Comments on robjsoftware.org: A Growable Language Manifesto

Very informative post on growable language... keep...

2009-06-08T00:18:56.494-07:00

Very informative post on growable language... keep it up Rob !!!

Software Development Company

I commend you on seeing the true future of program...

2007-12-27T11:31:19.918-08:00

I commend you on seeing the true future of programming! :)

My half-baked attempts along similar lines have shown me the immense barrier to implementation which, in my humble opinion, dwarfs other implementation concerns by an order of magnitude. You mention the need for an "analysis engine", but *only* say it is the "single most technically ambitious goal". You are technically correct, but the "analysis engine" deserves a diatribe detailing it's massive scope, if only to impress upon others that this is more than a whole week's worth of prototyping.

From what little compiler optimization work I have done, books I read, and the compilers I tried to understand, there appears to be little or no interest in implementing higher-level optimizations, and the powerful algorithms needed to support those transformations.

I had not heard of JastADD until you pointed it out, but it is a toy compared the heart of other optimizing compilers. Although JastADD is useful for known lattice operations, I doubt the extensibility will be useful for the higher level optimizations that have less to do with control flow and more to do with theorem proving. I might be wrong though.

In any case, keep up the good thoughts. If you, or someone you know, are tackling the "analysis engine" please let me know (kyle@arcavia.com). If I can't contribute, at least I can watch something wonderful being built.

Thanks

"Axiomatic language" may satisfy some of these goa...

2007-12-24T07:08:36.058-08:00

"Axiomatic language" may satisfy some of these goals. It is a minimal, extensible meta-language in which language features and paradigms could be defined. A type system, for example, is not built-in, but would have to be defined within the language.

See http://www.axiomaticlanguage.org/

watch the included video:complete computing system...

2007-12-12T15:39:18.994-08:00

watch the included video:

complete computing system in 20000 lines

Nice, Rob. It will be interesting to see where you...

2007-12-11T20:55:11.787-08:00

Nice, Rob. It will be interesting to see where you go from here!

BTW, some of your links came out bad -- specifically the ones pointing at your own blog posts.

For fine-tuned control on static typing, you might...

2007-12-11T07:05:03.490-08:00

For fine-tuned control on static typing, you might be interested by Qi: http://www.lambdassociates.org/

You might be interested by metalua (http://metalua...

2007-12-11T06:35:16.049-08:00

You might be interested by metalua (http://metalua.luaforge.net) or Converge (http://convergepl.org).

Metalua has most of the features you're expecting, except those related to static type checking. Or rather, nothing has been published yet about static type checking in Metalua ;-)

Imam, I wouldn't say I'll have no one with me. I'...

2007-12-10T22:14:18.434-08:00

Imam, I wouldn't say I'll have no one with me. I've posted this to a thread on Lambda the Ultimate, which is getting a fair bit of discussion. Also, all my references are to projects that are not dissimilar to what I have in mind.

metaprogrammer, that's very interesting and definitely right up my alley! I will certainly check it out in detail! Thanks :-)

As for "hard to learn" comment: it is a problem, b...

2007-12-08T05:09:32.643-08:00

As for "hard to learn" comment: it is a problem, but the same problem exists for a very fixed and simple languages. To learn Java you have to learn all that libraries and frameworks (including some DSLs, like for Hibernate). It is not easier than learning a bunch of well-designed domain specific languages.

So I support the author in his struggle. I am looking forward to see the result.

I wish you good luck in your project. This is a di...

2007-12-07T17:34:09.308-08:00

I wish you good luck in your project. This is a difficult task and you will have no one with you (at least, not before you get totally frustrated). A growable language is by definition hard to learn, have you considered that? I mean, the language might be easy to understand, but there will be so many libraries written in so many different styles that the question will be lifted from "learning the language" to "using this form of the language".

At any rate, apart from the syntactic and metaprogramming facilities you want, I think Scala (www.scala-lang.org) might help you a little. It has type inference, and in my opinion, prototyping is rather fast in it. But the main reason for me to ask you to take a look at Scala is, you probably want to borrow some implementation tricks from them. It pretends to be a parametrically polymorphic language, whereas actually it compiles to Java(/.NET) bytecode, so in that sense it is in no way "better". In a sense it's typing is someone gradual. You write your code in a very strongly typed language but in the end some of it gets erased so what you guarantee is that those parts written in your language guarantee more static checks (which is, in fact, the point of static typing). So you might be interested to compile, in the end, to a dynamically typed language. But your compiler, when optimizing, should make use of theorems that it can prove from the type information. This is not hard, but risky. So that if your theorems are wrong (suppose you didn't think of concurrency when you "proved" it) you might get into real trouble. Then again, you have to take risks in your life.

You may want to take a look at this growable langu...

2007-12-07T07:22:16.857-08:00

You may want to take a look at this growable language base:

http://www.meta-alternative.net/techpreview.html

Great post--and great ideas by the way. The seman...

2007-12-04T16:17:53.442-08:00

Great post--and great ideas by the way. The semantic web is doing pretty good with getting closer to this, people are programming without knowing it. That's cool.

HaHa.. Pity,IP's not out yet. Their demo looks good, I don't know what the hold up is...the date on this ad is !!2000!!?

http://youtube.com/watch?v=tSnnfUj1XCQ