The Next Lisp: Back to the Future | Hacker News Reader

The Next Lisp: Back to the Future

lambdassociates.org

36 pointsparenthesis13 comments

But it seems Mark Tarver (the author) is ending his work on Qi:

http://groups.google.com/group/Qilang/browse_frm/thread/5927...

I find it amusing that although Mr. Tarver himself seems to admit his ideas on the future of Lisp could be better presented by omitting his strawman arguments supposedly debunking the "ecomonics" of open source efforts, he nonetheless can't seem to restrain himself from including them.

My suggestion to Mr. Tarver: if you want me to take your language ideas seriously, save your "opinionating" until after your technical ideas are widely adopted. As much as I'd like to try out a promising language like Qi, an ideological rant against open source won't help convince me or anyone else it's a good idea.

And to make his rant painful on top of unconvincing, he confuses free-as-in-speech with gratis.

No. He starts talking about FOSS - MIT and BSD stuff. This is nearly in every case free economically. And the GPL stuff is almost always free economically too. In practice if you give people complete freedom over sources you really have to give them the same freedom you have to change and distribute. So by undercutting price tends aymptotically to zero. People know that and so they make it free from the outset. They make their money from hand-holding.

If that's what he actually meant to say, then it's a pity he didn't use those words since they make much more sense.

My understanding of what he said was - to paraphrase and re-interpret a bit - "Free software is BS, because nothing is (as you put it) 'free economically'."

That's two meanings of the word 'free' in one proposition. The first refers to the rights of the users, and the second refers to putting in labor or cash to create something. The fact that the two might be closely related in practice doesn't matter. For this to work, you'd have to make the additional claim "Free-as-in-Speech is identical to Free-as-in-Beer", which would be insane. See: OpenBSD's cd sales.

Additionally, he's throwing in a third concept, which we could safely call "Free-as-in-Lunch".

What he seems to believe is "Free-as-in-Speech software is BS because nothing is Free-as-in-Lunch". Which I still disagree with, but I think would be a more direct way of expressing the point without using confusion about "free" to a rhetorical advantage.

Actually, the biggest problem is his confusion of free (as in price) with cost-less. Nothing is or can be cost-less, but that has little to do with the price charged for it.

Nonsense. This is tantamount to saying that unless a person's ideas are widely accepted or he's rich he should shut up if his ideas are not popular. I don't think he cares much whether you use Qi or not.

He's linked the two essays because he sees there is a connection between funding and development and he's right.

I suspect you too must have strong feelings on the subject of open source economics since you've clearly misunderstood and mischaracterized my point: airing these views doesn't help promote Qi so he shouldn't link them all together.

I'm not an expert on lisp or functional languages so I'm willing to give him the benefit of the doubt on the first two pages where he presents the design tradeoffs of Qi. Then on the next page instead of learning more about Qi, I find myself in the middle of an argument where his word choice confuses the distinction between free/unsupported open source and free/unencumbered open source. WTF? Didn't he just tell us on the previous page one of his strategies for Qi development was "3. Infiltrate don't confront".

And then I get to learn how he's decided to resign his job and become a wandering Taoist.

>> 1. Recursion as the primary mean of expressing procedure calling.

Does anyone know what that statement means?

The author misses the real essence of Lisp, which is that:

1) a single data structure, the heterogeneous list, can be used to express everything in the language, including the program itself.

2) the linear nature of lists makes it easy to recurse on them, thereby enabling significant use of recursion as a control structure, thereby simplifying the reasoning process about the code.

3) the simplicity and pervasiveness of the list structure makes it easy to write macros in the language, thereby increasing the range of abstractions that the language can express.

4) bindings are dynamic, increasing language flexibility and making the use of a REPL even more valuable.

I think his point 1 was your point 2.

Makes some good points, but calling Python a sugaring on Lisp is a bit weird, especially since it misses 3 of his 5 philosophies of Lisp.

Do I hate that the only implementation of Qi is not fully FOSS? Yes.

Do I love this essay? Yes.

Reply on news.ycombinator.com