[OT] Compilable Python-like language?

Jacek Generowicz jacek.generowicz at cern.ch
Tue Mar 23 09:31:47 EST 2004


Ed Cogburn <edcogburn at hotpop.com> writes:

> Jacek Generowicz wrote:
>  > That rather depends on what you mean by "Python-like".
> 
> 
> Oh dear...
> 
> It was not my intention to start this kind of debate,

Debate?

I think a number of us wanted to point out that your question can be
interpreted in many very different ways. We also wanted you to explain
a bit more about what _you_ meant by the words you used, in order to
enable us to give you answers which are more useful to _you_, rather
than writing pages in an attempt to answer a question which you didn't
really meant to ask.

(Or were you referring to the "debate" about whether such requests for
clarification are appropriate? I hope that it is self-evident that
they are.)

> I'm not a troll,

I don't think anyone suggested that you are. _I_ certainly did _not_
mean to imply that you are a troll.

> I actually consider GvR to be something of a minor deity.  :)

Quite an unhealthy attitude, IMHO. If more people realized that GvR is
just a guy who happens to have designed and implemented a rather nice
language, then a some discussions around here would be more
constructive that they are :-)

> I was merely curious if any of you may have known about some
> language currently in development, or some still-young and
> little-known language that was like Python and aiming for something
> like C/C++-ish performance.

Great, so you've identified performance issues as a motivation for
your question ... but we're still a bit hazy on the "like Python"
part :-)

> By compile-to-binary, I meant create a binary executable in native
> machine code that does not involve an interpreter process (source or
> byte-code), and any substantial run-time overhead.  In other words my
> concern, is as always, about speed.

Are you aware that you are slighly contradicting yourself, or at least
unnecessarily constraining your definition of "compile-to-binary"? (I
don't mean to be condescending here, in the slightest.)

My point is, if your concern really is "as always, about speed", why
should you care whether you "create a binary executable"?

For example, if the PyPy project is wildly successful, then the day
will come (I'm optimistically hoping for this to happen before the
next C++ standard comes out :-) when Python itself will compile to
native binary, run faster than C ... all that, I suspect, without
creating an "executable binary". And you couldn't ask for a more
Python-like language than Python itself ... or could you ?

> Concerning Perl/Lisp.  I'm using Python because it is *nothing* like
> those two.  :)

Funny. Here's someone who thinks that Python is rather similar to
Lisp:

In http://paulgraham.com/lispfaq1.html Paul Graham writes:

PG> I like Lisp but my company won't let me use it. What should I do?
PG> 
PG> Try to get them to let you use Python. [...]  semantically
PG> [Python] has a lot in common with Lisp, and has been getting
PG> closer to Lisp over time.

(Cue PP: "No it isn't", etc. etc.)

Are you sure that you are using Python because it is *nothing* like
Lisp?  I bet that if you removed all Python's similarities to Lisp
from the language, then you wouldn't give the language a second look.

> Lisp has good performance, and is a compilable language, I'm aware
> of that, but I just can't handle all the parentheses.  To the
> individual in the other thread who said people complaining about all
> the parentheses in Lisp are being ludicrous, well, just consider
> yourself lucky or superior that you can look at Lisp code for more
> than 30 minutes without getting a headache.  :)

The point is that, if you spend more than 30 minutes looking at Lisp
code without an attitude of "paretheses suck", then you will almost
certainly cease to have any inclination to develop a headache. If you,
subsequently, actually try to edit some Lisp code, then you will
discover that the parentheses are actually help considerably.

The problem most people have is to look at Lisp code without prejudice
for those first 30 minutes. That's the hurdle where most people fall
over.

Still, you might like to take a look at Dylan. It is, essentially, a
Lisp without parentheses.



More information about the Python-list mailing list