| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11
Comments:
<0> Well, thanks all. Interesting discussion.
<1> starseeker: sorry, was/am away for dinner. more later. you're right, but I think they would have little other recourse besides shutting it down, BICBW.
<2> they wouldn't have much to gain by shutting it down - more people using common lisp means more software sales for lispworks
<3> yeah, until SCO buys lispworks
<4> they have to protect their rights or lose them, when they become aware of a violation
<4> at least trademarks work that way and I think copyrights have something similar, but IANAL
<3> dankna-lap: no. that's not true.
<2> unless they granted permission
<4> oh
<5> copyrights are not similar to trademarks
<4> I am pleased to be corrected, then
<3> depending upon something of questionable legality, even if the current owner doesn't seem to mind/notice isn't necessarily a good idea.
<5> indeed
<2> otoh if they think it will generate more sales they will readily grant permission
<2> and if they fail to grant permission, you can expect a good possibility of getting a cease and desist
<6> SCO is sniffing around lispworks? Goodie.
<3> this is the way rumors get started. =)
<2> no, SCO is too busy getting fried in court
<3> DrForr: yes, please submit the story to slashdot!
<3> breaking news! :)
<2> now that would be the ultimate in bad publicity
<2> common lisp sued for 6 billion dollars
<6> Heh, obviously my sarcasm detector needs to be retuned. Given all of the stupid moves SCO's made buying out LispWorks seemed perfectly rational, if not curtains for the entire language.
<7> hello
<8> hello voidengineer
<9> good afternoon spiaggia
<7> spiaggia hello
<8> hello arbscht
<8> arbscht: I understand cl***es have started again, right?
<9> that's right, I'm on campus now
<9> and a little annoyed; they seem to have removed emacs at the labs
<8> oh?
<8> I think they must have reinstalled the machines and did not install Emacs again, nor Lisp probably. I'll talk to them.
<8> we are talking Linux, right?
<1> starseeker: still around? it would be interesting to explore what the worst-case scenario of using the ANSI spec would be.
<8> arbscht: no, wait. That can't be right. Linux without Emacs is unthinkable.
<1> not to mention unusable (by me at least)
<8> hello slyrus
<1> hey beach
<9> yep, it is linux :(
<9> they seem to have reorganised and upgraded a lot of things. I'm searching for traces of emacs in case they've simply moved it, but am having no luck
<0> slyrus: back
<0> slyrus: still here?
<1> hey starseeker. it would be interesting to consider what the downside would be. a C&D order about what? the entire work? parts? etc...
<1> what remedies might one be obliged to make?
<8> arbscht: let me know if you find it. I'll try to find out what is going on.
<9> spiaggia: will do, thanks
<0> slyrus: The worst case would be a suit for damages, with financial compensation on the part of the violator, I think
<0> plus killing the project and voiding all the work put into it
<1> yes, and I wonder what damages one could reasonably claim in this case
<0> dunno. With the draft freely downloadable for ten years, I imagine that would be a bit tricky...
<1> lispworks might be able to claim some damages, others would probably have a harder time
<0> the company doing the hyperspec, too
<0> Harlequin(sp?)
<0> Symbolics Inc.
<0> Allegro
<1> who owns harlequin and symbolics? yes, allegro would be interesting to approach. I could imagine a more freely avaiable spec being something in their best interests.
<0> Guy Steele I think wrote much of one of the books the spec was based off of
<0> harlequin - not sure.
<0> Symbolics is a strange case - their ***ets were bought and they're back online, but it's some sort of weird holding company.
<8> I think we could handle writing a document called (say) The Common Lisp Reference Manual.
<10> Hmm. The weak hash table tests no longer crash, which is good. But the weak pointers tend to be held on for longer on x86 than non-x86. Not sure why, but I'll blame the conservativeness.
<0> We went round and round about that back when we were thinking about the Macsyma source
<1> spiaggia: I like the title
<0> We could, but that wouldn't be a spec in the sense of a language spec, correct?
<8> starseeker: it would be as good (provided we take care to keep it correct)
<0> Well, then it would be a spec :-)
<0> My original idea was to have a new spec which was "very similar" to ANSI Common Lisp, but instead call it Community Lisp
<8> starseeker: sort of, yes, but it would also be a much more complete document in terms of examples and explanations
<0> spaiggia: Right, but to my way of thinking that's just a spec done Right ;-)
<0> We agree.
<8> yes, I think we do.
<8> my estimate is that such a thing would be around 1k pages
<0> Yes
<0> at least
<0> I would like to see a unit test framework incorporated from the beginning, sort of like Paul Dietz's test suite
<8> most of the work could be distributed, like writing the spec and examples of each function. One person (or a few people) would have to work on the general organization and conventions and such.
<0> Yes
<1> perhaps we can eliminate the need for a spec with a detailed enough test suite
<0> Hmm - that a VERY intersting idea. Anythhing which can run the test suite without error would be, by definition, compliant
<0> what an elegant way to test compliance
<1> the normative part would be the docstrings :)
<0> Well, I'd suggest using LaTeX for the text part, and including the code in noweb or similar style - docstrings are a bit limiting ;-)
<0> Is there some kind of "specification notation" for things like argument order, type limitations, etc?
<8> I suggest texinfo, actually
<0> That would work too.
<1> smarkup!
<9> spiaggia: still nothing, but I've made a post on the faculty forums which may yield something. I'll let you know how that goes
<8> arbscht: OK.
<0> slyrus: Probably the least informative page I've seen for a project ;-)
<1> thanks! :)
<0> There is also cl-typesetting
<0> Oh, is it yours? Sorry!
<1> it will be more informative after I finish the thesis
<1> no need. the lack of documentation is appalling, considering that it's purpose is writing documentation and other text!
<0> Is it based on the cl-typesetting engine or is it something all its own?
<1> s/it's/its/
<1> it spits out latex
<11> isn't docbook the fashionable/politically correct way nowadays, as opposed to texinfo
<0> Cool.
<1> and xhtml
<1> and bibtex
<8> rr-: does docbook have macros?
<1> this used to be extensions to gigamonkey's markup, but I removed all of that stuff
<1> and made it its own package
<1> so it used to be able to do cl-tyepsetting
<0> I would suggest cl-typesetting for obvious reasons, but it has almost as little documentation as smarkup :-(
<11> i ***ume that is a rhetorical question
<8> no, I don't know
<12> hmm wtf is this some kind of freenode thing? autoleave after timeout?
<0> maxm - is your client set up to do that maybe?
<8> slyrus: the dogfood argument is in favor of something like smarkup, but it would have to be able to do at least what Texinfo does, i.e. produce indexes, cross references, and such.
<12> well its ERC which I just started using
<12> why would it have something like that by default
<12> plus it does not happen with #emacs, only with #lisp
<0> maxm : dunno
<12> very strange
<8> maxm: I am using ERC right now and I have been connected for weeks
<1> spiaggia: yes, that's reasonable. for me things like TOC, lists of figures, etc... are all just p***ed off to the layer below (latex) and ignored in xhtml. not ideal for something like this.
<12> maybe I'm hitting some kind of leave key by mistake... But I swear I left work and it was here, and I came home, just reattached my screen and it says I left
<12> ok lets leave it and see what happens
<0> I would suggest that the probable best way to proceed would be to pick texinfo or something else that has already the mature features we need, and a little awk magic can later convert it to whatever we ultimately decide on
<8> slyrus: well, that doesn't work, because we might want to produce all of those things in an online version of the manual in the form of text
<8> sounds very reasonable to me
<1> spiagga: yes, I agree.
<0> Obviously if a lisp solution matures we can hope on, probably providing a nice big debugging document for whatever system we switch to later ;-)
<0> er s/hope/hop
<1> starseeker: or you can do what I'm doing which is using smarkup to generate latex (or texinfo or what-have-you) with the idea that working with ***p's is going to be easier for the folks writing the lisp spec
<8> BBL
<0> That's a point.
<0> Is there an example smarkup document somewhere?
<1> it doesn't solve the problems spiaggia mentioned, but doesn't make things worse
<1> yes. or maybe there will be in a minute :)
<0> Can't texinfo produce text? If you can export to that that should work...
<8> slyrus: well, there is a good chance that we would have to extend Texinfo anyway, so we should weigh that against adding features to something like smarkup
<1> sure, after losing all the markup :)
<0> slyrus: How hard would it be to add back in the cl-typesetting targeting?
<13> rydis: you around?
<1> spiaggia: yeah, but my point is that having ***ps that are used to generate the texinfo is better than native texinfo. it's a debatable point, at best :)
<1> hmm... I don't know. I never targeted cl-typesetting, but gigamonkey's stuff did and I was using that as my format for a while.
<1> I imagine it wouldn't be too bad
<0> I don't know how ready for prime time cl-typesetting is (it certainly isn't in terms of math) but we could at least have an "in house" environment
<1> no worse than what gigamonkey had to do for his markup, which didn't seem too horrible
<1> I'm not sure what that buys us for online stuff though. I would prefer to target xhtml in lisp and latex for dead trees, personally
<0> Hmm... true...
<0> OK, let's start there.
<0> I downloaded smarkup - which lisp do you recommend?
Return to
#lisp
or
Go to some related
logs:
#debian
HTML::TreeBuilder Parsing of undecoded UTF-8 will give garbage when decoding ent
gentoo ditscc
nick0garvey
bcm4306 g4 dapper
debian multiple cpus
qcow vdi
#php
lamemp3 debian
PKCS5_PBKDF2_HMAC_SHA1 openssl
|
|