python a bust?

[Andrew Dalke]

Brandon J. Van Every:

Apple almost went under and is a decided minority computing platform

today.

I work in strange subfields by general computing standards:

Up until a year or two ago, computational chemisty was dominated by
IRIX. Linux has mostly replaced, except for some very high end
visualization (game graphics cards still have problems with lots of
triangle - they prefer textures, and the drivers for stereo displays are
still poor, esp. for stereo-in-a-window support). Though non-
computational chemisty (eg, compound registration and database
searches) is done on MS Windows, which has some quite excellent
ActiveX plug-ins for compound display. Still, one of my clients
has an SGI Octane *and* an NT box on the comp. chemists'
desktops.

Bioinformatics used to be Solaris for the server and Linux for
the development machines, but OS X has made suprising in-roads
for developer machines. (That's what I use.) Linux is replacing
Solaris for many of the servers, excepting some high-end ones
(large memory, many processors). Most of the bioinformatics
apps are web-based and don't require plug-ins so are quite
portable.

At the last bioinformatics open-source software developer meeting
I went to, everyone was running either Linux or OS X. People
booted into MS Windows mostly to check comptability.

And of course Python works wonderfully on all those platforms.

Andrew
(e-mail address removed)

 

[Andrew Dalke]

Brandon J. Van Every:

You've gotta be kidding me. Even 12 years ago, "Computer Graphics:
Principles and Practice" didn't teach texture mapping. Along came DOOM.

============
I have a 2nd ed. copy of Foley&van Dam (&Feiner & Hughes) which
has that title and is (c) 1990. I used for a class in that year. That's
13 years ago.

In the index under "Texture mapping" -- "See Surface Detail"
In the index under "Surface detail" -- references to pp 741-745

============ (all typos mine)
16.3 SURFACE DETAIL

Applying and of the shading models we have descrive so far to planar or
bicubic surfaces produces smooth, uniform surfaces -- in marked contrast
to most of the surface we see and feel. We discuss next a variety of
methods developed to simulate this missing surface detail.

16.3.1 Surface-Detail Polygons
...
16.3.2 Texture Mapping

As detail becomes finer and more intricate, explicit modeling with
polygons or other geometric primitives becomes less practical. An
alternative is to map an image ... a technique pioneered by Catmull
[CATM74b] and refined by Blinn and Newell [BLIN76]. This
approach is known as texture mapping or pattern mapping ...
...
The approach just described assumes square pixel geometry and
simple box filtering. It also fails to take into account pixels that
map to only part of a surface. Fiebush, Levoy, and Cook
[FEIB80] address these problems for texture-mapping polygons.
.... [It] can be quite inefficient [and we discuss other approaches
in] Section 17.4.3. Catumull and Smith's efficient technique [CATM80]
for mapping an entire texture map directly to a surface is discussed
in Excercise 17.10. Heckberg [HECK86] provides a thorough
survey of texture-mapping methods.

16.3.3 Bump Mapping
...
17.4.2 Other Pattern Mapping Techniques

[discussion of mip maps]

============

So it was discussed, with an overview of how it works and
the different approaches and pointers to literature references
for more info.

(Nit-picker? Me? Nahhh.

In addition, SGIs in 1990 could be bought with texture
mapping hardware, or emulated texture maps in software.
That was the 'VGXT' (or something like that) naming scheme.
"Vertex / Graph / ....? / Texture", depending on what was
done in hardware. As I recall -- memory fading after all
these years and I never was much of a hardware guy.

Andrew
(e-mail address removed)

 

[Michele Simionato]

Michele Simionato:

Try also bookpool.com for technical books. PiaN is 43% off. With
3-4 day UPS ground it's $24.12, which for a book with list price of
$34.95 means it's 31% off. And you can get cheaper shipping if you
are willing to wait longer.

Andrew
(e-mail address removed)

Unfortunately I am in Italy now, and we don't have Amazon.it or
something equivalent yet (I am hoping I am wrong ...)

Michele

 

[Alex Martelli]

Michele Simionato wrote:
...

Unfortunately I am in Italy now, and we don't have Amazon.it or
something equivalent yet (I am hoping I am wrong ...)

No, you're not. We do have place such as gorilla.it for online books. Not
particularly good.

But this (theoretically freemarket...) government passed a law forbidding
discounts of more than 15% on books, basically to impede the chances of
supermarket chains (such as Coop) and large bookstore chains (such as
Feltrinelli), which happen to be mostly left-ish wing, to compete with
small independent bookstores, which happen to be mostly right-ish wing. As
a side effect the chance of there ever being an amazon.it was destroyed.

Try amazon.de, .fr or .uk if you're in a hurry -- I normally still order at
amazon.com, wait a while, but get a bargain despite mail costs, and good
and wide choice too.


Alex

 

[A.M. Kuchling]

At the last bioinformatics open-source software developer meeting
I went to, everyone was running either Linux or OS X. People
booted into MS Windows mostly to check comptability.

Similar anecdotal evidence indicates the same is true for astronomers, at
least the astronomers my SO hangs out with. A friend also recently told me
of a family reunion where pretty much everyone had an Apple laptop except
him.

It makes sense -- these are technical people who want to do various
demanding jobs that can use a Unix infrastructure, but don't want to have to
wrestle Linux configuration into submission.

--amk

 

[Anand Pillai]

I think my O.P has been taken out of context here.

Python is a great language, it is flexible, agile,
great for newbies, good for experts at the samt time,
allows for prototyping blah blah... I am sure everyone
in this group will tend to agree with most of these
oft repeated statements.

My question was that, "If it is such a great language, why
it is not getting the recognition it deserves"?

Probably there is a need to change the attitude of Pythonistas
about their language. It needs to probably come out from that
"high pedestal" where books get written only if there is a definite
need percieved. One way to do that is may be, just may be to
actually create market for new python books.

The "right" way to do this is by synergy. First we need to popularize
the language, get the website in right shape, evangelize it,
get it accepted by the Suits, then it will get accepted by the Suits
as their language of choice in s/w projects and in due course
every Java or C++ guy will be buying new Python books from the
bookstall. Ah, but this topic often gets discussed again and again
and I was wondering if there is a tendency to come full circle
starting from the book topic, without really achieving anything.

In fact, I have been commisioned to write a series of articles on
Python in a popular computer magazine in India. I am trying to do what
I can to evangelize the language in and around where I live, as I
percieve potential in it, and perhaps potential for me to grow with
the language.

There is no way AFAIK any index to measure the popularity of a
language
by the plethora of books written on it. But it is common sense that
if there are many authors writing books on a language it has got to be
popular. Doesn't popularity mean the number of people interested in
a certain thing? So if 'n' guys are writing books on Python and
'n*100' guys on C++, I would say, not as a techie, but as an average
person, that C++ is more popular than Python.

There was a company called Wrox writing P2P (Programmer 2 Programmer)
books
on many technical topics. I think it got absorbed by Wiley recently.
Their choice of topics was "Popular" languages and technologies used
by
practicing software developers. I never saw a Python book in their
stable.
Of course, again dont start a thread about the technical perfection of
Wrox
books, since I know that their style & content cannot be compared with
an O'reilly nutshell book. But as again, I am not talking about
anything regarding
the "greatness" of the language, but just simple arithmetic.

-Anand

 

[Alex Martelli]

Anand Pillai wrote:
...

There was a company called Wrox writing P2P (Programmer 2 Programmer)
books on many technical topics. I think it got absorbed by Wiley recently.

They had gone bust, and the brand has been purchased.

Their choice of topics was "Popular" languages and technologies used
by practicing software developers. I never saw a Python book in their
stable.
Of course, again dont start a thread about the technical perfection of
Wrox books, since I know that their style & content cannot be compared
with an O'reilly nutshell book.

You're wrong: some of their books were truly excellent. Short of Don Box's
Addison-Wesley bible, they had the best on COM, and the best in particular
on ATL, the best way to do COM in Visual C++. Of course, many others
(while still technically OK) were "me too"'s in crowded fields.

But as again, I am not talking about
anything regarding
the "greatness" of the language, but just simple arithmetic.

Simple arithmetic tells us Wrox went bust (despite having some excellent
books and a very vast selection). I'm not sure how you plan to use this to
convince other publishers to put out plenty of "me too's" technically decent
but mostly undistinguished books in crowded fields.


Alex

 

[Alex Martelli]

Michele Simionato wrote:
...

the world is. So, I do promote Python, but the pacific way

Michele

P.S. in Italian "pacific" means "peaceful" but also, referred to a person,
somebody with a slow pace, and/or somebody who doesn't worry too much
(like me not sure if the English word has the same connotation, but

Yes, "pacific" has exactly the same two connotations in American English
as "pacifico" has in Italian -- http://www.bartleby.com/61/37/P0003700.html
It's just a rare-ish word, being overwhelmed by [a] "peaceful" and the
geographic-connoting "Pacific" (which refers not just to the Ocean, but to
the huge variety of regions bordering said Ocean...).


Alex

 

[Michael Hudson]

My question was that, "If it is such a great language, why
it is not getting the recognition it deserves"?

Oh, I think bit by bit we're doing fine. I personally have no urge to
rush things.

Cheers,
mwh

 

[Douglas Alan]

Similar anecdotal evidence indicates the same is true for astronomers, at
least the astronomers my SO hangs out with. A friend also recently told me
of a family reunion where pretty much everyone had an Apple laptop except
him.

I work for astronomers too, and we're slowly but steadily giving up on
Solaris and moving to Linux and OS X. I also went to a conference on
"Lightweight Languages" a week ago, and 8 out of 10 notebook computers
in the audience were Apples.

|>oug

 

[Anand Pillai]

It was a generalization, not with any malice
What I meant was that many of their books were "me-toos",
especially the Early Adopter series. I did not do
a blanket bombing of their books. But I will prefer
an O'Reilly nutshell book anyday to a Wrox one.

I am not planning to convince any publishers on the
market of Python books. At least, not yet.

-Anand

 

[John Howard]

Let's say some one walked into a book store and glanced at the books
in computers. He might think, "gee look at all the books on java and
perl and hardly any books on python. What is python? And it's not even
in the languages section. No sense in looking at that!" It's a matter
of perception. "Lots of perl and java - that's the way to go." I also
agree that maybe a lot of books on python are not necessary to learn
the langage. Heck, I learned almost all I know (which I admit is not
much) from web tutorials and samples. But I do have questions I cannot
find the answers to when I browse the book store. Eg, how do I do
admin stuff with python? How do I do cgi stuff with python? There are
books on how to do these in perl. Dozens.

 

[Anand Pillai]

Exactly my feelings on this topic

-Anand

Let's say some one walked into a book store and glanced at the books
in computers. He might think, "gee look at all the books on java and
perl and hardly any books on python. What is python? And it's not even
in the languages section. No sense in looking at that!" It's a matter
of perception. "Lots of perl and java - that's the way to go." I also
agree that maybe a lot of books on python are not necessary to learn
the langage. Heck, I learned almost all I know (which I admit is not
much) from web tutorials and samples. But I do have questions I cannot
find the answers to when I browse the book store. Eg, how do I do
admin stuff with python? How do I do cgi stuff with python? There are
books on how to do these in perl. Dozens.

 

[JD]

What is the best way to debug a Python program if when i run it, it
totally hangs up the
terminal console.

Control-c or control-d has NO effect, so it's hung up somewhere, but
we have no idea
where.

We are using Python 2.3, on an openBSD box, running the python script
from the shell.

Is it possible to somehow find out what is making it hang up?

 

[Skip Montanaro]

JD> What is the best way to debug a Python program if when i run it, it
JD> totally hangs up the terminal console.

JD> Control-c or control-d has NO effect, so it's hung up somewhere, but
JD> we have no idea where.

Does it use X? If so, try setting your DISPLAY to another box. (Maybe it's
hanging the X server.) If not, try ssh'ing into the box from another
machine. Can you run top in another window while it runs? Perhaps it's
chewing up gobs of memory (got any monster range() calls in there)?

If you give us some more details about how you think it's supposed to
interact with the system we might be able to provide some more debugging
clues.

Skip

 

[Donn Cave]

Quoth JD <[email protected]>:
....
| Control-c or control-d has NO effect, so it's hung up somewhere, but
| we have no idea
| where.

One way to look at it might be ktrace. Somewhere in there, you're
stuck in one of those damned kernel mediated services like NFS, and
some ioctl to the offending device will be the last syscall reported
for the process. lsof might be good enough.

Donn Cave, (e-mail address removed)

 

[Bengt Richter]

What is the best way to debug a Python program if when i run it, it
totally hangs up the
terminal console.

Control-c or control-d has NO effect, so it's hung up somewhere, but
we have no idea
where.

We are using Python 2.3, on an openBSD box, running the python script
from the shell.

Is it possible to somehow find out what is making it hang up?

have you tried gdb python and then run with your_python_prog.py as the arg?
If gdb will catch a ctl-c then you might be able to use bt to get a backtrace?
(I don't really know much about gdb, but someone here has probably done this
and can tell you the right way to do it ;-)

Of course, there's the lions-in-africa thing of putting print statements to
narrow down successively the area where it could be hanging ;-)

Regards,
Bengt Richter

 

[Michael Hudson]

What is the best way to debug a Python program if when i run it, it
totally hangs up the
terminal console.

Control-c or control-d has NO effect, so it's hung up somewhere, but
we have no idea
where.

We are using Python 2.3, on an openBSD box, running the python
script from the shell.

"gdb -p"?

Cheers,
mwh