freenode/#lisp - IRC Chatlog
Search
18:26:00
pjb
The thing is that since the specification is not formal, and is known to contain bugs, some level of consensual interpretation is left to the implementations. But it is better if all implementors agree on thos litigious points. Unfortunately, on some important parts, they don't. Eg. they could agree on how to do pathnames (both physical and logical) on POSIX systems, but they don't. :-(
18:30:21
stacksmith
pjb: oh are you saying that they indeed do return 2? I read it as 'I wish they would all return 2' ...
18:31:30
stacksmith
loli: I know. I was asking pjb to clarify a statement earlier that I misunderstood.
19:44:48
puchacz
hi; slime on slow device please. I am getting [Interrupted: time limit exhausted] on fuzzy completion
20:15:54
jackdaniel
I've wrote some rendering approximation illustrations with McCLIM: http://hellsgate.pl/files/c885c815 (I'll make it one of clim-examples after adding some interactivity to it)
20:36:05
flip214
puchacz: well, why not do the auto-completion against a faster system, and just run the code on the slow target system anyway?
23:35:14
malice
The problem is, it somehow does not work on the emacs level; indentation bugs and C-c C-c invoked to compile it fails somehow
23:35:30
malice
it looks as it treated the comment as the code, even though string does not end prematurely
23:38:21
pjb
malice: when you have forms or strings that are too big, font-locking can have difficulties processing them.
2:41:14
nydel
i don't think i've ever seriously used cmucl. i might be interested to try out CLX its built-in x11 gui interface, anyone used it to positive or negative experience?
2:49:28
Bike
and programming with clx is equivalent to programming with xlib, that is to say, extraordinarily unpleasant
4:43:26
beach
Lycurgus: So now that you know about all the activities toward a Lisp OS, have you decided to help out with some project?
4:48:44
Lycurgus
the lisp image portion would be the "upper part" of MCP married to a SSI linux kernel
5:02:53
krwq
beach: but you got something working, and I think it still might be less work to fix it than write a new one from the scratch
5:04:17
beach
The underlying Common Lisp implementation I want must have first-class global environments, for example.
5:09:32
krwq
beach: I'm not sure what do you mean by that but will try reading your article. have you started implementing that yet?
5:09:59
minion
krwq: SICL: SICL is a (perhaps futile) attempt to re-implement Common Lisp from scratch, hopefully using improved programming and bootstrapping techniques. See https://github.com/robert-strandh/SICL
5:10:05
minion
krwq: Cleavir: A project to create an implementation-independent compilation framework for Common Lisp. Currently Cleavir is part of SICL, but that might change in the future
5:10:36
beach
krwq: I am working top-down and I try to make my stuff modular and implementation-independent as much as possible.
5:11:56
krwq
beach: so your lisp OS idea has the different idea of how is data organized (i.e. no hierarchical paths) - that seems to be incompatible with CL - are you planning to update CL to like CL2?
5:12:56
krwq
beach: i haven't read the whole (http://www.shirky.com/writings/ontology_overrated.html) - too much data to digest with addition to your first-class global environments
5:16:53
beach
But, for the OS, it is not terribly important to me if things like pathnames are supported.
5:17:39
krwq
beach: it would be nice if you could translate that idea to paths as well and make both supported - filesystem can interpret slash as a separator for a list of categories
5:19:19
beach
Lycurgus: I'll say this again in case it hasn't been clear: I don't care if I will end up being the only user of what I create.
5:20:45
krwq
Lycurgus: I think it's better to share ideas like that - IMO this sounds really cool - I've spent a lot of time already trying to figure out how to organize my hierarchical paths
5:21:22
loke
Even if Beach doesn't implement pathnames it seems to me that such a concept can easy be built on top of the primitves he _do_ create.
5:21:40
Lycurgus
so sharing collides with the reality of the capital system, in which life actually occurs
5:21:53
krwq
but having said that if we apply this for paths we could do that as well for other things i.e. bindings - lexical bindings are also hierarchical
5:22:03
loke
I think they issue here is that pathnames themselves are not the idea fundamental mechanism to organise information.
5:23:18
Lycurgus
an OS is a fundamental system resource that is judged on it's merits as such, that's what I meant by that
5:23:20
beach
Lycurgus: Just to be clear, what I am designing is not just another operating system similar to existing ones, except written in Common Lisp.
5:24:05
beach
Lycurgus: I am a researcher. I invent new ideas, test them, and document them in the form of papers. The main test is whether the ideas are good or not. Not whether it is used. Of course, it is even better if it turns out to be useful to people.
5:24:54
beach
But, there are so many psychological barriers in the domain of software that I strongly suspect very few people will be interested.
5:25:55
Lycurgus
beach, I see. I am not a researcher, I'm more interested in practical extension of current technical horizons
5:26:59
Lycurgus
so I use research results for that purpose but research for the sake of research doesn't appeal to me
5:27:17
krwq
beach: if at any time in the future (file-read ...) and (read-file ...) will be the same you bought one user
5:44:18
whoman
beach: what other kind of languages have you used a lot of? just curious. thats a lot of lisp =)
5:48:47
beach
whoman: A lot? Pascal, PL/I, C, Scheme. A little: Python, C++, Java, PostScript, Simula, APL, Metafont, and maybe some more.