freenode/#lisp - IRC Chatlog
Search
4:23:43
SaganMan
Xach: I bought the Winston and Horn 3rd edition book cheaply from amazon. The book cost is 2$ but shipping to my country is 3$.
4:23:43
minion
SaganMan, memo from pjb: Books are heavy, that's why we invented electronic books. Have it scanned!
4:39:09
pjb
lisp was discovered in 1958, first implemented in 1959. C was designed in 1969, evolved from B and BPCL.
4:40:09
aeth
Languages are living. Early lisp is ALL-CAPS and full of goto. Early C looks totally different, too.
4:40:45
aeth
Lisp predates C but the Lisp you'd read in any random program has a very late-1980s feel to it, if not later.
4:42:55
pjb
There was GO (used in PROG) in LISP 1.5, but most functions were written in quite a functional style.
4:43:35
aeth
But I guess my point is that modern Lisp is heavily influenced by Scheme (1970) and Smalltalk (1972)
4:45:10
aeth
pjb: What's the one that's full of GO? I remember reading one, and you seem to know all the popular really ancient Lisp programs
4:46:30
pjb
I don't know. I was about to say that it would be interesting to collect old programs, and do some statistics…
4:48:38
aeth
I think it used prog. http://www.lispworks.com/documentation/HyperSpec/Body/m_prog_.htm
5:02:16
pjb
SaganMan: yes and no. It's not written directly by human programmers, but it's expanded by macros.
5:28:22
aeth
I prefer physical books because it's easier to focus reading a physical book than a book on a distraction device.
5:32:42
whoman
kindle apparently has nice screens for reading. i turn brightness down on mobiles which probably helps in one way but hurts in another
9:49:32
SaganMan
I'm reading this article https://tpapp.github.io/post/common-lisp-to-julia/ where the author says that standard make-array doesn't give you an array of double-float.
9:51:10
Shinmera
There's only a very limited set of types that the implementation needs to support as explicit element-types. Everything else might be upgraded to T or something else.
9:52:23
Shinmera
Right. The linked page says that the only types that need to be explicitly supported as element-types are bit, base-char, and character.
9:53:38
Shinmera
Which is bad for performance, since it means you need to either manually annotate every value you retrieve from the array, or suffer the consequences of run time type dispatch.
9:54:24
Shinmera
It also means the implementation is not going to pack the value into the array and instead store a pointer to a box, meaning additional unpacking cost on dereference.
9:55:25
Shinmera
the point he makes with it is kind of moot because if you really want performance, choosing an implementation that isn't going to be performant is bogus.
10:01:51
SaganMan
it would be waste of effort for him to switch to another language now after he's done so much with cl
10:02:43
SaganMan
and with so many years of using cl, that would also hinder one's style of programming, isn't it?
10:04:48
SaganMan
I mean when you write code in new languages, it follows the style of your old language. In lisp, we use extensive macros and functions
10:06:17
varjag
i think they just wanted to play around in julia and found some lame excuse to switch :)
10:06:29
SaganMan
really? my python code is like my C code. Though there a significant changes but I feel like I'm not using language libraries to it's fullest. I mean like that.
10:09:56
Shinmera
It's a fair point to make if you want to write portable code that's fast. But I don't think it's a fair comparison to strike between a group of implementations and a single implementation.
10:11:01
Shinmera
Anyway, Julia can do some neat things with packing memory and compile-time optimisation because it is much more rigid than CL in that approach. (static, inferred types is the new trend it seems)
10:13:12
Shinmera
I don't think Julia is going to convince the Lisp crowd, but I would welcome it instead of Matlab, which I hate with the power of a thousand exploding suns.
10:14:18
Shinmera
Unfortunately I also doubt its ecosystem and maturity are going to rival or excel Matlab any time soon.
12:08:59
Shinmera
sjl_: Please join me in #shirakumo for future libmixed/harmony discussions. https://irclog.tymoon.eu/freenode/%23shirakumo?from=1509019350#1509019350
12:11:56
_death
Shinmera: a few weeks ago I wrote lisp like the way someone wrote java :).. some time before that, I wrote lisp like the way someone C :).. direct ports are ugly
12:13:33
Shinmera
ACTION grinds his teeth about some of the C-isms that unfortunately slip through between the cracks in some of his bindings libraries.
12:15:36
_death
Shinmera: yes, I also went with this approach for some atari BASIC programs.. but it's incomplete