6 ms·

It's been so sad to see that while the AI winter ended, the Lisp winter that started around the same time never did. Maybe it's coming back after consolidation

It's been so sad to see that while the AI winter ended, the Lisp winter that started around the same time never did. Maybe it's coming back after consolidation on the languages end: SBCL and Clojure are coming ahead as clear winners, and are quite powerful. I'm sure I'm not the only guy who started learning Lisp after reading Paul Graham's essays, and I'm going to keep using Lisp for several reasons, but the momentum just isn't there.

mbrock 8y ago
Lisp certainly wasn't in vogue when Paul Graham did his startup and wrote those essays. "Robert and I both knew Lisp well, and we couldn't see any reason not to trust our instincts and go with Lisp. We knew that everyone else was writing their software in C++ or Perl. But we also knew that that didn't mean anything. If you chose technology that way, you'd be running Windows. When you choose technology, you have to ignore what other people are doing, and consider only what will work the best." "This is especially true in a startup. In a big company, you can do what all the other big companies are doing. But a startup can't do what all the other startups do. I don't think a lot of people realize this, even in startups." [...] "If other companies didn't want to use Lisp, so much the better. It might give us a technological edge, and we needed all the help we could get." http://www.paulgraham.com/avg.html http://www.paulgraham.com/avg.html
- tim333 8y ago
  >But a startup can't do what all the other startups do Seems like a lot of startups do do what the others do tech wise and use Ruby which seems to work out ok.
  - vidarh 8y ago
    It's more that a startup can't do what all the others do in every aspect. You need to do something differently (different niche) or better (if there are others in your niche).
s-shellfish 8y ago
AI winter is having a problem because not every problem can be solved with a number calculation. Not everything can be measured with a metric, and this especially runs into problems when measurements are being used to measure one another - becomes very convoluted. I have optimism for lisp.
- lgessler 8y ago
  What do you mean by "number calculation"? In at least one sense, any kind of computation at all is a "number calculation".
  - s-shellfish 8y ago
    Statistics, probability
armitron 8y ago
The momentum will never be there, Lisp is not geared for public consumption period. The commodification of programming and proliferation of languages like Javascript and Python means not only that the "masses" lack fundamental understanding of Computer Science but that they have been conditioned to go after superficial ease-of-use rather than actual value. Anything that requires upfront effort to reveal its secrets is rejected for what is fundamentally broken but covered in a veneer of user friendliness. Python is the new Perl (Go is already gearing up to replace it where reliability is at stake) and in a couple of years the masses will look at it as a colossal and obvious failure, but they will make the same mistakes again and again when the next shiny thing comes along because the entire domain is rigged in multiple ways. I suggest reading Richard Gabriel's essays, they are very depressing but also illuminating. In many ways he predicted what is taking place now, the rise of commodification and the obliteration of the field by companies like Google. https://www.dreamsongs.com/Essays.html https://www.dreamsongs.com/Essays.html
- sedachv 8y ago
  > Python is the new Perl (Go is already gearing up to replace it where reliability is at stake) and in a couple of years the masses will look at it as a colossal and obvious failure, but they will make the same mistakes again and again when the next shiny thing comes along because the entire domain is rigged in multiple ways. Before Perl, Tcl was supposed to be the Lisp replacement. As far as I know this "language X is an adequate/better replacement for Lisp" argument goes back to Ted Nelson ranting against Lisp while advocating TRAC in the 1974 book Computer Lib. So far all of the language Xs have disappeared, and the only thing to replace Lisp has been other Lisps.
  - mehh 8y ago
    Hang on I thought Scala was meant to be the new Perl ;)
Tarq0n 8y ago
The R language that gets used a lot for machine learning is actually very Lisp-like. So maybe that's introducing a lot of people to Lisp without them realizing it.
- workinfunk 8y ago
  Do you mind expanding? I don't know R, but on looking at several examples, I don't see it at all.
workinfunk 8y ago
Is SBCL winning in anything? One of my acquaintances, a Lisp zealot, convinced me to jump in and make my next project in SBCL. Short story, it was a joke. Nobody seems to have done any serious work in SBCL in over a decade. Basic things are either completely missing or broken. Clojure I can at least make working software in. The fact that you are putting SBCL up there with Clojure makes me wonder if you've ever used SBCL.
- olzd 8y ago
  Latest release is June 28, 2018. SBCL REPL kinda sucks by itself but nobody uses it that way.
- blue1 8y ago
  what's wrong with sbcl?
- jlg23 8y ago
  Not sured what you tried to do, but over the last decade I've done a lot of commercial lisp programming and wonder what is "completely missing or broken". EDIT: always in SBCL and CCL
- zodmaner 8y ago
  >Is SBCL winning in anything? It is one of the most active implementation of Common Lisp (other being Clozure CL) and one of the fastest thanks to its support of optional type hinting (which it had long, long before gradual typing become in vogue recently). With some type hinting, it is possible to generate a code that's almost as fast as C (in certain cases, of course, but still much better than other implementations). >Nobody seems to have done any serious work in SBCL in over a decade. Could you please elaborate on what you mean by this? SBCL releases a new version that contains both enhancements and bugfixes every couple of months, with the latest version (version 1.4.9, released on June 28, 2018, ~5 days ago) came one just one month after the previous version, 1.4.8, so I really don't understand what you mean when you said that no one "seems to have done any serious work in SBCL in over a decade". You can also see all previous releases and their changlog here: http://www.sbcl.org/all-news.html http://www.sbcl.org/all-news.html. Or did you meant to say that no one use SBCL to do serious work? Well, with some googling, I'm sure you'll be able to find that, while not many, there are companies that use SBCL (and other Common Lisp implementations) in production and other non-trivial works (not to mention various freelancers who use SBCL to put foods on the table). Some of those companies are also involves with the development of SBCL as well. >Basic things are either completely missing or broken. Care to list them? One of the strength of SBCL is that the devs are very responsive, especially to bug reports, so I'm sure they would be more than happy to fix the problems that you encountered.

mbrock 8y ago

Lisp certainly wasn't in vogue when Paul Graham did his startup and wrote those essays. "Robert and I both knew Lisp well, and we couldn't see any reason not to trust our instincts and go with Lisp. We knew that everyone else was writing their software in C++ or Perl. But we also knew that that didn't mean anything. If you chose technology that way, you'd be running Windows. When you choose technology, you have to ignore what other people are doing, and consider only what will work the best." "This is especially true in a startup. In a big company, you can do what all the other big companies are doing. But a startup can't do what all the other startups do. I don't think a lot of people realize this, even in startups." [...] "If other companies didn't want to use Lisp, so much the better. It might give us a technological edge, and we needed all the help we could get." http://www.paulgraham.com/avg.html http://www.paulgraham.com/avg.html

tim333 8y ago
>But a startup can't do what all the other startups do Seems like a lot of startups do do what the others do tech wise and use Ruby which seems to work out ok.
- vidarh 8y ago
  It's more that a startup can't do what all the others do in every aspect. You need to do something differently (different niche) or better (if there are others in your niche).

s-shellfish 8y ago

AI winter is having a problem because not every problem can be solved with a number calculation. Not everything can be measured with a metric, and this especially runs into problems when measurements are being used to measure one another - becomes very convoluted. I have optimism for lisp.

lgessler 8y ago
What do you mean by "number calculation"? In at least one sense, any kind of computation at all is a "number calculation".
- s-shellfish 8y ago
  Statistics, probability

armitron 8y ago

The momentum will never be there, Lisp is not geared for public consumption period. The commodification of programming and proliferation of languages like Javascript and Python means not only that the "masses" lack fundamental understanding of Computer Science but that they have been conditioned to go after superficial ease-of-use rather than actual value. Anything that requires upfront effort to reveal its secrets is rejected for what is fundamentally broken but covered in a veneer of user friendliness. Python is the new Perl (Go is already gearing up to replace it where reliability is at stake) and in a couple of years the masses will look at it as a colossal and obvious failure, but they will make the same mistakes again and again when the next shiny thing comes along because the entire domain is rigged in multiple ways. I suggest reading Richard Gabriel's essays, they are very depressing but also illuminating. In many ways he predicted what is taking place now, the rise of commodification and the obliteration of the field by companies like Google. https://www.dreamsongs.com/Essays.html https://www.dreamsongs.com/Essays.html

sedachv 8y ago
> Python is the new Perl (Go is already gearing up to replace it where reliability is at stake) and in a couple of years the masses will look at it as a colossal and obvious failure, but they will make the same mistakes again and again when the next shiny thing comes along because the entire domain is rigged in multiple ways. Before Perl, Tcl was supposed to be the Lisp replacement. As far as I know this "language X is an adequate/better replacement for Lisp" argument goes back to Ted Nelson ranting against Lisp while advocating TRAC in the 1974 book Computer Lib. So far all of the language Xs have disappeared, and the only thing to replace Lisp has been other Lisps.
- mehh 8y ago
  Hang on I thought Scala was meant to be the new Perl ;)

Tarq0n 8y ago

The R language that gets used a lot for machine learning is actually very Lisp-like. So maybe that's introducing a lot of people to Lisp without them realizing it.

workinfunk 8y ago
Do you mind expanding? I don't know R, but on looking at several examples, I don't see it at all.

workinfunk 8y ago

Is SBCL winning in anything? One of my acquaintances, a Lisp zealot, convinced me to jump in and make my next project in SBCL. Short story, it was a joke. Nobody seems to have done any serious work in SBCL in over a decade. Basic things are either completely missing or broken. Clojure I can at least make working software in. The fact that you are putting SBCL up there with Clojure makes me wonder if you've ever used SBCL.

olzd 8y ago
Latest release is June 28, 2018. SBCL REPL kinda sucks by itself but nobody uses it that way.
blue1 8y ago
what's wrong with sbcl?
jlg23 8y ago
Not sured what you tried to do, but over the last decade I've done a lot of commercial lisp programming and wonder what is "completely missing or broken". EDIT: always in SBCL and CCL
zodmaner 8y ago
>Is SBCL winning in anything? It is one of the most active implementation of Common Lisp (other being Clozure CL) and one of the fastest thanks to its support of optional type hinting (which it had long, long before gradual typing become in vogue recently). With some type hinting, it is possible to generate a code that's almost as fast as C (in certain cases, of course, but still much better than other implementations). >Nobody seems to have done any serious work in SBCL in over a decade. Could you please elaborate on what you mean by this? SBCL releases a new version that contains both enhancements and bugfixes every couple of months, with the latest version (version 1.4.9, released on June 28, 2018, ~5 days ago) came one just one month after the previous version, 1.4.8, so I really don't understand what you mean when you said that no one "seems to have done any serious work in SBCL in over a decade". You can also see all previous releases and their changlog here: http://www.sbcl.org/all-news.html http://www.sbcl.org/all-news.html. Or did you meant to say that no one use SBCL to do serious work? Well, with some googling, I'm sure you'll be able to find that, while not many, there are companies that use SBCL (and other Common Lisp implementations) in production and other non-trivial works (not to mention various freelancers who use SBCL to put foods on the table). Some of those companies are also involves with the development of SBCL as well. >Basic things are either completely missing or broken. Care to list them? One of the strength of SBCL is that the devs are very responsive, especially to bug reports, so I'm sure they would be more than happy to fix the problems that you encountered.