Talk:NewLISP

From Wikipedia, the free encyclopedia

[edit] Userbox for NewLISP fans

For those of you who are interested, I've created a NewLISP userbox that can be added to your user page.

Required code:

{{User:Tylerdmace/Userboxes/newlisp}}

--Tylerdmace 11:20, 15 October 2007 (UTC)

[edit] First shot at cleanup

I'd have to say that it's about due time someone took a look at newLISP here at wikipedia :) --HoH 23:40, 16 May 2006 (UTC)

LIke the logo! This entry gets occasional updating, and I added a few lines back in March. I wasn't sure, though, how much information generally went in Wikipedia about a prog language, if there's plenty of information on external sites. (Which there is!) I could write a book but wasn't sure whether to... :-) Cormullion 22:04, 17 May 2006 (UTC)

Anything you got that will further this article, is welcomed. Thanks for the previous work too! --HoH 00:19, 18 May 2006 (UTC)

I'm adding a request for cleanup, starting today. --HoH 19:07, 20 May 2006 (UTC)

[edit] The TODO List

Possible ideas (please comment on):

• Add History Section - I'm emailing Lutz for this information as the site/documentation provide little
• Further infobox information - The structure is there for more information, I just didn't know the facts
• Add Philosophy Section - I think one of the strong points of newLISP is it's unique philosophy, and we should portray that if possible
• Syntax Examples - What do you guys think?
• We should bust out current sections into subsections of a Features section
• Implementations/Available Platforms section
• Notes, See Also, and External Links sections
• A section discussing the future of newLISP - Not too sure if I'm on board with this idea, but it's a suggestion none-the-less

I immediately thought:

• Add Philosophy Section!

NewLISP is obviously neither a Scheme nor a CommonLISP (??), and so it's something else. In the text I can figure out that CommonLISP is created for lightweight and productivity. But that need not be the main philosophy. Said: Rursus ☺ ★ 18:59, 24 July 2007 (UTC)

About the future of newLISP, there's some policy of Wikipedia that makes divinations and prognoses unfit for an article. However: if the "future of newLISP" regards planned features, then the plan is current, not future - an so a plan is a fact as good as any. Said: Rursus ☺ ★ 19:02, 24 July 2007 (UTC)

[edit] Older

hey guys, a working knowlege of tk is not assumed. Newlisp is a language, so what the article calls a gui is really the ide. You can use newlisp, and be productive with it, without tk.

corrected - added GUI programming

[edit] "Excellent language" for distributed computing projects?

A recent edit to the article added the claim that:

"newLISP is an excellent language for implementing distributed, agent-based computing projects, where dozens of different newLISP instances run concurrently and communicate with one another over a network, in response to changing demands."

From the perspective of someone interested in distributed/concurrent computing and programming languages, this claim seems a bit uninformed, to say the least: newLISP is many things, but a distributed programming language it is not.

Under the impression that it would be a fairly uncontroversial correction, i removed the claim from the article, with an explanatory comment. However, Cormullion disagreed on my talk page, and eventually reverted the change without, i feel, answering any of the objections i raised in the ensuing discussion. Could interested users please read the conversation there, and help establish consensus on this? --Piet Delport 00:18, 5 June 2006 (UTC)

"Excellent" and "Ideal" are unattributed, sounds like an advertisement, and value-laden and has no place in wikipedia in this case. See WP:NPOV. I've made a suggestion for rewording. Jens Nielsen 14:02, 5 June 2006 (UTC)

I agree. Maybe 'optimized for' (if true), or 'described by Mister Big as excellent', if notable and cited. Tom Harrison ^Talk 14:10, 5 June 2006 (UTC)

I would agree, but as far as i can tell, neither of those are the case. Like most of its peers (Io, Lua, Pike, Python, Ruby, Scheme...), newLISP does not provide any particular support or consideration for distributed programming techniques aside from simply being a general-purpose high-level dynamic programming language. I think that suggesting otherwise is misleading, at best, and belittling of the efforts of other distributed programming languages, at worst. --Piet Delport 16:21, 6 June 2006 (UTC)

Piet, I disagree -- I believe that newLISP is extremely easier to use for distributed programming than any of the peers you listed. Specifically, you can send an entire program for one newLISP node to another newLISP node using a single command net-eval [1]. Most other programming languages require some libraries to be loaded, handshaking, etc. The ease in which I can distribute newLISP processing across any number of nodes is, IMHO, stunning. In addition I'd point to the entire net-* series of functions as evidence of the ease of use of networking and distributed computing within newLISP. [2] Full disclosure: I am heavily involved with the development of newLISP - so I will leave it to other wikipedians to establish the NPOV on this. --Sarken 16:34, 15 June 2006 (UTC)

The only language features demonstrated by net-eval are support for networking (sockets) and run-time code serialization/evaluation, and the only functionality it achieves is simple (some would say "trivial") request/response-style remote code execution. I'm not sure i understand what makes newLISP's implementation of this so stunning; the same is not any more difficult to do in Scheme and Python, for example, aside from maybe implementing a wrapper function with the same signature as net-eval. (I don't know the other languages well enough to be sure.)

However, the point i'm trying to make is not about whether newLISP's support is better or more unique than its peers (and i don't think it is); but rather that that this functionality is not notable to begin with: there is much, much more to distributed/concurrent programming than remote evaluation. For some pespective, you should look at a few actual distributed programming languages, the problems they face, and the techniques they employ.

To attempt an analogy, this claim is similar to saying that (hypothetically) that C is an excellent language for object-oriented programming because it supports structs and pointers (for composition and indirection, respectively): while it's true that those primitives can, and are, used to implement powerful and fully-featured object-oriented programming environments^[1], you still cannot put C in the same class as "real" object-oriented languages (let alone characterizing C's object-orientation support as "excellent", among them).

[1] just like various kinds of distributed programming features can, and are, implemented on top of Lua [3], Python [4], Scheme [5][6], and newLISP, by analogy

--Piet Delport 05:32, 16 June 2006 (UTC)

You make some valid points about other distributed languages solving larger problems. eg. Plan 9 is a distributed operating system with a default distributed FS, whereas newLISP provides easy way to distribute computation across distinct operating systems. May I suggest changing "excellent" and "ideal" to something less grandiose. Perhaps "newLISP can easily implement ..." I maintain that the net-eval functionality is much easier to use than other high-level programming languages -- it may simply be remote execution, but it allows by default a single node to send multiple requests to other nodes and collect and collate the responses (note func-handler) [7]. This can be done in other high-level languages but it usually requires writing all the socket management and interrupts to handle multiple nodes responding to the master node. I believe this construct is unique to newLISP. Does anyone know if there is something like this in similar languages (Ruby/Python/Perl)?

--Sarken 15:33, 23 June 2006 (UTC)

I think you greatly overestimate the effort required to implement net-eval in other languages. It's not the kind of thing you would see outside of a toy cookbook example, probably. --Piet Delport 02:16, 28 June 2006 (UTC)

I'm going to go ahead and remove the claim again in the next day or two, if no one still objects. --Piet Delport 22:48, 21 June 2006 (UTC)

[edit] Name

The name of the language should be written Newlisp, newLISP is logo; even trademarked. 89.172.56.89 (talk) 12:38, 29 May 2008 (UTC)