@# Quotes DB     useful, funny, interesting





Google
 
Web www.quotesdb.info
Undernet  |  EFnet  |  Quakenet  |  Freenode  |  Dalnet  |  Ircnet  |  Galaxynet
Page: 1 2 3 4 5 6



Comments:

<0> at low resolutions the hinting of bitstream vera is awesome. hail to freetype.
<1> at any decent solution, hinting/subpixel rendering is crap
<0> resolution?
<1> er yes, resolution
<2> for what it's worth, I'm starting a blog that I intend will become involved with discussion coincident to CL programming: http://tioga-rd.blogspot.com . I'll probably wind up trying to take CORBA as something of a subject, there, hopefully soon - got some errands to take care of, today.
<0> subpixcel rendering is the best thing ever seen in font development in the last 5 years.
<2> fwiw, Tioga is a mountain p***, on the east end of yosemite valley, facing onto the great basin. blogspot wouldn't let me get tioga.blogspot.com, so I took the name of the road through the p***
<3> therp: I still violently prefer bitmapped fonts for hacking.
<1> indeed
<3> I really don't want to stare at huge amounts of fuzzy text, but I haven't got a flat screen.
<1> therp: eh? I hear people say it works on LCDs, but to me, it just makes a yellowish halo around blurry text on my LCD
<0> rydis: I take the burdon of running xemacs to get xft support. I find everythign too ugly to look at. bitstream vera sans mono is perfect for hacking.
<2> gibson research - fwiw, that's the name of the agency with a felllow running it who has made some discourse about subpixel rendering
<2> wasn't able to recall it at first
<0> mbishop: maybe your rgb ordering is wrong, but I doubt that. try to choose a different hinting style. I like medium.
<2> vz http://www.grc.com/ctwhat.htm



<1> http://modeemi.fi/~tuomov/b/archives/2006/03/17/T20_15_31/
<0> mbishop: the gnome panel gives a nice selector.. you can also you for ~/.fonts.conf directly
<3> therp: I can't really use decently priced LCD's, since they have far too low resolution. I get annoyed at seeing all the blockiness.
<0> rydis: that's a good point. do you know any vendors that sell high dpi LCDs?
<0> rydis: I only find high DPI panels at laptops
<3> therp: Not at reasonable prices.
<0> rydis: at unreasonable prices?
<1> therp: I use ratpoison, and what was that about .fonts.conf? I don't think that was english :P
<3> therp: Then look for manufacturers of stuff for medical imaging. :)
<0> mbishop: I use ion. I even patched ion myself to get xft. Otherwise I wouldn't be able to use ion.
<0> rydis: can you give me a vendor name?
<3> therp: Not off the top of my head, no.
<4> mbishop: sub-pixel anti-aliasing doesn't blur anything
<0> spiaggia: I'd say it depends on the setting :)
<5> ... Not the sub-pixel aa discussion -again-... :-/
<4> therp: if done wrong, it might, yes, if that's what you mean
<6> the subpixel aliasing under linux is vastly inferior to that on osx
<0> nyef: better than no discussion
<3> therp: www.planar.com seems to have some nice screens.
<4> nyef: oops, first time I saw it. I'll quit right now.
<5> Nah, don't worry.
<5> I'll just go looking for food or something.
<4> nah, stay. I'll think about LispOS instead.
<6> on linux, it looks like it doesn't even try to make the color correct.
<1> spiaggia: perhaps, but I knnow with auto hinting and sub pixel rendering on, fonts on my LCD monitor look horrible, they are blurry, and have colored halos
<6> on osx, it appears to color compensate, so that it is never like a black line with a yellow shadow
<3> mbishop: That does sound like you either have a low-resolution screen or the wrong subpixel-ordering.
<6> rydis: no really, it's just what it looks like. the algorithm is wrong, I've looked at it in a blown up screen capture.
<6> maybe it's better in more recent versions, I'm using FC3 so I'm a little out of date
<1> rydis: 1280x1024 on a 17" LCD monitor
<2> mmm lispos
<1> someone should write an X server in lisp and then they can make sure the sub-pixel rendering is correct :P
<3> mbishop: I'd consider that a too-low DPI.
<2> like, point me to a company that can back a foss project for one, then I could try some foot-work about drivers
<0> rydis: that resolution on 17" is pretty normal. so rather normal DPI.
<2> imo, there's grounds for a new company about it. I know, this ain't #business but what the hey, heh.
<0> I doubt that my 15" on 1024x768 has more and the fonts are perfectly readable dispite their tiny size (yeah which gives me additional lines in my emacs buffer!)
<2> that new company could take an approach like apple has - develop the OS and the hardware - and could make sure it works, firstly, on intel arch
<3> therp: Yes, I know it's normal. I use 1600x1200 on good CRT 17", and 1400x1050 on fuzzy ones.
<1> rydis: 96 dpi
<7> gimbal`: that's not the approach apple took - they took an existing OS and threw all their legacy architectures on it
<8> rydis: What size font do you then use for such a display?
<3> rtoy: 6x13, mostly.
<8> Isn't that, like, really tiny?
<2> xarq: ahah; well, if we'd argue on it, they develop an OS and a hardware architecture; that's all I'd said they do, really. It's interesting to hear some history on it, though
<7> I starting liking tiny fonts after running out of space
<3> Not really. But it's a bit small for most people, I guess.
<7> gimbal`: so you want to form a LispOS company? xD
<2> did apple have anything to do with the development of os/2 ? all I notice is that it has a name similar to os/9
<7> gimbal`: no
<2> xarq: I would probably not be the right person for it, but there is a room for it
<3> For Hemlock I use 6x10, though, so I can fit a bit more. :)
<2> I ain't "the ceo type"
<2> xarq: <nod>
<7> gimbal`: I think Ubunutu makes it very hard to be profitable as an OS company
<8> rydis: Ok, that's not too bad. At least on my 21" CRT running at 1920 by something.
<7> gimbal`: they make it profitable to run *on* ubuntu and make money, like google
<8> But 6x10 is a bit too small, even for me.
<7> gimbal`: novell, which is close to the top of linux companies, is struggling, as is redhat, as is microsoft



<2> xarq: I've noticed that ubuntu has a .com domain, and it's a project seeming to have been formed by a company, Canonical LTD, but I haven't thought of ubuntu as being much involved with the business sphere, save in that they have a good looking linux distro, "good looking" & it's based on debian
<7> gimbal`: so to develop a LispOS as a company you would need a stategy like google
<2> xarq: strategy, there's a good thing to hear of a need & potential for, in regards to business; aggreed
<7> gimbal`: what I'm saying about ubuntu is they are free-as-in-beer so it's hard to be an OS vendor
<2> xarq: naively, but I've hoped that some interest might become involved of IBM, Sun, and Franz
<7> gimbal`: they might give some startup backing if you have a wicked idea
<2> xarq: hm; might you know whom would i need to potray an idea to, for it?
<7> gimbal`: no
<1> wishful-thinking@ibm.com? :)
<2> there's the rub, hehe
<9> gimbal`: PG, obviously.
<3> Or possibly Ron Garrett; he's been talking about having some dough to throw around. ;)
<2> heh; i tried bringing cl to the attention of the Eclipse project, once. it's a project seeming to be headed by IBM, but producing some FOSS java components, I don't recall what the systematic focus is. fwiw, they have an interesting nntp setup
<2> pkhuong: PG?
<7> gimbal`: develop a nice google-style business plan and a LispOS prototype with about as many features as QNX and you've got a shot
<1> or talk to norvig about getting google to do something for lisp :)
<2> xarq: I've found it challenging to seperate business interest from systems-development interest; I guess I'll have to work on that
<7> I wonder how much they're doing with lisp already...
<9> mbishop: that(CL and Smalltalk)'s a FAQ at google. The answer is "No".
<2> wondering if Y Combinator would take a look at it; they're HQ'd in Boston, though, which is a continent away from the city where I been stuck
<3> gimbal: Y Combinator is PG:s VC firm, so you probably do know who he is. ;)
<2> PG, Paul Graham, ok
<10> in the next version of lisp, can we make maphash accumulate values?
<7> pkhuong: where might that google FAQ be?
<5> Why? Isn't LOOP enough?
<10> being the hash-keys blah blah blah
<9> xarq: some random blog. Possible garret's.
<2> honestly, thinking of presenting the proposal to Paul Graham, I could try to portray how I look at it : like a novice monk endeavoring to propose a plan to Confucius, "or so"
<5> How would you monetize a LispOS?
<3> slyrus: Why not just make the function you call push the results onto a list?
<10> rydis: I just expect it to work llike mapcar, that's all. it's a minor thing.
<2> nyef: good question, especially if it's made in a FOSS project
<5> Let me know what you come up with?
<2> fwiw, google has managed to monetize a foss search engine system, and they've used averts during it, in their web-based stuff, and they do search-engines for companies, that much I know
<2> er .. s/foss/free/ and they use linux for it, though; they even developed a special filesystem on linux, for running their search engine
<11> LispOS talking?
<7> the problem with being ad-based is you go extinct when the next depression hits, like excite, lycos, etc. etc.
<2> yas
<7> or seriously downsized and bought by other companies, like yahoo and friends
<10> gimbal`: google's search engine is FOSS?!?
<7> slyrus: built on FOSS
<2> yahoo is owned by someone not yahoo? i did not know
<10> xarq: that's a huge difference
<2> slyrus: nah it was a typo, sorry heh; they do run it on linux though, afaik, judging by that they made a linux filesystem for their systems
<7> slyrus: very
<10> lightning and a lightning-bug
<5> "This lightning has a bug in it!"
<12> hmm, is there a way to help adjust-array resize an array efficiently (i.e. without creating a new one, and copying the old contents over)?
<9> make it an adjustable array?
<12> pkhuong: tried it, then it's not simple anymore, and the penalty is even bigger
<9> penalty for what?
<13> foom: sb-debug:var works for me
<12> well adjusting might be faster, but access seems to be slower (position, subseq, ...)
<6> jsnell: hmmm (defun foo (x y) (let ((z (+ x y))) (/ z 0)))
<5> So, nobody has any better idea for monetizing a LispOS than selling ad space on the website?
<14> jsnell pasted "example" at http://paste.lisp.org/display/22822
<3> michaelw: If you know the maximum size in advance, you can allocate an array of that size and do the book-keeping/bounds checking yourself, but that's the only way I can see.
<7> well, that brings us back to the "lisp needs a sabdfl" discussion, to make a free-as-in-beer lispOS from which revenue can be realized later
<6> it only shows x and y for me, not z, no matter what debug level
<12> rydis: i know, buf i do not know the size in advance
<13> you're only using z once, it gets optimized away?
<5> "sabdfl"?
<7> hmm.. is it just bdfl?
<7> forgot what the sa stood for
<7> "benevolent dictator for life"
<12> fwiw, i can get a 6x speedup if i do buffer management myself (keeping a list of fixed-size chunks instead of using adjust-array)
<2> if anoyne would happen to come across discourse by anyone at google, about how they've managed to keep the gumption to make money on their free search-engine service, or would know of a place to look maybe in regards to such, I could appreciate to hear of it
<15> "keep the gumption"?
<6> hm, yes, seems to work there, if I use it twice...
<2> taking google as a paradyne (?) here might not be exactly the right approach, but that's the first thing I can think to get a hold on
<7> gimbal`: I could, but it'll cost you :P
<2> hah
<3> michaelw: Then I don't think it's possible. You need to either use a non-simple array, which makes for another level of indirection and more expensive access, copy, or do manual book-keeping.


Name:

Comments:

Please enter the result of the sum 63 + 46 (to avoid spam):






Return to #lisp
or
Go to some related logs:

#css
load_privatekey import
SAXParserFactoryImpl gentoo
gentoo _cfg net.lo
python __getatt__
install vbulletin +oracle +create database
#debian
distupdate+debian
mythtv lspci -v not listed
inspiron 9300 xorgconfig



Home  |  disclaimer  |  contact  |  submit quotes