, and different groups may be quite different in
this respect.


Numbers of postings as provided by the nntp server.

Right.


Maybe. I am more worried about the Java groups, though. If there are
lots of crosspostings between the various Java groups, the total
number of different postings may be less than for Python.

- anton
--
M. Anton Ertl http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
New standard: http://www.complang.tuwien.ac.at/for...forth200x.html

john passaniti

Anton's list measures the number of messages in comp.lang newsgroups for
a language. Since there isn't a comp.lang newsgroup for C#, it isn't
surprising it isn't in Anton's list.

Microsoft has their own support and discussion newsgroups for C#, of
course. But the large (and growing) community using open source
implementations of the language and .NET environment (like Mono) also
have web-based discussion forums, code exchange sites, technical blogs,
and mailing lists.

C# isn't the only language that doesn't have a comp.lang newsgroup.
Another very popular language (especially with game programmers) is Lua.
The Lua community prefers instead to center around a mailing list and
a wiki. If you visit http://lua-users.org/lists/lua-l/ you'll see an
archive of the Lua mailing list, with message counts per month. There
were 717 messages last month, which put is somewhere between
"functional" and "awk" in Anton's list.

I personally find that I am spending far less time in newsgroups than I
did in the past-- and not just comp.lang.forth, but various other
newsgroups I participate in. I find that more-focused mailing lists
offer more signal than noise, and I find that wiki's offer an immediate
and direct sense of community and collaboration that I don't get in
newsgroups.

And when you couple all this with search engines, newsgroups become less
attractive for people trying to find information (verses just open-ended
discussion).

So, I don't really believe Anton's programming language popularity has
much relevance, and what relevance it may have is diminishing over time
as other forms of communication and collaboration are taking over.

anton@mips.complang.tuwien.ac.at (anton

Andrew Haley <andrew29@littlepinkcloud.invalid> writes:

I can agree with that. Wow:-).


Why not? BTW, just in case you missed it, this is not a reposting,
but an update (new data).

If they are all bugus, how do you explain the positive correlation
between them? If the common factor is not popularity, what is it? Or
do you think that they agree just by coincidence?

You would find an improperly conducted survey more far more credible?

If you find the properly conducted one, let me know. I don't know how
they would be able to achieve that.

- anton
--
M. Anton Ertl http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
New standard: http://www.complang.tuwien.ac.at/for...forth200x.html

anton@mips.complang.tuwien.ac.at (anton

John Passaniti <nntp@JapanIsShinto.com> writes:


Actually, since the number of postings I recorded is from a 100-day
period, the number for the Lua list would be around 2600, i.e.,
between Scheme and Smalltalk.

- anton
--
M. Anton Ertl http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
New standard: http://www.complang.tuwien.ac.at/for...forth200x.html

andrew haley

Because it's nonsense, that's why. And it's also off-topic. And it
in no way contributes to the debate in this group. The only effect it
might have -- apart from prompting arguments like this one -- is to
promote yet another bout of hand-wringing angst over Forth's "lack of popularity".

Oh, that's OK then.

How could I? I haven't seen these other data.

I don't know. As I said, I haven't seen them.

You are the one who says that these numbers do not disagree too much
with other ways of estimating programming language popularity. OK, if
these other ways, whatever they are, are based on a methodological
foundation more substantial that counting Usenet postings, why not post them instead?

Of course not. That sentence, read in a straightforward way, suggests
that I would find a properly conducted survey far more credible than
this bizarre numerology.

Andrew.

stephenxxx@mpeforth.com (stephen

Anton is posting numbers and the context in which they occurred.
Nothing else. Give him a break. It isn't hand-waving, it's a
measurement. What of is a different matter.

Both Elizabeth and I know that most of our clients do not use
newsgroups.

Stephen

--
Stephen Pelc, stephenXXX@mpeforth.com
MicroProcessor Engineering Ltd - More Real, Less Time
133 Hill Lane, Southampton SO15 5AF, England
tel: +44 (0)23 8063 1441, fax: +44 (0)23 8033 9691
web: http://www.mpeforth.com - free VFX Forth downloads

yossi kreinin

In short, Python is an infix LISP. The syntax and semantics are very
uniform - numbers and strings, some syntactic sugar for lists, and most
complex things work using calls and dereferencing
(module.obj.member.method(args).member2...).

Some things I like about Python:

typing and simple generic data structures are supported, and the
uniform grammar makes it easy to mentally break the code into
statements.

* it's easy to write, since most simple things are easy to do, and it
compiles fast and can be tested interactively. Dynamic languages which
are equally easy to write are usually harder to read.

* it's very realiable - almost never fails without explicit apology.
Exception handling and garbage collection work nice together. It's
fairly easy to recover from errors and write robust programs. I think
that "undefined behaviour" such as overwriting memory is very barbaric
- much like getting a knife in your back for not greeting a person
properly (it happens in undeveloped societies). I don't understand how
did I ever allow to brainwash myself into accepting this on everyday
basis.

* it supports metaprogramming very well. it's easy to load or generate
code at runtime, traverse information about the program, traverse and
serialize data structures, and play with the binding of words.

* it's "real" - it's easy to write large projects in it. it has a
simple and good module system, there are few compilation dependencies
because of dynamic binding, and it's easy to read code written by other
people.

* it has a lot of very good and well documented libraries.

* it's fairly portable since there's a lot libraries abstracting an OS,
and a single implementation. C/C++ programs, for instance, are hard to
port, since they become dependent on byte ordering, system calls, and
the dialect recognized by the compiler as time goes by. Of course you
can always use os.system() or the like to break portability.

* most programmers pick it up really fast because it resembles the
mainstream languages, and is easier then those. So it's easy to promote
it in an organization, unlike many good languages that look very
different from today's mainstream.

Some things I don't like:

* Lack of static typing (c.l.f readers are not very likely to identify
with this...). It limits performance, so you need C/C++ connectivity to
handle bottlenecks, and I don't like these languages very much (but
there's a dialect called Pyrex which AFAIK tries to deal with it).
Programs failing because of stupid errors at runtime are also annoying
(zealots claim programs should be tested, but the annoying thing is
that *the test* runs for 5 minutes and fails - not much better then
spending these 5 minutes while waiting for a C++ test to compile, or
debugging core dumps in C).

* It can't be used in embedded systems. That's a pity. I'd like to have
a fast, elegant and reliable language for embedded systems. It's a hard
problem, though.

* They break things between releases, and in general I'm sort of at
Guido Van Rossum's mercy when it comes to my Python code. A bloated,
obscure and slowly changing standard can be worse, though.

* Minor annoying syntactic issues - Python has an order of magnitude
fewer of these then other languages I know.