freenode/#lisp - IRC Chatlog
Search
18:47:16
jasom
sorry, the disciminant (and B, depending on your thoughts of having the sgn function in your equation).
18:48:00
jasom
It's the most well known function that involves catastrophic loss of precision, so it's the canonical example.
18:48:23
jasom
The pythagorean theorum can show some oddities when you are far from the origin as well, which is yet another argument for not using floats for coordinates.
18:49:11
White_Flame
yeah, and I've done some 3d graphics stuff far away from the origin which starts aliasing badly as well
18:49:25
jasom
There are approximately zero cases in which it makes more sense to use a double float than a 64-bit integer for a coordinate system. (doubles made some sense when it was faster to add 2 doubles than 2 64-bit integers).
18:51:14
aeth
If you want performance over accuracy and it's for a game and you're going to be converting to float dozens of times a second...
18:51:41
White_Flame
also, you'll be dealing with fixed point multiplication & reducing from 128-bit results
18:52:29
aeth
jasom: the thing is, in games you want performance, not precision, and in scientific calculations, you can just use a fancier algorithm to get your error small enough since it's inexact anyway
18:53:22
Bike
i think most scientists' engagement with numeric algorithms is ignoring matlab warnings that a matrix is near singular
18:53:32
aeth
(and many of the issues with float are the issues with not using hardware decimal float, which is the fault of hardware makers, not float)
18:54:48
jasom
Bike: more seriously the field is split between those super familiar with the limitations of the computational hardware, and the rest of the field, where the best you can hope is that they use libraries written by the former.
18:54:50
aeth
Bike: well, most of this thought should be moved to a library and not thought about at each use, e.g. a CL implementation of BLAS
18:55:24
Bike
i mean, i've generally been impressed by matlab's treatment. documentation with citations is pretty nice
21:06:27
jasom
Hmm, CLISP's long-float pi value is off by more than one ULP for very large precision values
21:35:33
pjb
Bike: it's different in physics. For example, you will find that the circumference of the Earth is not 2πr, but it's smaller. Because of relativity.
21:57:45
no-defun-allowed
I try to program about 12 hours a week, but I'd say I'm only productive for 4 to 6 hours.
22:00:41
clothespin
are you able to get something good done when you only have an hour or so before you have to go somewhere?
22:02:08
jasom
clothespin: also, keep notes *somewhere* about what needs to be done next before you leave; that means stopping work 5 minutes earlier than you would otherwise, but if it saves you an hour of time to zone-in on the otherside, it's easily worth it.
22:02:30
Oladon_work
It's a fallacy that you have to be "zoned in" in order to be productive — to maximize productivity, perhaps, but if you've scoped out the project (or part of the project) well and appropriately, you can pick up one piece much more easily.
22:03:23
jasom
clothespin: I'm more thinking a brain-dump; I tried X, Y, and Z, they didn't work because Q, so I'm going to change directions and investigate W next.
22:03:42
jasom
enough touchstones to get your brain back to where it was the last time you were working.
22:06:45
jasom
That's partly practice, and partly self-introspection. You will get better at getting in the flow with practice, but you'll also want to analyze why can't you do so without (for example) checking your e-mail?
22:08:45
jasom
IMO that's partly due to them having poor time management skills ;) but also some types of coding are more routine than others. I can speed up a poorly performing program with my brain half-off, and similar for porting code. For debugging something (particularly something hard to reproduce), I need maximum focus.
22:09:37
clothespin
yes, i agree. when i was running test scripts ob builds i could hack for 5 minutes
22:10:16
clothespin
now that i'm doing more architecture and i want elegant solutions i have to slow down
22:11:17
pjb
clothespin: https://www.amazon.fr/Coders-Work-Reflections-Craft-Programming/dp/1430219483/ref=sr_1_1?__mk_fr_FR=ÅMÅŽÕÑ&keywords=coders+at+work&qid=1569363071&s=gateway&sr=8-1
22:12:59
jasom
clothespin: for that, it can help to invert things a bit; collect problems and whenever you learn a new elegant solution, try each problem against it.
22:16:45
clothespin
jasom: what do you mean exactly? I code crap until i see the elegant solution, then i rewrite until the crap goes away and repeat
22:19:51
clothespin
I literally dreamed of a solution to avoid extra destinations for coordinates in generating vertex buffers
22:32:40
clothespin
well thanks for making me feel normal peeps, those guys with the macbooks on the train had me worried i was a less productive hacker
22:35:17
clothespin
I think Andreas Fuchs and Carl Shapiro doe that too but those guys are just exceptional
4:40:48
beach
Wow, the Wikipedia article on CMS is so general that it could work as a description of a text editor too.
4:42:56
p_l
the way my (primitive so far) blog works would be very well a representative of "headless" CMS (I use Emacs + org-mode + custom lisp code to output data that is then compiled into HTML)
4:43:56
no-defun-allowed
The first thing I think of is Coleslaw <https://github.com/coleslaw-org/coleslaw>
5:46:46
pjb
beach: that's unfortunately the problem with most natural language definitions, to be too general. See for example, how difficult it is to characterize "river", to be able to designate the shortest river of the world: https://www.youtube.com/watch?v=WEGzvZ85dgs
5:49:49
pjb
see also: Zawinski's Law: “Every program attempts to expand until it can read mail. Those programs which cannot so expand are replaced by ones which can.”, and Greenspun's tenth rule: "Any sufficiently complicated C or Fortran program contains an ad hoc, informally-specified, bug-ridden, slow implementation of half of Common Lisp."
5:58:42
beach
pjb: It sometimes a problem with the definition itself, and sometimes a problem with the concept not being precise enough.
6:00:15
pjb
beach: granted, but people (neural networks) don't seem in general have any problem classifying instances. What happens is that two different neural network may differ on extreme cases.
6:00:21
beach
pjb: And for programs that expand, I think that is a phenomenon related to the difficulty of making programs work together in primitive systems such as Unix. There would be no need to do that kind of thing in CLOSOS.
6:01:46
beach
Which reminds me, instead of imagining a "web browser" in Common Lisp, it would probably be better to imagine separate systems for formatting HTML, for showing PDFs, for viewing MPEG movies, etc.