News for hackers. Stuff that matters.
Re:Why Lisp Will Never Win (2003) by Jd on Saturday November 05, 2011 @02:02PM (#3200405)
I think the general attitude of lispers is frequently a large turn off. There is a lot of elitism, but it is often a needy Mensa-like elitism. There is almost a hidden plea attached "please respect us as the smartest" and the problem is, of course, that they can't point to any tangible, quantifiable method for gauging intelligence with respect to a language. I guess PG's "let's see if anyone is using it in 50 years" is a good measure in some case, but that implies that lisp as a language is less of a "computer language" (e.g. a language that is made for making machines do what we humans want them to) and more of an artform or philosophical/mathematical depiction of some deeper truth obscured to normal people.

I think this is ultimately always a problem between ultra-smart people who live in something of an abstracted world that frequently approaches fantasy and the reality of industry, which generally doesn't care about how pretty something is unless someone will buy it because it is pretty. That's a pretty large gap, and, while in my opinion there is no need for arrogance, it is pretty clear why most of the smartest people in history end up on the former side of that divide. They want to contemplate some sort of eternal idea which usually they ascribe beauty to (e.g. Bohr, Einstein, etc.) -- and yet hackers, unlike theoretical physicists, are halfway between the people probing the mysteries of the universe and those putting together blueprints for the atom bomb.

In my opinion that's just life. Einstein hated the bomb and rightly so in my opinion -- but that doesn't mean that it was any less a necessity for the Manhattan project to go forward. The sad thing, however, is that lots of brilliant people trying to navigate this divide end up with a huge chip on their shoulder. PG talks about this in H&P; the jocks were popular and (surprise!) the nerds were not. So you go out and get a few million and say, "ha, screw the jocks." Fine, generally speaking I think that is a good idea, but let's not turn it into a wanker fest since there are lots of needed functions in the world that should probably be performed by those who can't understand Lisp -- and I don't just mean cleaning toilets, I mean politics.

Abstraction, which is really what lisp is about in many ways, is only ever IMO one pole. You have to zoom out to see the bigger picture and zoom in to address the minutia. Generally people miss the forest for the trees, so I have every sympathy for those who stick with the forest and ignore the trees. But when it comes down to it, someone has to re-write your code in blub to make a functional product (even true with Viaweb) so have some sympathy for the little man who has to do a sh't job, probably his mother never gave him fairy tales to read and he only vaguely knows who Godel, Escher, and Bach are. That's sad, but that's life and we who are Americans should especially expect it as an immigrant nation.

Anyways, I'm one who discovered Lisp rather late despite and not because of the Lisp community as it is usually presented. I think it is beautiful and esp. I think SICP is a fantastic book to introduce one to computer science concepts and I think that industry should serve great passions and ideas and not vice versa. Hell, I even think lisp machines are a great idea and hope someone builds new ones and they take over the world. But despite all of that, I would never hang out with "the Lisp community."

Why? I want to build great products. Great products need to scale. If you want to go to the moon you need a f'ing spaceship and not emacs and an open source notebook. Spaceships are big. They take a lot of people to run, and they have to be smart people. One mistake and you are smithereens at Cape Canaveral.

The thing that irks me the most about all of this whining is that America used to be capable of building automobiles and space shuttles and (omg!) computers. It was amazing. Now it is hello Taiwan, goodbye America. People talk about "Rework" and how they can't imagine being in a company with more than ten people and glorify the life of tribes in third world countries without contraception or toothpaste (or, in pg's case, the lifestyle of lions on the African prairie).

I know it may be nuts to say this, but I think we need to build cool stuff for the sake of building it, and that our tools should be our tools -- not some magic fairy dust. Use whatever works to solve the problem. Get your hands dirty. A parentheses is not a hammer if what you've got in front of you is a nail.

And if what you've got in front of you is a problem bigger and nastier than that, build a bigger and better tool. There's no problem that can't be solved if people put aside their differences and work together.

Re:Why Lisp Will Never Win (2003) by sedachv on Saturday November 05, 2011 @06:00PM (#3201047)
> If you want to go to the moon you need a f'ing spaceship and not emacs and an open source notebook. Spaceships are big. They take a lot of people to run, and they have to be smart people. One mistake and you are smithereens at Cape Canaveral.

The moon is not that far away from orbital flight. These two guys might disagree with you: http://www.copenhagensuborbitals.com/

Re:Why Lisp Will Never Win (2003) by nknight on Sunday November 06, 2011 @01:33AM (#3201954)
That site indicates they're attempting to launch a human on a suborbital trajectory. What does that have to do with either orbital flight or going to the moon?
Re:Why Lisp Will Never Win (2003) by rbanffy on Sunday November 06, 2011 @04:39AM (#3202261)
If you can get to space, even a suborbital flight, you must have gotten most of the life support right. If you reach orbit, you got your rocket right. If you can change orbits, you got manevering right. If you can stay a week in space, the rest of your life support is OK. All you need now is a second booster to launch you to the Moon. You can either circle it in a free return trajectory, or brake an enter orbit.

The orbit part seems the most complicated.

Re:Why Lisp Will Never Win (2003) by nknight on Sunday November 06, 2011 @06:30AM (#3202401)
Wow. Just wow.

Every single system you name is made up of a large number of complex sub-systems designed and built to exacting requirements, and any one of them can still break at any time. Thinking you "got it right" just because it didn't break one time in one set of circumstances is unbelievably naïve. There's a reason all man-rated launch systems have redundant backup systems, and even the fallbacks don't always work right.

By the way, let me know when you get micrometeoroid shielding "right" -- every space agency on Earth will want to talk to you. You might even win a Nobel Prize.

You're also completely ignoring navigation. Just because you can fire up a rocket doesn't mean shit if you're at the wrong point in your orbit, or pointed in the wrong direction.

You're also failing to distinguish between "I can get THIS rocket into orbit" and "I can get a rocket big enough to carry people and a useful payload into orbit and/or to the moon". Think you can just "scale up"? You're wrong. Effects that were too small to measure before will suddenly kill you. Materials that could handle the strain of the smaller rocket will crumble under the increased forces.

Oh, by the way, every time the weight or configuration of your spacecraft changes, you're gonna have to rework your navigation assumptions.

Might I suggest playing with Orbiter for a while? Once you've done that, spend a few minutes thinking about the fact that you've just been overwhelmed by a vastly simplified simulation in which nothing ever goes wrong. Now picture the real world.

Oh, by the way, your heat shield is cracked, but nobody noticed because your budget didn't allow for more than one guy to inspect it, and he was too drunk. Have fun getting home.

Re:Why Lisp Will Never Win (2003) by rbanffy on Sunday November 06, 2011 @06:58AM (#3202437)
The original point is that, once you achieve LEO, a free-return tragectory to the moon is just a question of delta-v. And not that much - the Soyuz-based moon roundtrip being offered for US$ 100M is a Soyuz docked with a mini-bus-sized booster stage.

Micrometeorites? Did Apollo have that? I am not implying it's easy - I'm implying LEO is the hardest part when it comes to a moon-orbit trip.

And in the case of an unmanned craft, it's easier: you don't have to bother with life support, atmospheric braking and most of the radiation shielding.

Re:Why Lisp Will Never Win (2003) by nknight on Sunday November 06, 2011 @08:02AM (#3202551)
Except that wasn't the original point at all. It started with Jd saying:

"If you want to go to the moon you need a f'ing spaceship and not emacs and an open source notebook. Spaceships are big. They take a lot of people to run, and they have to be smart people. One mistake and you are smithereens at Cape Canaveral."

Followed by sedachv saying some guys who are trying to launch humans on suborbital flights would disagree.

Nothing you've said explains what suborbital flights have to do with going to the moon. On the contrary, you seem to be admitting they have nothing to do with going to the moon, so I still have absolutely no idea what sedachv's point was.

Re:Why Lisp Will Never Win (2003) by rbanffy on Sunday November 06, 2011 @10:25AM (#3202853)
Well... I maintain that the hardest part of going to the Moon is achieving Earth orbit. The rest is, comparatively, easy.

But, then, this discussion is completely unrelated to the point regarding emacs and an open source notebook (whatever that is). In general, Jd's rant felt confusing. I am not sure how Lisp and spacecraft are related, BTW. He (or she) implied somehow that you couldn't go to space with something simple and elegant like Lisp, but I can't be sure.

Re:Why Lisp Will Never Win (2003) by Jd on Sunday November 06, 2011 @04:52PM (#3203981)
Re: open source notebook see: http://stallman.org/stallman-computing.html

I'll make the relation very clear. There is a segment of the Lisp community, which most certainly includes Paul Graham, that maintains that projects with large team sizes are bad. There is a certain logic here, as it is hard to maintain a high average intelligence distributed across a large number of people. Most true geniuses were loners. Regardless, projects of a certain scope (i.e. building a space shuttle) require large teams. You simply can't do it with two guys in a basement.

In a certain sense the computer language employed is irrelevant, except that the lisp community has never, largely because of the facts that should be obvious but are apparently not to the introverted "genius" types they pray on, produced good and extensive enough libraries to be viable for most real world tasks. In case you didn't notice, for example, it took what, 4 years for Hacker News to get HTTPS support.

I don't want to rib anyone too badly, because I am a strong supporter of what YCombinator is doing in general and in the incorporation of lambdas and closures in newer languages, but I think generally speaking it is clear that the Lisp culture is not to cooperate but to position one's self as some sort of genius and thus to create one's own dialect, etc.

I'll give you an example that will probably irritate some people, but is relevant here. Microsoft Word is, in my opinion, one of the greatest architectural marvels of modern civilization. The feature set of that single app is truly extraordinary, and there are all sorts of wonky ways that you can do lots of tasks that makes millions (and potentially even billions) of people's lives better every day. I understand that there is a point to OpenOffice and "fighting the man" and whatever since Microsoft certainly crushed some competitors in mean ways, but realistically there is nothing in the world that competes.

Now you might say, oh sure but I can just use vim and use latex for formatting (or even google docs). And in a certain sense you might be right too, but sooner or later you will find a feature that is not accessible to you (I recently noticed that change tracking still absolutely blows in Google Docs compared to Word).

Why do I bring this up? Well, even if you are right that "the hardest part of going to the Moon is achieving Earth orbit," you still have to have all of those other things. You need life support or you dead; this is non-negotiable, just in the way that mail merge is not negotiable for millions of people at millions of office worldwide. And if you want life support (or mail merge), you need to work with the people who design these other systems.

By the way, I'm not saying you need to like any of this. I'm simply saying that if you want to build something great, you have to at times slog through the mud. My experience with Lispers is they are frequently afraid to get their hands dirty. And that doesn't cut it except in a pretty-ditty fantasy castle.

Re:Why Lisp Will Never Win (2003) by rbanffy on Monday November 07, 2011 @08:50AM (#3206203)
> Regardless, projects of a certain scope (i.e. building a space shuttle) require large teams.

The shuttles are not engineering successes. Most large projects (large as in "lots of brains required to think about how they are going to work") usually aren't either. I'll side with pg on this. Humans cannot organize that well and less ambitious steps are a better approach to solve really big problems.

Languages are tools you use to express and think about problems. There are problems that are more amenable to be expressed in terms of Lisp and others that are more suited for C. Or SQL, or Forth.

It may be hard to express mud in Lisp, but I don't think the whole thing is as bad as you do.

But I have to wonder... Does anyone still use mail merge?

Re:Why Lisp Will Never Win (2003) by Jd on Monday November 07, 2011 @05:27PM (#3208447)
What are the shuttles if not engineering successes? PR stunts? How would you categorize them? There are certainly problems (like providing a good and functional high-speed rail system) that are not really questions of "genius" and more practical questions of how to apply already developed technology, which, given the politics of our day, might require a different type of "genius" to approach.

It is interesting that you describe language as a tool to think about a problem. Is it really so? Is it only because you have the language that you think about the problem in terms of the language?

Humans cannot organize that well? Well, I guess Cape Canaveral in 1969 was smoke and mirrors -- or what do you mean? I think you probably mean that the cost of organizing at a large scale like that is not a cost you are personally (or, collectively in terms of the culture of elites in the West) willing to bear. I would even suggest that it very well could be in actuality an unwillingness to bear any sacrifice for the sake of a collective goal.

"Less ambitious steps" for "really big problems." Sounds great. What steps do you mean and what problems are you possibly hoping to solve? How the private sector might be able to get someone in space -- an "engineering success" that was already accomplished approximately 50 years ago, now potentially finally accomplished by the private sector, so long as individual Russian tycoons are able to pay for $100M tickets with money stolen from starving peasants.

A brave new world indeed. As for mail merge, I have very little interest in the common man. I live in Silicon valley and watch the world burn from a carefully-cultivated distance.

Re:Why Lisp Will Never Win (2003) by brlewis on Saturday November 05, 2011 @09:55PM (#3201548)
When you say that you would never hang out with "the Lisp community" because you want to build great products, I take that as an implication that the language, the community, or both, are incompatible with building great products.

That attitude, to me, is a large turn-off.

I disagree with your notion that Lisp code has to be re-written into another language to address the minutia. I address minutia in Scheme a lot.