From: Kenny Tilton
Subject: Re: NYC LOCAL: Tuesday 11 February 2003 Lisp in New York City: Lispers gather to eat and drink
Date:
Message-ID: <3E494B58.9010308@nyc.rr.com>
Alan Krueger wrote:
> "Kenny Tilton" <·······@nyc.rr.com> wrote in message
> ·····················@nyc.rr.com...
>
>>With the kind assent of the author I am forwarding this announcement to
>>perhaps more comp NGs than he assented to. (oops) Any blame for
>>over-posting must fall on me.
>>
>>My thinking is that the nascent lispnyc.org [...]
>
>
> Perilously close to 'lipsync.org'.
I like it!
>
>
>>and associated user group
>>are casting the net wide, reaching out to functional languages of all
>>kinds,
>
>
> Which is not C++.
>
So you're not coming? :) Hey, you conveniently chopped: "...and other
folks currently using imperative languages for production work but who
might well be interested in learning of Lisp's use in production
environments." tsk, tsk.
(1) C++ has a big problem (I hear) when it comes to the time required to
rebuild large apps after certain interesting changes to the source. I
hear tales of multi-hour builds. Back to the Seventies?
With Common Lisp I can crash into a backtrace, spend a couple of hours
debugging, refactoring, recompiling and then often continue from the
backtrace by picking a call frame a couple of steps higher than the
point at which things came unglued. When I re-enter the failed branch I
dynamically pick up the new function definitions (and any class
instances automatically get updated if I changed the class definition).
(2) You say C++ is not a functional language. Interesting, they took out
recursion? Too bad, because if I had to program in C/C++ I am sure I
would carry over the functional style. That's another reason someone
doing language X should play with oddball languages Y and Z. Hell, I
remember kicking ass on a Cobol project by carrying over techniques from
Basic. Learn Lisp to write better C++. But...
(3) ...C++ has had its day. Java couldn't save it.
Python is just Lisp done badly and slowly.
You might want to get Lisp on the resume.
Peace, out.
--
kenny tilton
clinisys, inc
http://www.tilton-technology.com/
---------------------------------------------------------------
"Cells let us walk, talk, think, make love and realize
the bath water is cold." -- Lorraine Lee Cudmore
From: Alan Krueger
Subject: Re: NYC LOCAL: Tuesday 11 February 2003 Lisp in New York City: Lispers gather to eat and drink
Date:
Message-ID: <v4jc832t8oaq2e@corp.supernews.com>
"Kenny Tilton" <·······@nyc.rr.com> wrote in message
·····················@nyc.rr.com...
> Alan Krueger wrote:
> > "Kenny Tilton" <·······@nyc.rr.com> wrote in message
> > ·····················@nyc.rr.com...
[...]
> >>and associated user group
> >>are casting the net wide, reaching out to functional languages of all
> >>kinds,
> >
> > Which is not C++.
>
> So you're not coming? :) Hey, you conveniently chopped: "...and other
> folks currently using imperative languages for production work but who
> might well be interested in learning of Lisp's use in production
> environments." tsk, tsk.
This does not make it on-topic in comp.lang.c++
> (1) C++ has a big problem (I hear) when it comes to the time required to
> rebuild large apps after certain interesting changes to the source. I
> hear tales of multi-hour builds. Back to the Seventies?
Please take this to an advocacy newsgroup.
[...]
> (2) You say C++ is not a functional language. Interesting, they took out
> recursion?
An ability to write functions and recurse does not a functional language
make. While it's possible to write in a functional programming style in an
imperative/procedural language like C++, it's also possible to bail water
with a shovel.
From: Tim Bradshaw
Subject: Re: NYC LOCAL: Tuesday 11 February 2003 Lisp in New York City: Lispers gather to eat and drink
Date:
Message-ID: <ey3el6etuk5.fsf@cley.com>
* Kenny Tilton wrote:
> (1) C++ has a big problem (I hear) when it comes to the time required
> to rebuild large apps after certain interesting changes to the
> source. I hear tales of multi-hour builds. Back to the Seventies?
We tried to build a fairly well-known C++ CORBA ORB a while ago. The
machine was somewhat small (2xx MHz, 100-200MB memory), but still, a
week seemed excessive. We actually gave up before it completed (we
had the libs and IDL compiler which is what we needed), and I haven't
tried again on the larger machine we have now.
--tim
From: Ron Natalie
Subject: Re: NYC LOCAL: Tuesday 11 February 2003 Lisp in New York City: Lispers gather to eat and drink
Date:
Message-ID: <C1e2a.18$AR5.12@fe02.atl2.webusenet.com>
"Tim Bradshaw" <···@cley.com> wrote in message ····················@cley.com...
> We tried to build a fairly well-known C++ CORBA ORB a while ago. The
> machine was somewhat small (2xx MHz, 100-200MB memory), but still, a
> week seemed excessive. We actually gave up before it completed (we
> had the libs and IDL compiler which is what we needed), and I haven't
> tried again on the larger machine we have now.
CORBA's insanity isn't C++'s fault.
From: Kenny Tilton
Subject: Re: NYC LOCAL: Tuesday 11 February 2003 Lisp in New York City: Lispers gather to eat and drink
Date:
Message-ID: <3E4974AD.9040200@nyc.rr.com>
Ron Natalie wrote:
> "Tim Bradshaw" <···@cley.com> wrote in message ····················@cley.com...
>
>
>>We tried to build a fairly well-known C++ CORBA ORB a while ago. The
>>machine was somewhat small (2xx MHz, 100-200MB memory), but still, a
>>week seemed excessive. We actually gave up before it completed (we
>>had the libs and IDL compiler which is what we needed), and I haven't
>>tried again on the larger machine we have now.
>
>
> CORBA's insanity isn't C++'s fault.
Well how about a 100kloc app? Suppose I change a core header, ie, the
precompiled header has to go. How long does that take to build?
I remember when ThinkC on the Mac became Symantec C++ and incremental
linking went away. 20 sec to link on every edit-build-test iteration.
yechhh.
I think that is when I went looking for another language. Tried
SmalltalkAgents, but that never stabilized. Ended up on Mac Common Lisp,
never looked back.
--
kenny tilton
clinisys, inc
http://www.tilton-technology.com/
---------------------------------------------------------------
"Cells let us walk, talk, think, make love and realize
the bath water is cold." -- Lorraine Lee Cudmore
From: Jerry Coffin
Subject: Re: NYC LOCAL: Tuesday 11 February 2003 Lisp in New York City: Lispers gather to eat and drink
Date:
Message-ID: <MPG.18b33407b9cb2b7b98980e@news>
In article <················@nyc.rr.com>, ·······@nyc.rr.com says...
[ ... ]
> Well how about a 100kloc app? Suppose I change a core header, ie, the
> precompiled header has to go. How long does that take to build?
I just did a quick check of a total rebuild of [pause to sounds of wc
running furiously...] about 96,000 lines. It takes about two minutes
with full optimization, and less than one minute without (as you'd
normally do if you were modifying core headers).
This is with no pre-compiled headers at all, and on a machine that
hasn't been leading edge for a couple of years or so -- it wouldn't
surprise me too much if a current top-end machine could do the job twice
as fast. Of course, under normal circumstances, a total rebuild
probably doesn't happen more than once very few months or so...
--
Later,
Jerry.
The universe is a figment of its own imagination.
From: Tim Bradshaw
Subject: Re: NYC LOCAL: Tuesday 11 February 2003 Lisp in New York City: Lispers gather to eat and drink
Date:
Message-ID: <ey38ywljw9u.fsf@cley.com>
> CORBA's insanity isn't C++'s fault.
Well, that's an interesting statement. I have Lisp-based CORBA
applications, and C++ based CORBA applications, and the C++ ones take
a *lot* longer to build than the Lisp ones. I mean a *lot* longer - a
sub-300 line C++ program using CORBA takes 5x as long to compile and
link as a 20,000 line Lisp application that uses CORBA. The C++
binaries are a little smaller though. I don't have experience of
implementations of CORBA bindings for other languages, but I suspect
that they are often significantly quicker to build than the C++ ones
too.
So, well, perhaps it `isn't C++'s fault' in some sense, but it doesn't
seem to be the only sense that matters, namely that you can't actually
get systems which build in reasonable times for C++, while you can for
other languages.
--tim
From: Daniel Barlow
Subject: Re: NYC LOCAL: Tuesday 11 February 2003 Lisp in New York City: Lispers gather to eat and drink
Date:
Message-ID: <874r7albct.fsf@noetbook.telent.net>
Tim Bradshaw <···@cley.com> writes:
> * Kenny Tilton wrote:
>> (1) C++ has a big problem (I hear) when it comes to the time required
>> to rebuild large apps after certain interesting changes to the
>> source. I hear tales of multi-hour builds. Back to the Seventies?
>
> We tried to build a fairly well-known C++ CORBA ORB a while ago. The
> machine was somewhat small (2xx MHz, 100-200MB memory), but still, a
> week seemed excessive. We actually gave up before it completed (we
> had the libs and IDL compiler which is what we needed), and I haven't
> tried again on the larger machine we have now.
A week is a lot of drinking time. We could meet in London, Cambridge,
Oxford _and_ Edinburgh (and probably include the travel time, if we
avoid Virgin trains) before it got back to the shell prompt.
-dan
--
http://www.cliki.net/ - Link farm for free CL-on-Unix resources
From: Tim Bradshaw
Subject: Re: NYC LOCAL: Tuesday 11 February 2003 Lisp in New York City: Lispers gather to eat and drink
Date:
Message-ID: <ey34r79jw7q.fsf@cley.com>
* Daniel Barlow wrote:
> A week is a lot of drinking time. We could meet in London, Cambridge,
> Oxford _and_ Edinburgh (and probably include the travel time, if we
> avoid Virgin trains) before it got back to the shell prompt.
We did (well, London, Devon and Edinburgh I think).
--tim