M-expressions

Ulrich Hobelmann wrote:
>
> I think a language needs to be *really* crappy, like Java, in order to
> succeed (you know, worse is better!).

I don't intend to repeat all the stuff here that I just went through in
comp.lang.lisp. But I have a theory that languages aren't what's driving
people's language choices. Rather, it's the internet. All the web stuff.
When I look at all these "dumb" books on the University Bookstore shelf, the
common thread of obscure things like "Squid" and "Ruby" is they're about the
web. Anything, everything, that could conceivably make a profit on the web.
So, if you can package up a competitive advantage for web development, you
might gain language adherants.

"Games" have also become a respectable category of books. 3 shelves full at
B&N.

--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA

When no one else sells courage, supply and demand take hold.

Ray Dillinger wrote:
>
> I guess the question is whether you want the "best" paradigm for
> an individual program or programmer, or the "best" paradigm for a
> widespread community of programmers.

I want the best paradigm for an individual programmer. I think our society
is based on incredible wasteages of work, because team development is always
assumed. We can have some super-bloated process for developing software
because we've got a lot of bodies and a lot of money to throw at
programmers. If 1 programmer *has* to get the important things done, then
the important things are going to become more sharply defined. I hope the
tools will arise that allow the individual to get legions of work done.

--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA

When no one else sells courage, supply and demand take hold.

Brian Harvey wrote:
> But you have to imagine that we're back in the days when the only
> real competition was Fortran. Maybe then there was an opening for
> a high level language -- although maybe then the world wasn't ready
> for one. (Yeah, I know, maybe today the world still isn't ready.)

Remember, we're talking about:
a) fifties hardware, and
b) fifties lisp compiler tech.

Not a language that would fit all tasks, at the time.

Ray Dillinger wrote:
> Or maybe they're not really harder to climb, and just _seem_
> inaccessible because there's no city with all the amenities on
> top right now.

Or, almost no signs pointing to those hilltops. (Only SICP, _On Lisp_,
and few more.)

I'm very happy that Java wasn't my first language; I'm not very fond of
that way of organizing programs - as a bunch of "structs with variables
and functions glued to them", and that every part of the program has to
be expressed that way (and, in Java's case, kinda verbose syntax).

There are so many other ways of expressing programs - layers upon layers
is one that often works for me.

I started programming using a hideously imperative subset of basic. I
used plenty of goto and variables. While familiarity with an imperative
language helped me out when I moved to assembly, it's not a very
versatile way of programming.

Lisp-style languages hit the sweetspot for me in many ways (and I've
always hated infix, so I adored the syntax when I first saw it).

I don't get the idea that, say, Scheme, should be more difficult than
Java or C++, or more arcane/esoteric than Ruby.

(I've Ruby folks rant that Scheme is so "academic"/"inaccessible", then
in the next breath start to talk about Ruby's MOP, and/or invent weird
data languages like YAML instead of using s-exps. (And I do think Ruby
has many nice qualities - it's among my favourite languages - I just
don't think it competes with Scheme variants.))

> I want the best paradigm for an individual programmer.

I want the best paradigm for an individual _program_. Lisp
languages, to me, seem the most generally useful languages
because every paradigm finds fairly convenient expression.

Well, that's the dream anyway.

-Jordan Henderson
jordan.henderson@gmail.com

Sunnan <sunnan@handgranat.org> schrieb:
> (I've Ruby folks rant that Scheme is so "academic"/"inaccessible", then
> in the next breath start to talk about Ruby's MOP, and/or invent weird
> data languages like YAML instead of using s-exps. (And I do think Ruby
> has many nice qualities - it's among my favourite languages - I just
> don't think it competes with Scheme variants.))

Practically every time I see the word "Ruby" online these days, it's
because somebody is raving about Ruby on Rails. Here are some potential
translations for anybody wishing to start the equivalent scheme project:

Scheme Full-Sail
Scheme to Scale
Scheme on the Scene (Like a SEXP Machine)
Scheme that Screams
Scudding Scheme
Scheme on Skis
Spiked Scheme

I was very disappointed that I couldn't capture the dual "guide/support
structure" and "something that makes you go fast" meanings of "rails".

Ray Dillinger wrote:
> It's a funny thing. I'm starting to think of OO as a "hill" in the
> fitness space of programming paradigms. Not the highest hill, but
> one with a nice easy slope to climb that lets a lot of people get
> fairly high up with a reasonably small amount of effort.
>
> And of course, these days there's a city built on the summit, which
> means there are roads along the way and vendors to help you along
> with whatever you need to make the journey up the hill, and bankers
> willing to finance you when you buy a house there and so on.
>
> There are higher peaks, more expressive paradigms, things that allow
> you to get higher and see more and do more -- but they're relatively
> hard to climb, and may not be accessible enough to a large population
> of programmers to build a city on top of. Instead you get scattered
> cottages, and maybe a small village here and there.

I don't think Lisp is hard to climb (or ML, for that matter). I
would describe OO as something like this: if you want to build a
standard house, you get a credit from your bank. Lisp and other
"exotic" languages are a small town somewhere out, where the
houses are round, and so nobody wants to live there, because
they're all too conventional (i.e. bound by the opinions of other
people in the programming community). I would like to plug Ayn
Rand's "The Fountainhead" here (beautiful book), the situation
reminds me a little of that.

> Or maybe they're not really harder to climb, and just _seem_
> inaccessible because there's no city with all the amenities on
> top right now.

Ah, ok. Let's say, things are different in another culture, and
some people are just too prejudiced against (or grossed out by)
other cultures (might happen to me too, if a culture is TOO
different).

> OO demonstrably is not the most expressive paradigm in computing.
> But at the moment, it looks as if it may be the most expressive
> paradigm that can be easily mastered by a very widespread community,
> and that, whatever else is going on, does have real value.

No. It's the only thing that's really ever explained to people.
How respectable university people can teach Java of their own free
will (or are they forced to?), is beyond my comprehension. I
tutored some students one semester in beginning Java, and the
language is the WORST EVER language -- in terms of didactics --
for beginning students. May ML and Scheme (even Haskell) spread
further in academics.

Brandon J. Van Every wrote:
> Ulrich Hobelmann wrote:
>
>>I think a language needs to be *really* crappy, like Java, in order to
>>succeed (you know, worse is better!).
>
>
> I don't intend to repeat all the stuff here that I just went through in
> comp.lang.lisp. But I have a theory that languages aren't what's driving
> people's language choices. Rather, it's the internet. All the web stuff.
> When I look at all these "dumb" books on the University Bookstore shelf, the
> common thread of obscure things like "Squid" and "Ruby" is they're about the
> web. Anything, everything, that could conceivably make a profit on the web.
> So, if you can package up a competitive advantage for web development, you
> might gain language adherants.

Of course the web is important. But there's dynamic web servers
for most languages, I believe. Lisp people just don't talk too
much about them.

You might say, that people don't look for a technical quality, but
they listen *to the people around them*, which is just another
form of (self-inflicted) slavery. The internet breaks this
locality rule a bit, as we're all on usenet now, but in general
our culture is still in the middle ages.

(Lispen to your heart)

Adrian Kubala wrote:
>
> I was very disappointed that I couldn't capture the dual
> "guide/support structure" and "something that makes you go fast"
> meanings of "rails".

Actually, for me personally, the phrase "X on Rails" does not imply that X
goes fast. In fact I really have no idea what it means at all. Railroad
metaphors just aren't in the brains of most Americans. We drive cars.
Well, lately I take the bus as my car was totalled, but please no bus
metaphors!

In the negative associations dept. that's so crazy it just might work, how
about

PyramidScheme


--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA

20% of the world is real.
80% is gobbledygook we make up inside our own heads.

On Fri, 18 Mar 2005 20:43:02 -0800, "Brandon J. Van Every"
<try_vanevery_at_mycompanyname@yahoo.com> wrote:

>Ray Dillinger wrote:
>>
>> I guess the question is whether you want the "best" paradigm for
>> an individual program or programmer, or the "best" paradigm for a
>> widespread community of programmers.
>
>I want the best paradigm for an individual programmer. I think our society
>is based on incredible wasteages of work, because team development is always
>assumed. We can have some super-bloated process for developing software
>because we've got a lot of bodies and a lot of money to throw at
>programmers. If 1 programmer *has* to get the important things done, then
>the important things are going to become more sharply defined. I hope the
>tools will arise that allow the individual to get legions of work done.

Team development *must* be assumed in general. It's true that some
virtuoso programmers exist (if only in their own minds) but the fact
is - if you consider as participants any library developers - nearly
all applications are produced by team. Outside the small system
embedded world, it's almost unheard of to find an application written
completely, top to bottom, by a single person. Even in the embedded
world, such efforts are becoming increasingly rare.

You have posted numerous times expressing frustration about lacking
FFI support in various high level languages. What did you need that
for again? I remember now ... integration with somebody else's code.

George
--
for email reply remove "/" from address

George Neuner wrote:
>
> Team development *must* be assumed in general.

Today. In our current era of technological history.

> You have posted numerous times expressing frustration about lacking
> FFI support in various high level languages. What did you need that
> for again? I remember now ... integration with somebody else's code.

2 reasons for FFI's:
1) interface to C++ 3D engines that other people have written
2) interface to my own low level C and ASM code


--
Cheers, www.indiegamedesign.com
Brandon Van Every Seattle, WA

When no one else sells courage, supply and demand take hold.

George Neuner wrote:
> You have posted numerous times expressing frustration about lacking
> FFI support in various high level languages. What did you need that
> for again? I remember now ... integration with somebody else's code.

Integration with other people's code, as well as free software-style
distributed developement, are different things from "a group of people
sitting in an office, coding together".

It's become difficult to be a lone programmer (albeit one working with
libraries and/or other people's code), even if some new tools (like
darcs) help.

For example, I wouldn't undertake doing a large-scale Java application
on my own, since I'd be doing a job that was meant to be done by a group
of people. I'd probably have to bother with ant, classes, UML, some
weird introspective IDE, a complex unit testing framework, strange
database libraries, and all the other boring boilerplate stuff that Java
is know for. Which would take a lot more time than, as well as being a
lot more work than, undertaking a large-scale Scheme application which I
could code using the "layers upon layers" design style that lisp encourages.

That's why I think that much of current software development is geared
towards a "wasteful" style (and I think that's what the original poster
meant, as well).

And yes, I do encourage "a group of people sitting in an office, coding
together", as well as pair programming and similar practices - it's just
that the tools and developement processes that are made for that kind of
environment are commonly pretty heavy. It could be done in a simpler way
(lisp, smalltalk), even for a group.

Sunnan schreef:
> Which would take a lot more time than, as well as being a
> lot more work than, undertaking a large-scale Scheme application which I
> could code using the "layers upon layers" design style that lisp
> encourages.

Is there some paper about 'layers upon layers'?

Adrian Kubala wrote:
> Practically every time I see the word "Ruby" online these days, it's
> because somebody is raving about Ruby on Rails. Here are some potential
> translations for anybody wishing to start the equivalent scheme project:

In many ways it'd be better to have a name that isn't reminiscent of
Ruby on Rails at all.

Hans Oesterholt-Dijkema wrote:
> Is there some paper about 'layers upon layers'?

I mean the design style that the SICP lectures encourage, as well as
Graham's _On Lisp_.