In article <·····@castle.ed.ac.uk>
····@castle.ed.ac.uk (Jeff Dalton) writes:
> pain to use. It's a mistake to think that if Lisp merely provides a
> standard FFI then everyone will think Lisp is OK. On the contrary,
> lots of people will attack it for being too limited or too much
I think the assumption here is that the code must be loaded into
the Lisp enviroment? This is not true under Microsoft Windows,
where it's possible to dynamically link to another program module.
The only important problem left is the calling convention used,
and this has been solved by Whitewater, Microsoft, and various
others I'm sure. They do it in different ways, but the just a
matter of naming.
I assume that the Lisps being discussed here run under Unix. Can
anyone please explain why this is such a big problem with Unix?
> I sometimes think we should change the charter of Comp.lang.lisp
> to say the purpose of the newsgroup is to discuss why Lisp sucks.
It looks to me like people think some implementations "suck."
-- Cyber Surfing on CIX --
From: Jeff Dalton
Subject: Re: Are Lisp Implementations Stagnant?
Date:
Message-ID: <38750@castle.ed.ac.uk>
In article <···········@cix.compulink.co.uk> ············@cix.compulink.co.uk writes:
>In article <·····@castle.ed.ac.uk> ····@castle.ed.ac.uk (Jeff Dalton) writes:
-- It's a mistake to think that if Lisp merely provides a
-- standard FFI then everyone will think Lisp is OK. On the contrary,
-- lots of people will attack it for being too limited or too much
- I think the assumption here is that the code must be loaded into
- the Lisp enviroment? This is not true under Microsoft Windows,
- where it's possible to dynamically link to another program module.
It might be true under windows; it would depend on the Lisp
implementation. Moreover, if you want a standard it can't
just work with Windows.
- The only important problem left is the calling convention used,
- and this has been solved by Whitewater, Microsoft, and various
- others I'm sure. They do it in different ways, but the just a
- matter of naming.
What do you mean "solved"? It's easy enough to have a standard
convention used by C, Pascal, etc. For instance, there was a
de facto standard on the VAX, and such languages more or less
used it. But Lisp often didn't. So you could call from Lisp
to C in the "standard" way, provided the C implementation
used the standard, but you couldn't call the other way so
easily.
- I assume that the Lisps being discussed here run under Unix. Can
- anyone please explain why this is such a big problem with Unix?
1. A Lisp standard can't be just for Unix.
2. Lisp can't compell other languages to follow particular calling
conventions.
In article <·····@castle.ed.ac.uk> ····@castle.ed.ac.uk (Jeff Dalton) writes:
> pain to use. It's a mistake to think that if Lisp merely provides a
> standard FFI then everyone will think Lisp is OK. On the contrary,
> lots of people will attack it for being too limited or too much
Primarily, I want to see an FFI standardized because it would make my
work easier.
In addition, a significant fraction of Lisp programmers that I know
have (reluctantly) switched to other languages because of issues such
as FFI and optimization. I fear that if that trend isn't reversed
soon, there won't be enough users to fund commercial development or
encourage research grants for Lisp development. Maybe you don't care,
but I do.
Thomas.
From: Jeff Dalton
Subject: Re: Are Lisp Implementations Stagnant?
Date:
Message-ID: <38856@castle.ed.ac.uk>
In article <·················@arolla.idiap.ch> ···@idiap.ch writes:
-In article <·····@castle.ed.ac.uk> ····@castle.ed.ac.uk (Jeff Dalton) writes:
-
-- It's a mistake to think that if Lisp merely provides a
-- standard FFI then everyone will think Lisp is OK. On the contrary,
-- lots of people will attack it for being too limited or too much
-Primarily, I want to see an FFI standardized because it would make my
-work easier.
-In addition, a significant fraction of Lisp programmers that I know
-have (reluctantly) switched to other languages because of issues such
-as FFI and optimization. I fear that if that trend isn't reversed
-soon, there won't be enough users to fund commercial development or
-encourage research grants for Lisp development. Maybe you don't care,
-but I do.
I think the FFI is important, and I wanted to have a standard one
in Common Lisp. Unfortunately, there can't be one in this round
of standardization. The opportunity, if it was a real possibility
at all, has passed.
I don't think a minimal FFI would bring many people back to Lisp.
I don't think there's a trend against FFIs. What we've seen instead
is the development of better ones.
I'm entirely in favor of portable FFIs being made available on
the net.
What I object to in this discussion is the suggestion that Lisp
in effect deserves to lose because it lacks a standard FFI and
the suggestion that there's no good reason not to have a standard
FFI and so Lisp designers and implementors must be in the grip
of some losing way of thinking.
-- jd