Please join us for our next meeting on Tuesday, August 14th from 7:00 to 9:00 at Trinity Lutheran Church.
Perry Metzger Presents: Otter, a new dialect of Lisp.
Otter is strongly influenced by Scheme and Python, with some bits of inspiration from Common Lisp, Perl and other languages thrown in, as well as some strange ideas Perry came up up with on his own. The results may be pleasing to some and are guaranteed to be repulsive to others. If you're tired of waiting for Paul Graham to release Arc, you may find Otter interesting. If you don't care if Paul Graham ever releases Arc, you may also find Otter interesting. In the unlikely event that all goes according to plan, a prototype Otter interpreter.
Perry Metzger is a long time ne'er-do-well and agent provocateur. He subsists largely on cruciferous vegetables, and rarely turns down a chance to nap. Besides designing Otter, he is currently learning how to distinguish different organic solvents by scent alone. He lives in New York with a purple haired woman and many elderly computers whom he cares for in their golden years.
From N,R,Q,W (8th Street NYU Stop) and the 4,5 (Astor Street Stop): Walk East 4 blocks on St. Marks, cross Tompkins Square Park.
From F&V (2nd Ave Stop): Walk E one or two blocks, turn north for 8 short blocks
From L (1st Ave Stop): Walk E one block, turn sounth for 5 short blocks
The M9 bus line drops you off at the doorstep and the M15 is near get off on St. Marks & 1st)
To get there by car, take the FDR (East River Drive) to Houston then go NW till you're at 9th & B. Week-night parking isn't bad at all, but if you're paranoid about your Caddy or in a hurry, there is a parking garage on 9th between 1st and 3rd Ave.
Discussion subject changed to "Arc about to be released!!!!!! [was Re: NYC LOCAL: Tuesday 14 August 2007 Lisp NYC: Perry Metzger on Otter, the new salty slick whiskered snow-bellying programming system}" by Ken Tilton
Subject: Arc about to be released!!!!!! [was Re: NYC LOCAL: Tuesday 14 August 2007 Lisp NYC: Perry Metzger on Otter, the new salty slick whiskered snow-bellying programming system}
Otter is advertised by its author to be something one waiting on Arc would like. The author is a hard-core schemer, and Arc's Prime Directive is to be minimalist. Like Scheme.
> Please join us for our next meeting on Tuesday, August 14th from 7:00 > to 9:00 at Trinity Lutheran Church.
> Perry Metzger Presents: Otter, a new dialect of Lisp.
> Otter is strongly influenced by Scheme and Python, with some > bits of inspiration from Common Lisp, Perl and other > languages thrown in, as well as some strange ideas Perry > came up up with on his own. The results may be pleasing to > some and are guaranteed to be repulsive to others. If you're > tired of waiting for Paul Graham to release Arc, you may > find Otter interesting. If you don't care if Paul Graham > ever releases Arc, you may also find Otter interesting. In > the unlikely event that all goes according to plan, a > prototype Otter interpreter.
> Perry Metzger is a long time ne'er-do-well and agent provocateur. He > subsists largely on cruciferous vegetables, and rarely turns down a > chance to nap. Besides designing Otter, he is currently learning how to > distinguish different organic solvents by scent alone. He lives in New > York with a purple haired woman and many elderly computers whom he cares > for in their golden years.
> From N,R,Q,W (8th Street NYU Stop) and the 4,5 (Astor Street Stop): > Walk East 4 blocks on St. Marks, cross Tompkins Square Park.
> From F&V (2nd Ave Stop): > Walk E one or two blocks, turn north for 8 short blocks
> From L (1st Ave Stop): > Walk E one block, turn sounth for 5 short blocks
> The M9 bus line drops you off at the doorstep and the M15 is near get > off on St. Marks & 1st)
> To get there by car, take the FDR (East River Drive) to Houston then > go NW till you're at 9th & B. Week-night parking isn't bad at all, > but if you're paranoid about your Caddy or in a hurry, there is a > parking garage on 9th between 1st and 3rd Ave.
Discussion subject changed to "NYC LOCAL: Tuesday 14 August 2007 Lisp NYC: Perry Metzger on Otter, the new salty slick whiskered snow-bellying programming system" by Stefan Scholl
Well, it will get better. Mailing list was shown on a slide, but I thought the above link would be enough to remember. You might pester Perry on the LispNYC list.
Basically the goal is to be able to do high-performance computing in something other than C and something like Scheme, Common Lisp, Python, Ruby, and Arc. Two notations, ala Dylan, so one can go more succinct at the cost of extra syntax and !%^[]#!@ stuff.
> Well, it will get better. Mailing list was shown on a slide, but I > thought the above link would be enough to remember. You might pester > Perry on the LispNYC list.
> Basically the goal is to be able to do high-performance computing in > something other than C and something like Scheme, Common Lisp, Python, > Ruby, and Arc. Two notations, ala Dylan, so one can go more succinct at > the cost of extra syntax and !%^[]#!@ stuff.
> kt
New Lisp dialects seem to be the new hobby. They are mostly incompatible to existing and emerging dialects.
On 16 Aug, 09:40, Rainer Joswig <jos...@lisp.de> wrote:
> New Lisp dialects seem to be the new hobby. They are mostly > incompatible to existing and emerging dialects.
This makes one worry that no replacement for CL will ever become sufficiently widespread as to be truly viable. Is it perhaps in the nature of people who are attracted to Lisp that they tend to go off each in his own direction?
> Any others I have missed (and is not L#, Arc, 'newLISP', > Croma, Goo)?
> On 16 Aug, 09:40, Rainer Joswig <jos...@lisp.de> wrote:
> > New Lisp dialects seem to be the new hobby. They are mostly > > incompatible to existing and emerging dialects.
> This makes one worry that no replacement for CL will ever become > sufficiently widespread as to be truly viable. Is it perhaps in the > nature of people who are attracted to Lisp that they tend to go off > each in his own direction?
Perhaps it's in the nature of Lisp. This seems to be one more characteristic Lisp has in common with Forth - ease of creating at least an interpreter and very uniform syntax may be the main culprits.
> > Any others I have missed (and is not L#, Arc, 'newLISP', > > Croma, Goo)?
are wrote: > On 16 Aug, 09:40, Rainer Joswig <jos...@lisp.de> wrote: >> New Lisp dialects seem to be the new hobby. They are mostly >> incompatible to existing and emerging dialects.
> This makes one worry that no replacement for CL will ever become > sufficiently widespread as to be truly viable. Is it perhaps in the > nature of people who are attracted to Lisp that they tend to go off > each in his own direction?
Yes, and a few cannot resist the temptation. It seems to be part of the learning curve at some stage, and it seems to take considerably more learning to understand that developing a new dialect doesn't really pay off, or requires an enormous amount of work to make it _substantially_ better than existing ones.
Pillsy wrote: > On Aug 16, 4:40 am, Rainer Joswig <jos...@lisp.de> wrote: > [...] >> Any others I have missed (and is not L#, Arc, 'newLISP', Croma, >> Goo)?
Discussion subject changed to "NYC LOCAL: Tuesday 14 August 2007 Lisp NYC: Perry Metzger on Otter, the new salty slick whiskered snow-bellying programming system" by Perry E. Metzger
Ken Tilton <kennytil...@optonline.net> writes: > Basically the goal is to be able to do high-performance computing in > something other than C and something like Scheme, Common Lisp, Python, > Ruby, and Arc.
Not really. :)
> Two notations, ala Dylan, so one can go more succinct at the cost of > extra syntax and !%^[]#!@ stuff.
Actually, there is only one notation, but I added on some new syntactic sugar (similar in flavor to other forms of lisp syntactic sugar, such as the use of 'a to mean (quote a).
The reason for the new syntactic sugar is to make it easier to deal with complicated data structures using notation compact enough to make them convenient. I think the video of my talk explains it pretty well.
The web site, http://www.otterlang.org/, currently is empty, but that should change in a week or so. Meanwhile, one can subscribe to the announcements mailing list by sending mail to announce-requ...@otterlang.org
Rainer Joswig <jos...@lisp.de> writes: > New Lisp dialects seem to be the new hobby. They are mostly > incompatible to existing and emerging dialects. Propon...@gmx.net writes: > This makes one worry that no replacement for CL will ever become > sufficiently widespread as to be truly viable. Is it perhaps in the > nature of people who are attracted to Lisp that they tend to go off > each in his own direction?
New Lisp dialects are not a new hobby -- they are a very old hobby. From 1960 to about 1985, new lisp dialects flourished, and with them new ideas and new visions of what Lisp could do and what it could become.
On a regular basis, new Lisps would appear, and with them, new frontiers were crossed. Lexical scoping, object systems, even such basic tools as strings as a first class type, were tried out in new dialects, and ultimately spread and made the community stronger.
Since the development of Common Lisp and the AI winter, it has become unfashionable to create new dialects, and a dim view has been taken of experimentation. Some claim that new dialects divide the community, and that there is nothing to learn from the new dialects. Orthodoxy has set in, as well as stagnation, and the main dialect has remained largely untouched in fifteen years, on a basis that is 25 years old. The community has also, sadly, turned inwards and looked back to the past instead of towards the future -- one hears constant laments for Lisp Machines and other parts of the glorious golden age of yore, instead of about what we can do next.
The emergence of new dialects of Lisp is not a sign of trouble. It is a sign that the community is finally regaining its health. Springtime has come at last after a long ice age, new flowers are growing where the glaciers have receded. I'm glad of it.
Let a hundred flowers bloom: let a hundred schools of thought contend.
> Rainer Joswig <jos...@lisp.de> writes: > > New Lisp dialects seem to be the new hobby. They are mostly > > incompatible to existing and emerging dialects.
> Propon...@gmx.net writes: > > This makes one worry that no replacement for CL will ever become > > sufficiently widespread as to be truly viable. Is it perhaps in the > > nature of people who are attracted to Lisp that they tend to go off > > each in his own direction?
> New Lisp dialects are not a new hobby -- they are a very old > hobby. From 1960 to about 1985, new lisp dialects flourished, and with > them new ideas and new visions of what Lisp could do and what it > could become.
> On a regular basis, new Lisps would appear, and with them, new > frontiers were crossed. Lexical scoping, object systems, even such > basic tools as strings as a first class type, were tried out in new > dialects, and ultimately spread and made the community stronger.
I doubt that you need a new Lisp dialect to try out dot notation for access of array data. Not in 2007. Not in 1980.
> Since the development of Common Lisp and the AI winter, it has become > unfashionable to create new dialects, and a dim view has been taken of > experimentation.
The experimentation has moved to a different level. Once you have a stable base language, you still can experiment with libraries, interfacing, tools, ...
> Some claim that new dialects divide the community, > and that there is nothing to learn from the new dialects.
I won't say that there is nothing to learn from new dialects. The implementor of a new dialect will learn how difficult it is to create a good language, a good implementation, good documentation and good libraries. Maybe some person years (ten?) to get basics going. I have high respect for that.
> Orthodoxy > has set in, as well as stagnation, and the main dialect has remained > largely untouched in fifteen years, on a basis that is 25 years > old. The community has also, sadly, turned inwards and looked back to > the past instead of towards the future -- one hears constant laments > for Lisp Machines and other parts of the glorious golden age of yore, > instead of about what we can do next.
Personally I'd like to move back and forward at the same time. There is the vision of a workstation that is built for the (Lisp) programmer. From the ground. Then there is the vision of a software environment that is powerful enough to support programming for most current purposes. Then there is the vision of being able to develop state-of-the art software that integrates with current operating systems, so that you get the power of interactive programming even for those. These visions have some overlap. Some people are trying to work on different versions of the vision. For example to make Lisp implementations good citizens in a GNU/Linux environment (there is a log activity in this area). But these different visions can be moved forward while using the same base language. It makes the exchange of ideas and code much easier.
I'm personally totally uninterested in 'new' 'languages' that focus on micro-advancing the state of syntactic expressibility. Also, if a new language is developed without software, it is not worth it. If it claims to be better, more modern, whatever, it should be able to demonstrate that and to compete with other Lisp systems. I'm somehow interested in a Lisp implementation, but I'm MUCH more interested in using/writing software applications written and extended in (Common) Lisp.
That's also what has happened. Quite a few people have 're-' discovered Lisp and tried to apply it to some (new) domains. There are now several active implementations and many people able to maintain them. There are now people experimenting how to develop software, what the environment should look like for that (given the current surroundings), what tools one needs in the Internet age (project sites, chat channels, web forums, bots, paste boards, ...).
> The emergence of new dialects of Lisp is not a sign of trouble. It is > a sign that the community is finally regaining its health. Springtime > has come at last after a long ice age, new flowers are growing where > the glaciers have receded. I'm glad of it.
For me the sign of spring would be Lisp software, not Lisp dialects.
På Sat, 18 Aug 2007 04:05:30 +0200, skrev Rainer Joswig <jos...@lisp.de>:
> For me the sign of spring would be Lisp software, not Lisp > dialects.
I see both. At least in the sense that there has been a explosion in the number of libraries in the last few years. Perhaps there is room for some new dialects?
In article <op.tw8no0fqpqz...@pandora.upc.no>, "John Thingstad" <john.things...@chello.no> wrote:
> På Sat, 18 Aug 2007 04:05:30 +0200, skrev Rainer Joswig <jos...@lisp.de>:
> > For me the sign of spring would be Lisp software, not Lisp > > dialects.
> I see both. > At least in the sense that there has been a explosion in the number of > libraries in the last few years. > Perhaps there is room for some new dialects?
But what for? What is not possible with CL or Scheme, ISLISP? You could even revive some old attempt like EuLisp and Dylan. More thought went into them than you can ever hope to have time.
(Side note: Perry should not be discouraged. If he wants to do that, he will do it anyway. But some criticism/discussion upfront can't be bad? ;-) ).
His plan is basically not realistic for a single person. Getting a cool Lisp language and a high-performance implementation is certainly a nice goal. But far from easy to achieve. He admits that in his presentation. Let's see how far he gets.
Influenced by Scheme, Python, Arc, Common Lisp, and Perl.
So that basically means you get a bunch of ideas which are not easily made to work well together.
It has been too long since there has been a new Lisp
This misses all the other attempts. EuLisp, ISLISP, Dylan, plus quite a lot of smaller attempts like e7, l#, Pico Lisp, and a bunch of others... Plus new versions of Scheme (the core) with lots of new language extensions (see SRFI).
(! somehash["myname"].val "Perry")
Please. You need a new dialect for that? In which way is that 'radical'?
Why can't I have fun too?
That's a valid reason.
Optimizing compiler good enough for heavy math (like computational chemistry).
That's tough.
Interpreter + system & text hacking features good enough that no one need ever use Perl again.
Currently if you use CL or something like scsh, you won't need that much PERL.
If people keep having to go to the reference manual for something, that something is badly designed.
You can only say that if you start. After a while any project needs documentation or easy to understand source code (self documenting).
no format strings
Fine for me. For that I don't need a new langauge. Just programmer guidelines and a few lines of library code.
def instead of defun
Talking about 'radical'.
compact names?
Obfuscation built in. Great. Not.
It is too difficult to use real data structures in Lisp.
Is it?
REGEXPs
I thought that topic is done. Edi Weitz has a CL lib for that. Olin Shivers explained how REGEXPs could be better done in Scheme (1998).
Compilation, No rebinding of identifiers in other modules
See the ANSI CL pages about compilation.
And so on. I don't find most of that very radical. There is nothing which you can't do with a bit of coding for, say, SBCL to have it look like what he describes.
In article <op.tw8no0fqpqz...@pandora.upc.no>, "John Thingstad" <john.things...@chello.no> wrote:
> På Sat, 18 Aug 2007 04:05:30 +0200, skrev Rainer Joswig <jos...@lisp.de>:
> > For me the sign of spring would be Lisp software, not Lisp > > dialects.
> I see both. > At least in the sense that there has been a explosion in the number of > libraries in the last few years.
That's definitely good. Some consolidation might be needed, plus some more apps using it. But there is lots of progress, that I think new dialects are not that much needed. What could be more interesting is to take the implementations and improve them in some (multi-core, distributed computation, better maths performance, better app packaging, ...) or create some nice 'distributions' with libs, environment and compiler.
> Perhaps there is room for some new dialects?
I have been reading the latest discussions about the new Scheme R6RS.
To me R6RS looks like a new language. It moves further away from Lisp ideas and morphs into something that is hard to describe.
There again with R6RS is a language where the motivations for the design are not clear to me and are not proven by experimentation.
> Please. You need a new dialect for that? In which way > is that 'radical'?
Even though I don't know much about language design, I don't understand what is wrong with prefix syntax like (aref myarray 12). I find it more convenient than the "traditional" myarray[12], and I have worked in languages that use the latter syntax (C, Fortran, R, Mathematica) for years. I don't want to start another prefix-infix thread, but to me the Otter syntax would not be an improvement.
Moreover, the [] syntax misses the point that sometimes there are extra arguments besides the key or the index. Eg
(gethash (cons i j) sparse-matrix 0)
that I used for sparse matrices. You need to use different operators, or push all kinds of ugliness into []. See where this lead for R:
There are many badly designed Lisps around, eg Lush (dynamic scoping! *shudder*). OTOH, I find CL extremely well designed: whenever I run into something I consider a shortcoming, a closer look reveals it was not bad choice.
På Sat, 18 Aug 2007 10:25:05 +0200, skrev Rainer Joswig <jos...@lisp.de>:
> But what for? What is not possible with CL or Scheme, ISLISP? > You could even revive some old attempt like EuLisp and > Dylan. More thought went into them than you can ever > hope to have time.
Because we are not still solving the same problems. I was thinking more about dotlist (http://dotlisp.sourceforge.net) A Lisp written over CLR and with full access to .NET. Naturally .NET class system is radically different from CLOS and it would be interesting to explore how to do this. Then perhaps bring some of the experience back to Common Lisp.
John Thingstad wrote: > Perhaps there is room for some new dialects?
It is very hard to create a Lisp dialect that is _substantially_ better than the ones we already have (Common Lisp, Scheme, ISLISP, Dylan and some others). Yes, it is possible to make shallow improvements, but almost all such improvements can be made from within those dialects, most of the time in a very straightforward way. (I would consider an improvement to be substantial only if it cannot be done from within one of those dialects.)
I don't see a value in new Lisp dialects that are only better at such superficial levels. What's even worse is that these new Lisp dialects are typically implemented from scratch, which means that decades of experience with efficient implementation techniques are lost and have to rediscovered. Only few examples take the 'right' approach and embed new dialects in existing ones - but due to the Lisp nature, it's hard to consider them really new dialects, they are typically considered "just" libraries. [1]
Are libraries not sexy enough? Is that the reason why inventors of new Lisp dialects prefer to start from scratch? Why do they put up with implementing new Lisp dialects in languages like C, when they could as well take advantage of Lisp from the start?
Where do you see value in having new Lisp dialects?
Pascal
[1] Is UCW a new Lisp dialect? Cells? ContextL? Qi?
Pascal Costanza <p...@p-cos.net> writes: > [...] > Are libraries not sexy enough? Is that the reason why inventors of new > Lisp dialects prefer to start from scratch? Why do they put up with > implementing new Lisp dialects in languages like C, when they could as > well take advantage of Lisp from the start?
> Where do you see value in having new Lisp dialects?
There's even another dimension that's underexploited IMO, it's that Common Lisp allow for subsets of the language.
A lot of these "new" lisp implementations would be much more interesting if they were developed as a subset of Common Lisp (with possibly some implementation specific packages).
NOTE: The most fundamental particles in this product are held together by a "gluing" force about which little is currently known and whose adhesive power can therefore not be permanently guaranteed.
Pascal Bourguignon wrote: > Pascal Costanza <p...@p-cos.net> writes: >> [...] >> Are libraries not sexy enough? Is that the reason why inventors of new >> Lisp dialects prefer to start from scratch? Why do they put up with >> implementing new Lisp dialects in languages like C, when they could as >> well take advantage of Lisp from the start?
>> Where do you see value in having new Lisp dialects?
> There's even another dimension that's underexploited IMO, it's that > Common Lisp allow for subsets of the language.
> A lot of these "new" lisp implementations would be much more > interesting if they were developed as a subset of Common Lisp (with > possibly some implementation specific packages).
I'd consider ISLISP an example of a subset of Common Lisp. And I think ISLISP would be a good starting point for such dialects, at least better than starting from scratch.
> Pascal Bourguignon wrote: > > Pascal Costanza <p...@p-cos.net> writes: > >> [...] > >> Are libraries not sexy enough? Is that the reason why inventors of new > >> Lisp dialects prefer to start from scratch? Why do they put up with > >> implementing new Lisp dialects in languages like C, when they could as > >> well take advantage of Lisp from the start?
> >> Where do you see value in having new Lisp dialects?
> > There's even another dimension that's underexploited IMO, it's that > > Common Lisp allow for subsets of the language.
> > A lot of these "new" lisp implementations would be much more > > interesting if they were developed as a subset of Common Lisp (with > > possibly some implementation specific packages).
> I'd consider ISLISP an example of a subset of Common Lisp. And I think > ISLISP would be a good starting point for such dialects, at least better > than starting from scratch.
It was not that unusual to base a Lisp on a subset of CL.
Interleaf Lisp is similar to CL. (now called Quicksilver, a publishing suite for large scale documents) " "Interleaf 6" is a document authoring and composition package. It provides an integrated set of tools for creating compound documents: word processing, graphics, data-driven business charts, tables, equations, image editing, automated page layout, book building- including automatic index and TOC, conditional document assembly. It includes several features engineered to support the production of large and complex document sets, including: centralized control over parts or all of a document (format and/or content), global search and replace/change on individual graphics objects regardless of specific orientation or position, revision management.
Also available (on some platforms) is the optional Developer's Toolkit (DTK) for customizing or extending the capabilities of the above authoring tool. Developer's Toolkit is used to write programs in Interleaf Lisp. Interleaf Lisp is similar to CommonLISP, but it also contains an extensive set of classes, methods, and functions for examining and changing almost all Interleaf objects, including documents and their contents. DTK includes an editor, debugger, compiler, listener, interpreter, and on-line documentation. Lisp code developed with DTK, or even written with an ordinary editor, can be executed by the stock system, so that customization or the provision of special functionality is not limited to installations with DTK. In fact, much of the distributed system is written in Lisp."
Discussion subject changed to "NYC LOCAL: Tuesday 14 August 2007 Lisp NYC: Perry Metzger on Otter, the new salty slick whiskered snow-bellying programming system" by Ken Tilton
Perry E. Metzger wrote: > Ken Tilton <kennytil...@optonline.net> writes:
>>Basically the goal is to be able to do high-performance computing in >>something other than C and something like Scheme, Common Lisp, Python, >>Ruby, and Arc.
> Not really. :)
You better go read your slides.
>>Two notations, ala Dylan, so one can go more succinct at the cost of >>extra syntax and !%^[]#!@ stuff.
> Actually, there is only one notation, but I added on some new > syntactic sugar...
So it's a two-notation system with one-notation. Excellent.
> The web site, http://www.otterlang.org/, currently is empty, but that > should change in a week or so.
Is your objective to avoid as much as possible the adoption boost you would get from having the web site ready at the same time you are driving people to it with public appearances, NG postings, and that Times Square billboard I saw? (hint)
