From: ········@gmail.com
Subject: one laptop per child
Date: 
Message-ID: <b652414e-1df8-466d-a42e-f658c75ab0ca@w73g2000hsf.googlegroups.com>
Ivan Krstik, the Director of Security Architecture for One Laptop Per
Child answers a question about why they didn't go w/ lisp or smalltalk
when making olpc xo at around 59:00 in this google tech talk

   http://video.google.com/videoplay?docid=-4285568518538296189

question person: "you mentioned that you'd like to do the sort of
things lisp or smalltalk can do... like say dynamically change
things ... so it made me wonder why you didn't choose something like
lisp or smalltalk which have a history of dynamic high performance
systems including lisp machines or smalltalk machines that are close
to the metal?"

Ivan Krstik: "the question is: if we want to be able to do more lispy
or smalltalky things why not just use lisp or smalltalk? ...and what
about lisp machines? and things like this... so if you guys have paid
attention when I said we really want to implement as much as possible
with everything in python that comes pretty close to the lisp machine
idea... now, we're not running python on the bare metal as lisp
machines did with lisp, but we're coming pretty close to it. I mean,
if you take away the kernel and look at everything else... then your
looking at python... everything is python... we are also shipping
squeak on these machines... so there is going to be a a smalltalk
implementation on these macines... but I think, for now, for any
number of reasons, such as size of the community momentum of the
community... I certainty think something like lisp has a much higher
learning curve for kids  than python does for kids.. a lot of the same
reasons apply to smalltalk... I really think the best thing we can do
is use something like python..."

Nick

ps

pls don't start another p vs l flame war w/ this post or I WILL EAT
YOUR SOUL

From: Tayssir John Gabbour
Subject: Re: one laptop per child
Date: 
Message-ID: <74ed1213-24a4-4f4d-95eb-5a76203fbfc4@w34g2000hsg.googlegroups.com>
On Nov 16, 1:23 am, ········@gmail.com wrote:
> Ivan Krstik, the Director of Security Architecture for One Laptop Per
> Child answers a question about why they didn't go w/ lisp or smalltalk
> when making olpc xo at around 59:00 in this google tech talk [...]
>
> "I certainty think something like lisp has a much higher
> learning curve for kids  than python does for kids.. a lot of the same
> reasons apply to smalltalk... I really think the best thing we can do
> is use something like python..."

Sounds sensible to me. These could be culture-specific problems
though. Personally, I (and some in my office) find Common Lisp much
clearer than other languages. Children may be free of certain damning
preconceptions which make Lisp appear difficult.

Two days ago, I met someone with serious concerns about Lisp. Did it
have iteration? Could you make GUIs? Well, I showed him loop and
dotimes on my Mac (using SBCL/Slime), then switched to my Windows VM
where I show how I made fairly complex GUIs (using Lispworks). It took
only five minutes, but it gave him a positive impression. He seemed to
accept the unconventional syntax, if it's really true that it provides
power.

The specific nature of his objections made me think that he got a
misleading impression of Lisp from somewhere. I've seen REALLY
terrible presentations of lisp, where the code was mindbogglingly
littered with SETQ, CDDADR, parens everywhere with willy-nilly
indentation, etc.


> pls don't start another p vs l flame war w/ this post or I WILL EAT
> YOUR SOUL

Considering you're sitting about three feet from me at the moment,
I'll take this seriously.


Tayssir
From: ········@gmail.com
Subject: Re: one laptop per child
Date: 
Message-ID: <0a45ea4b-b597-4a47-8010-f870d3ac21b2@c30g2000hsa.googlegroups.com>
> Sounds sensible to me. These could be culture-specific problems
> though. Personally, I (and some in my office) find Common Lisp much
> clearer than other languages. Children may be free of certain damning
> preconceptions which make Lisp appear difficult.

I'm not sure I would wish something like Common Lisp upon children...
Can you imagine some poor child from a developing nation having a cl
question and having read the hyperspec to answer? ;-) or why there is
a SET-PPRINT-DISPATCH but not a GET-PPRINT-DISPATCH or why it takes a
function that takes a stream and an object but PRINT-OBJECT takes an
object and a stream, etc...

now scheme on the other hand seems to me to fit the bill: a
pedagagically excellent language capable of doing anything. but still,
even cl-ers have an opinion that scheme is a "toy" (which is weird to
me)..

which brings me to another question: why is there no scheme to python
bytecode interpreter/compiler already? is there something about python
bytecode that makes it unsuitable for this purpose?


> Considering you're sitting about three feet from me at the moment,
> I'll take this seriously.
>

omg, I had no idea =-)

take care

Nick
From: Richard Szopa
Subject: CL and Parrot (was Re: one laptop per child)
Date: 
Message-ID: <2460cec3-a444-4640-ac76-a925926280d9@o6g2000hsd.googlegroups.com>
On Nov 16, 9:12 pm, ········@gmail.com wrote:

> which brings me to another question: why is there no scheme to python
> bytecode interpreter/compiler already? is there something about python
> bytecode that makes it unsuitable for this purpose?

I don't think compiling Scheme to Python would make much sense. Python
by now only an interpreter, which makes Python code run a lot slower
than any decent (by which I mean compiled) CL or Scheme. Of course,
most of Python's great standard library is written as C modules, which
gives them reasonable speed (some people think of Python as not much
more than a cool and easy to use glue between native code libraries).

However, there's something you may want to take a look at, nom. Parrot
[1]. Parrot is a virtual machine for dynamic languages (its name comes
from an April Fool's joke of Larry Wall and Guido van Rossum merging
Perl and Python). As I have read, there's support for some kind of
schemish language Pheme and for normal Scheme. The compiler for the
latter is written in Perl, so I have the intuition there's plenty of
room for improvement ;-)

I think it would be really cool if there were decent Common Lisp tools
for Parrot, written in Common Lisp, of course. One possible
application of a CL->Parrot compiler would be writing libraries you
could easily use with Python/Perl/PHP/Ruby/Your-favorite-mainstream-
dynamic-language.

If you didn't have any idea on how to spend the long, boring winter
evenings then hey, now you one: CL-PARROT :-).

Bests,

    -- Richard

[1] http://www.parrotcode.org/
From: ········@gmail.com
Subject: Re: CL and Parrot (was Re: one laptop per child)
Date: 
Message-ID: <2014980d-4557-439e-9f13-e3d23c3ea319@f13g2000hsa.googlegroups.com>
> I don't think compiling Scheme to Python would make much sense. Python
> by now only an interpreter, which makes Python code run a lot slower
> than any decent (by which I mean compiled) CL or Scheme. Of course,
> most of Python's great standard library is written as C modules, which
> gives them reasonable speed (some people think of Python as not much
> more than a cool and easy to use glue between native code libraries).
>

the benefit of a python-interoperable Scheme, I think, would be the
ability to involve Scheme in Python environments (such as the olpc xo)
and to have access to its standard library written in c


> for Parrot, written in Common Lisp, of course. One possible
> application of a CL->Parrot compiler would be writing libraries you
> could easily use with Python/Perl/PHP/Ruby/Your-favorite-mainstream-
> dynamic-language.

however not with their main platforms

take care

Nick
From: Rainer Joswig
Subject: Re: CL and Parrot (was Re: one laptop per child)
Date: 
Message-ID: <joswig-8DC021.19121217112007@news-europe.giganews.com>
In article 
<····································@o6g2000hsd.googlegroups.com>,
 Richard Szopa <·············@gmail.com> wrote:

> On Nov 16, 9:12 pm, ········@gmail.com wrote:
> 
> > which brings me to another question: why is there no scheme to python
> > bytecode interpreter/compiler already? is there something about python
> > bytecode that makes it unsuitable for this purpose?
> 
> I don't think compiling Scheme to Python would make much sense. Python
> by now only an interpreter, which makes Python code run a lot slower
> than any decent (by which I mean compiled) CL or Scheme. Of course,
> most of Python's great standard library is written as C modules, which
> gives them reasonable speed (some people think of Python as not much
> more than a cool and easy to use glue between native code libraries).
> 
> However, there's something you may want to take a look at, nom. Parrot
> [1]. Parrot is a virtual machine for dynamic languages (its name comes
> from an April Fool's joke of Larry Wall and Guido van Rossum merging
> Perl and Python). As I have read, there's support for some kind of
> schemish language Pheme and for normal Scheme. The compiler for the
> latter is written in Perl, so I have the intuition there's plenty of
> room for improvement ;-)
> 
> I think it would be really cool if there were decent Common Lisp tools
> for Parrot, written in Common Lisp, of course. One possible
> application of a CL->Parrot compiler would be writing libraries you
> could easily use with Python/Perl/PHP/Ruby/Your-favorite-mainstream-
> dynamic-language.
> 
> If you didn't have any idea on how to spend the long, boring winter
> evenings then hey, now you one: CL-PARROT :-).

There is some work on a Common Lisp for Parrot: Kea Common Lisp.

  http://rgrjr.dyndns.org/lisp/kea-cl/

Btw. Keas are extremely intelligent parrots.
They are the mechanics among the birds. 


Also worth to mention that there is a Perl 6 compiler
written in a subset of Perl 6. The compiler can
generate Lisp code(!).

http://www.pugscode.org/kp6.html


> 
> Bests,
> 
>     -- Richard
> 
> [1] http://www.parrotcode.org/

-- 
http://lispm.dyndns.org/
From: Richard Szopa
Subject: Re: CL and Parrot (was Re: one laptop per child)
Date: 
Message-ID: <f2c410b4-6a18-4105-9ea8-7dea8841f88b@d50g2000hsf.googlegroups.com>
On Nov 17, 7:12 pm, Rainer Joswig <······@lisp.de> wrote:

> There is some work on a Common Lisp for Parrot: Kea Common Lisp.
>
>  http://rgrjr.dyndns.org/lisp/kea-cl/
>
> Btw. Keas are extremely intelligent parrots.
> They are the mechanics among the birds.
>
> Also worth to mention that there is a Perl 6 compiler
> written in a subset of Perl 6. The compiler can
> generate Lisp code(!).
>
> http://www.pugscode.org/kp6.html

Wow, this looks interesting. I'll definitely take a look at Kea.

Thanks!

Bests,

    -- Richard
From: Ken Tilton
Subject: Re: one laptop per child
Date: 
Message-ID: <vLn%i.444$Tl3.193@newsfe09.lga>
········@gmail.com wrote:
>>Sounds sensible to me. These could be culture-specific problems
>>though. Personally, I (and some in my office) find Common Lisp much
>>clearer than other languages. Children may be free of certain damning
>>preconceptions which make Lisp appear difficult.
> 
> 
> I'm not sure I would wish something like Common Lisp upon children...
> Can you imagine some poor child from a developing nation having a cl
> question and having read the hyperspec to answer? ;-)

And wait till I get my hands on the little devils.

> or why there is
> a SET-PPRINT-DISPATCH but not a GET-PPRINT-DISPATCH or why it takes a
> function that takes a stream and an object but PRINT-OBJECT takes an
> object and a stream, etc...

Actually kids do not have the problem of pre-formed obsessions with 
consistency that would have them arguing with CL, they would just use 
it. Grown-ups are the ones with rules.


kenny


-- 
http://www.theoryyalgebra.com/

"In the morning, hear the Way;
  in the evening, die content!"
                     -- Confucius
From: Tayssir John Gabbour
Subject: Re: one laptop per child
Date: 
Message-ID: <5691eff3-b36e-4c43-82d5-f4101e933a4a@d50g2000hsf.googlegroups.com>
On Nov 16, 9:12 pm, ········@gmail.com wrote:
 > I'm not sure I would wish something like Common Lisp upon
children...
> Can you imagine some poor child from a developing nation having a cl
> question and having read the hyperspec to answer? ;-) or why there is
> a SET-PPRINT-DISPATCH but not a GET-PPRINT-DISPATCH or why it takes a
> function that takes a stream and an object but PRINT-OBJECT takes an
> object and a stream, etc...

I agree that Lisp has a lot of gotchas, and I even collected a page of
them:
http://wiki.alu.org/Lisp_Gotchas

OTOH, Python was precisely the language I was dissatisfied with when I
started looking for other languages and found CL. ;) Python brings its
own share of pain.

A politer pain than CL's, but still problematic and perhaps less
fixable...


All best,
#\Tayssir
From: Brian Adkins
Subject: Re: one laptop per child
Date: 
Message-ID: <6d52d7a6-73c7-4661-90cf-3d306130ed62@l1g2000hsa.googlegroups.com>
On Nov 17, 4:43 pm, Tayssir John Gabbour <············@googlemail.com>
wrote:
> I agree that Lisp has a lot of gotchas, and I even collected a page of
> them:http://wiki.alu.org/Lisp_Gotchas

FYI - the wiki seems be having problems. I only see a blank page with
"Lisp Gotchas" and search box with associated buttons. I'm using
Firefox 2.0.0.6 on Ubuntu Linux. Similar result on the main wiki page
also.
From: Tayssir John Gabbour
Subject: Re: one laptop per child
Date: 
Message-ID: <69cfd4e2-f141-4db6-8ad3-832ef645e14c@i37g2000hsd.googlegroups.com>
On Nov 18, 12:41 am, Brian Adkins <···········@gmail.com> wrote:
> On Nov 17, 4:43 pm, Tayssir John Gabbour wrote:
> > I agree that Lisp has a lot of gotchas, and I even collected a page of
> > them:http://wiki.alu.org/Lisp_Gotchas
>
> FYI - the wiki seems be having problems. I only see a blank page with
> "Lisp Gotchas" and search box with associated buttons. I'm using
> Firefox 2.0.0.6 on Ubuntu Linux. Similar result on the main wiki page
> also.

Seems to be an intermittant problem. Tested fine until this sec; now
it manifests your problem across all my platforms. The ALU wiki has
always been flaky.

(BTW, I should make my position clear: the decision to go with python
was probably the wisest, just I would've thought a few more
milliseconds about it. ;)


#\Tayssir
From: Rob Warnock
Subject: Re: one laptop per child
Date: 
Message-ID: <la2dncKlQfziLKLanZ2dnUVZ_oCvnZ2d@speakeasy.net>
Tayssir John Gabbour  <············@googlemail.com> wrote:
+---------------
| Brian Adkins <···········@gmail.com> wrote:
| > FYI - the wiki seems be having problems. I only see a blank page with
| > "Lisp Gotchas" and search box with associated buttons.
+---------------

Yup. It outputs everything it should, up through the page title,
then abruptly closes the HTTP connection. [But the HTTP response
header "Content-Length:" *does* match the length sent.]

+---------------
| Seems to be an intermittant problem. Tested fine until this sec; now
| it manifests your problem across all my platforms. The ALU wiki has
| always been flaky.
+---------------

It's probably not the Lisp server or the Wiki code per se, but
rather its connection with the database behind it. [At least,
that's what it was the last time this exact symptom manifested
itself.] I have sent a message to one of the maintainers asking
for this to be checked...


-Rob

-----
Rob Warnock			<····@rpw3.org>
627 26th Avenue			<URL:http://rpw3.org/>
San Mateo, CA 94403		(650)572-2607
From: Robert Maas, see http://tinyurl.com/uh3t
Subject: Re: one laptop per child
Date: 
Message-ID: <rem-2007nov19-012@yahoo.com>
> From: Tayssir John Gabbour <············@googlemail.com>
> I agree that Lisp has a lot of gotchas, and I even collected a page of
> them:
> http://wiki.alu.org/Lisp_Gotchas

Um, that's only 4 lines (plus the title), with no actual content:

#                                                        ALU Wiki: Lisp Gotchas
   #ALU Wiki : Recent Changes
   ____________________
   Search Go Random
   Lisp Gotchas

So is that the joke, empty set of gotchas??
Or is it like an adventure (colossal cave) game except with not a
clue what is possible to explore such as "you are standing in front
of a REPL which has been split into three parts"?

I tried clicking on "Random", but it just gave me:
   You are being redirected to a random page. If it doesn't work, click
   here.
and when I clicked on "here" it just gave me that back again.
Why can't it just say "you can't scare the bird into the cage by
shaking the stick, because you aren't currently holding the stick"??
From: Rob Warnock
Subject: Re: one laptop per child
Date: 
Message-ID: <nMOdnRkD_cerRd_anZ2dnUVZ_qKgnZ2d@speakeasy.net>
Robert Maas, see http://tinyurl.com/uh3t <·······@yahoo.com> wrote:
+---------------
| > http://wiki.alu.org/Lisp_Gotchas
| 
| Um, that's only 4 lines (plus the title), with no actual content:
...
| I tried clicking on "Random", but it just gave me:
|    You are being redirected to a random page. If it doesn't work, click
|    here.
| and when I clicked on "here" it just gave me that back again.
+---------------

According to one of the ALU Wiki maintainers, there was a problem with
the interface from the Lisp app server to the SQL database behind it.
He thinks the problem has been resolved. I just tested it, and both
<http://wiki.alu.org/Lisp_Gotchas> and the "Random" button seem to
be working fine now. Please try it again and let us know if it still
doesn't work for you.


-Rob

-----
Rob Warnock			<····@rpw3.org>
627 26th Avenue			<URL:http://rpw3.org/>
San Mateo, CA 94403		(650)572-2607
From: JohnW
Subject: Re: one laptop per child
Date: 
Message-ID: <75868ef8-28c8-4850-8ee4-338e5dae66cc@w73g2000hsf.googlegroups.com>
On Nov 15, 8:23 pm, ········@gmail.com wrote:
> so if you guys have paid
> attention when I said we really want to implement as much as possible
> with everything in python that comes pretty close to the lisp machine
> idea... now, we're not running python on the bare metal as lisp
> machines did with lisp, but we're coming pretty close to it. I mean,
> if you take away the kernel and look at everything else...

When people say things like this, I wonder if they really understand
what Lisp and Smalltalk are, or what the Lisp Machines offered in
terms of an interactive, dynamic environment.  I've been using Python
for years, but I still find it's capabilities at the REPL quite basic,
and there is no way to dynamically update a remotely running
installation unless I've specifically created functionality to do so.

His statement sounded a bit like someone who wasn't entirely sure of
the right answer, but since he had been "sold" on Python, was ready to
believe on faith that it's equivalent to whatever came before.

John
From: Ken Tilton
Subject: Re: one laptop per child
Date: 
Message-ID: <aKv%i.1584$Tl3.581@newsfe09.lga>
JohnW wrote:
> On Nov 15, 8:23 pm, ········@gmail.com wrote:
> 
>>so if you guys have paid
>>attention when I said we really want to implement as much as possible
>>with everything in python that comes pretty close to the lisp machine
>>idea... now, we're not running python on the bare metal as lisp
>>machines did with lisp, but we're coming pretty close to it. I mean,
>>if you take away the kernel and look at everything else...
> 
> 
> When people say things like this, I wonder if they really understand
> what Lisp and Smalltalk are, or what the Lisp Machines offered in
> terms of an interactive, dynamic environment.  I've been using Python
> for years, but I still find it's capabilities at the REPL quite basic,
> and there is no way to dynamically update a remotely running
> installation unless I've specifically created functionality to do so.
> 
> His statement sounded a bit like someone who wasn't entirely sure of
> the right answer, but since he had been "sold" on Python, was ready to
> believe on faith that it's equivalent to whatever came before.

But you in truth find it hard to imagine that sequence. Much easier is 
tossing aside anything anyone ever says in favor of any position they 
might hold on the basis that it is no more than ad hoc inefficient 
incomplete justification of a position they hold on no stronger basis 
than did our beloved ancients who understood /some/ agent must be 
reliably dragging that bright yellow disk from here to there each day 
because the only time they had observed such a thing happening was when 
some agent they could observe was doing so, loop connected.

Bringing us back to Tilton's Law: Never ask or listen to the answer to 
the question Why?

kenny
From: ········@gmail.com
Subject: Re: one laptop per child
Date: 
Message-ID: <e2bb7dab-8ce4-4349-9280-40d944d8d22d@d50g2000hsf.googlegroups.com>
> His statement sounded a bit like someone who wasn't entirely sure of
> the right answer, but since he had been "sold" on Python, was ready to
> believe on faith that it's equivalent to whatever came before.

but isn't prior experience  a perfectly valid reason for using it in a
project?

"right tool for the right job" is  ideal but "right tool for the right
job, keeping in mind the current resources and experience of the
development team" is also applicable

Nick
From: Mark Tarver
Subject: Re: one laptop per child
Date: 
Message-ID: <03c8dd3f-f6f0-480b-afeb-f09060716883@w73g2000hsf.googlegroups.com>
On 16 Nov, 00:23, ········@gmail.com wrote:
> Ivan Krstik, the Director of Security Architecture for One Laptop Per
> Child answers a question about why they didn't go w/ lisp or smalltalk
> when making olpc xo at around 59:00 in this google tech talk
>
>    http://video.google.com/videoplay?docid=-4285568518538296189
>
> question person: "you mentioned that you'd like to do the sort of
> things lisp or smalltalk can do... like say dynamically change
> things ... so it made me wonder why you didn't choose something like
> lisp or smalltalk which have a history of dynamic high performance
> systems including lisp machines or smalltalk machines that are close
> to the metal?"
>
> Ivan Krstik: "the question is: if we want to be able to do more lispy
> or smalltalky things why not just use lisp or smalltalk? ...and what
> about lisp machines? and things like this... so if you guys have paid
> attention when I said we really want to implement as much as possible
> with everything in python that comes pretty close to the lisp machine
> idea... now, we're not running python on the bare metal as lisp
> machines did with lisp, but we're coming pretty close to it. I mean,
> if you take away the kernel and look at everything else... then your
> looking at python... everything is python... we are also shipping
> squeak on these machines... so there is going to be a a smalltalk
> implementation on these macines... but I think, for now, for any
> number of reasons, such as size of the community momentum of the
> community... I certainty think something like lisp has a much higher
> learning curve for kids  than python does for kids.. a lot of the same
> reasons apply to smalltalk... I really think the best thing we can do
> is use something like python..."
>
> Nick
>
> ps
>
> pls don't start another p vs l flame war w/ this post or I WILL EAT
> YOUR SOUL

Looking at my mbox; up until April 2007, I had a series of dialogues
with SJ Klein including a transatlantic telephone call on the subject
of Lisp for the OLPC.  The plan was to port CLisp to the machine as a
prep to porting Qi - SJK wrote

QUOTE
Hello, Mark -- I'd also be glad to hear more about those ports.
Python
has many advantages in our eyes, in terms of its toolchains and
environments for viewing what is going on around one in the system
without breaking out of the program... but of course it would be
interesting and useful to have a lisp environment available as well.
UNQUOTE

Unfortunately I couldn't get the interest on the CLisp side and this
opportunity was lost.  I think that SJK did not realise that CLisp was
small and had the mindset that CL was large and was surprised to learn
that it was not.  IMO, Lisp with its faster performance would have
been better than Python for this slow machine.  Qi would have provided
the facilities of a modern FP and is easier to learn than Lisp without
sacrificing any of the power of Lisp.

OLPC are tho', in some trouble right now, as their marketing policy
has dropped round their ankles.  Trying to hoist it up and run is
proving difficult.   A great effort has been spent on hardware design,
but distribution, sales, software and everything else have not been
planned well.   The G1G1 policy is probably a mistake.

Mark
From: Slobodan Blazeski
Subject: Re: one laptop per child
Date: 
Message-ID: <21f61786-7518-476b-ba5a-3e45da6c324a@l22g2000hsc.googlegroups.com>
Come on people, cut the crap everybody knows how outdated is lisp. Of
course they should choose Python. No Ruby is probably a better choice.
Or maybe Lua. Or F#. Or...  what's the hottest fad today?

Slobodan
From: ··············@gmail.com
Subject: Re: one laptop per child
Date: 
Message-ID: <78cf0eb5-05ab-4645-b355-abfa5a9d0e4a@i12g2000prf.googlegroups.com>
On Nov 15, 5:23 pm, ········@gmail.com wrote:
> Ivan Krstik, the Director of Security Architecture for One Laptop Per
> Child answers a question about why they didn't go w/ lisp or smalltalk
> when making olpc

Directly using Lisp, Scheme or Smalltalk, or python for that matter,
on the OLPC as programming environment for kids does not make much
sense. Precisely what would they do with that? Kids are very
interested in building simulations and games but one needs to scaffold
that with high level programming languages and integrated environments
including graphics etc. Of course these environments could be and
perhaps should be implemented in Lisp. It would be great to have CL
available on the OLPC allowing developers to make some game design
tools for kids. These tools could be open in the sense that if the
kids exhausted the high level interfaces they could drop into Lisp to
gain access to more functionality. That would be cool and useful.

Is there any Lisp out there that is close to be ported or at least
portable to the OLPC?

Alex
From: Mark Tarver
Subject: Re: one laptop per child
Date: 
Message-ID: <ae7f748c-e653-4c25-bc9b-c0a54dd51a18@b32g2000hsa.googlegroups.com>
On 19 Nov, 17:16, ··············@gmail.com wrote:
> On Nov 15, 5:23 pm, ········@gmail.com wrote:
>
> > Ivan Krstik, the Director of Security Architecture for One Laptop Per
> > Child answers a question about why they didn't go w/ lisp or smalltalk
> > when making olpc
>
> Directly using Lisp, Scheme or Smalltalk, or python for that matter,
> on the OLPC as programming environment for kids does not make much
> sense. Precisely what would they do with that? Kids are very
> interested in building simulations and games but one needs to scaffold
> that with high level programming languages and integrated environments
> including graphics etc. Of course these environments could be and
> perhaps should be implemented in Lisp. It would be great to have CL
> available on the OLPC allowing developers to make some game design
> tools for kids. These tools could be open in the sense that if the
> kids exhausted the high level interfaces they could drop into Lisp to
> gain access to more functionality. That would be cool and useful.
>
> Is there any Lisp out there that is close to be ported or at least
> portable to the OLPC?
>
> Alex

CLisp might be the best bet for a small sized CL if space is crucial.
It also runs under RedHat.  But you have to convince somebody to do
the port.

Mark
From: gavino
Subject: Re: one laptop per child
Date: 
Message-ID: <745d9ede-ff04-4288-8f28-a9c1981801f5@41g2000hsh.googlegroups.com>
On Nov 15, 4:23 pm, ········@gmail.com wrote:
> Ivan Krstik, the Director of Security Architecture for One Laptop Per
> Child answers a question about why they didn't go w/ lisp or smalltalk
> when making olpc xo at around 59:00 in this google tech talk
>
>    http://video.google.com/videoplay?docid=-4285568518538296189
>
> question person: "you mentioned that you'd like to do the sort of
> things lisp or smalltalk can do... like say dynamically change
> things ... so it made me wonder why you didn't choose something like
> lisp or smalltalk which have a history of dynamic high performance
> systems including lisp machines or smalltalk machines that are close
> to the metal?"
>
> Ivan Krstik: "the question is: if we want to be able to do more lispy
> or smalltalky things why not just use lisp or smalltalk? ...and what
> about lisp machines? and things like this... so if you guys have paid
> attention when I said we really want to implement as much as possible
> with everything in python that comes pretty close to the lisp machine
> idea... now, we're not running python on the bare metal as lisp
> machines did with lisp, but we're coming pretty close to it. I mean,
> if you take away the kernel and look at everything else... then your
> looking at python... everything is python... we are also shipping
> squeak on these machines... so there is going to be a a smalltalk
> implementation on these macines... but I think, for now, for any
> number of reasons, such as size of the community momentum of the
> community... I certainty think something like lisp has a much higher
> learning curve for kids  than python does for kids.. a lot of the same
> reasons apply to smalltalk... I really think the best thing we can do
> is use something like python..."
>
> Nick
>
> ps
>
> pls don't start another p vs l flame war w/ this post or I WILL EAT
> YOUR SOUL

lisp has no apps to back it up as cmpared to python?