freenode/#sicl - IRC Chatlog
Search
22:47:17
|3b|
alandipert: in 2.3, "the limitations of the pre-reader appear not to encumbered those" -> "not have encumbered" or "not encumber"... not sure which is better there (or maybe rewriting more of the sentence would be better?)
23:01:47
|3b|
alandipert: the async reader example in 3.1 might be more interesting if it wrote the number in more than 1 piece
23:23:37
alandipert
maybe i should rationalize that. i wrote them that way because nodes have arbitrary keys, but the methods/behaviors of Env are sealed
23:25:48
|3b|
possibly just put "node" in normal font and capitalize it in "node and Env objects...", if it isn't really a js type?
23:28:44
|3b|
js implementation details like actual class or not aren't really important to the paper
23:29:39
|3b|
but having lower case "node" at beginning of a sentence makes it stand out, which is a bit distracting
23:30:45
|3b|
Switching to "Env and node..." would also avoid that, if it does refer to something specific in js code worth having a different font (though that's only time it gets different font in the section that i could find)
4:35:07
alandipert
in even better news, my also-admittedly-small but highly educated family has agreed to edit. red pens have been produced
4:35:10
beach
Great! I feel privileged to be surrounded by so many smart and knowledgeable people, who, in addition, are willing to help out so much.
4:36:27
beach
You know, mine worked for ILOG that produced LeLisp, and she translated Lisp in Small Pieces from French to English. She will sign your copy if you bring it to ELS. :)
4:36:51
beach
oh, and she has a PhD in English, and she was a distinguished technical writer at ILOG and then IBM. :)
4:39:14
alandipert
that's awesome. yeah like many programmers i make casual comparisons between programming and writing, but every time i try to do it, i realize: writing is a higher art
4:42:24
beach
It is the first style guide that is not about silly personal opinions, but based on his research in linguistics and cognitive psychology.
4:43:23
beach
Plus, he is an excellent writer himself, and an excellent speaker, as you can see from his many talks available on YouTube.
4:46:30
alandipert
recently i watched a short series of videos featuring Patrick Winston on lecturing and applied them in a talk. good stuff i think
4:53:50
beach
Not necessarily, that's a different issue. The unified memory is about semantics, not about speed.
4:55:40
beach
The unified memory is about pretending that all you have is RAM. No files, no file system.
4:56:29
beach
But Multics did that already, in the 1970s. Unix destroyed that idea, and the minds of generations of software developers.
4:57:19
beach
Combined with the fact that young developers do not seem to want to (or have the opportunity to) study computing history, we know use systems that are inferior to what we had 50 years ago.
5:01:31
beach
ebrasca: I seem to fail to express my arguments, so I guess I'll stop for now. Let me know when you are willing to spend some more and effort understanding my arguments.
5:02:26
alandipert
beach do you have a theory about why unix came to dominate in spite of the superiority of multics that you perceive?
5:03:52
beach
alandipert: It's fairly simple. Bell labs gave away the source code for free so that universities could install it on small computers and let students study it and hack on it.
5:04:53
beach
ebrasca: Modern software is possible because of hardware performance. But modern software also takes a lot more effort to write, especially if you want it to be safe, than it would have with better OS technology.
5:06:02
beach
ebrasca: Original Unix was pretty crappy compared to the predecessors. I suggest you read the "Unix haters handbook". It is written by very smart and very knowledgeable people.
5:06:12
alandipert
beach presumably the students went into industry and, understanding unix to be synonymous with computing, created demand for commercial unix products?
5:07:00
beach
ebrasca: It took a very long time for Unix/Linux to catch up. A simple thing like shared libraries took forever to incorporate, and the way it is done is pretty bad. Especially for performance.
5:10:57
alandipert
beach my suspicion around that argument is based on how competitive the industry is. it seems like the ignorance would have to be unreasonably vast in order for a superior technology not to succeed
5:10:58
beach
Learn about Multics PL/I so that you know that it was an excellent programming language, and it inspired the condition system of Common Lisp.
5:11:44
beach
alandipert: I can guarantee that the ignorance of industry is incredible, and it is very scary indeed.
5:12:31
beach
Not only did I work for industry for a while, but I have consulted for several companies, and I continue to give talks in industry.
5:13:32
beach
ebrasca: Sort of. It was the good idea when they started. But then, commodity hardware caught up and their model was no longer competitive. It was hard for them to predict that.
5:14:10
beach
ebrasca: But a contributing factor was the "AI winter" which was largely blamed on Lisp.
5:14:55
beach
ebrasca: You really really need to study computing history. You will understand much more about why things are the way they are.
5:15:17
beach
alandipert: So you don't work for industry? I mean, if you did, you would know first hand, right?
5:16:28
alandipert
beach i have many friends become wildly successful with tech and techniques i considered stupid, so now i try to keep an open mind
5:17:23
alandipert
well, i think that's maybe a philosophical line of argument that bottoms out at personal values
5:18:09
beach
There is not always, perhaps not even often, any particular reason why things turn out the way they do. So there is no great point in speculating why this or that thing failed or succeeded, and certainly no reason to connect success or failure to quality.
5:50:55
beach
We often get the question "If Common Lisp is so good, how come it is not more popular". Underlying that question are several assumptions that are false. One is that if something is good, it automatically becomes popular. Another is that people are rational.
5:52:53
beach
Industry is no better. When they say "All our programmers already know <language-x>, so that's what we will choose for our project", they implicitly suppose that 1. all languages are roughly equivalent, no matter the task at hand, and 2. training our programmers will cost so much that it will be much more expensive than the advantage represented by the gain in productivity.
5:54:54
beach
And when they say "We need all the speed we can get, so we choose <language-x>", they are implicitly saying that they are willing to spend an unbounded amount of money for even an epsilon increase in performance.
5:55:33
beach
In other words, industry is not even capable of doing an elementary risk analysis involving elementary mathematics.
5:58:45
beach
ebrasca: I am not going to retype all the computing history I think you ought to know here in #sicl. You really need to take the time to study it offline. You will understand much more about what you are missing, and how to fix the current dismal situation.
6:02:16
beach
ebrasca: And ask yourself the following questions: Why do I have to stop execution of an application such as Firefox in order to install a new version of it? And why do I need to stop execution of all my applications AND the operating system in order to install a new version of the operating system?
6:04:16
beach
Do you know what incredible kludges that are needed to make such a thing possible with an operating system like Linux?