freenode/#lisp - IRC Chatlog
Search
12:24:20
decent-username
My code becomes really messy if I have to write (cell-low-max (aref cell *octant*)).
12:31:29
heisig
decent-username: My suspicion is you want to bind the place, not the value, so that (setf c ...) works, too. In that case, you should have a look at SYMBOL-MACROLET.
13:51:16
Kabriel
aeth: in my experience numpy is very slow compared to Fortran or C. Lisp is not as fast as Fortran or C, but way faster than Python.
13:51:48
Kabriel
I have tried the lapack routines in the f2cl packages in quicklisp; they work pretty well
15:39:15
minion
The URL https://gitlab.common-lisp.net/users/sign_in?secret=cabf5054 will be valid until 15:45 UTC.
15:44:34
goofist
Hmm.. I have tried to register with the common-lisp.net gitlab via my github account, but cannot. First I was directed here. But when I followed the link that minion gave me and clicked github, I was directed to an error saying that my account has been blocked.
15:56:06
goofist
In a DM with minion, i requested a second link and this time just tried to register with my email address. The registration process informed me that my address is taken. I thought that perhaps some kind of registration attempt was aborted so that my email is in the db already. So I tried to get a password reset link, and was informed by the process
15:56:06
goofist
that *if* my password exists in the database I will be sent an email. This is yet to occur. How can my email both be and not be in the database ?
16:12:14
goofist
arrrg. This is my punishment for not following through with the registration process earlier this year when I was directed to interact with the minion bot back then. I was like: ugg, how convoluted, no thank you. perhaps my registration state is still hanging from back then... sigh...
17:24:33
aeth
MichaelRaskin: I'm not sure if there is one, but if it was widespread there probably would be one. I don't do much web stuff, though
17:25:59
aeth
Kabriel: Ah, okay, I guess there's a high python-to-numpy-to-C/Fortran overhead, then. Not surprising.
17:26:10
MichaelRaskin
Well, maths where where just a properly considered alt text is not sufficient gives an upper bound on the spread…
17:26:14
aeth
Kabriel: f2cl lapack might "work pretty well", but run disassemble on them in SBCL and they don't come close to stuff written by hand with SBCL's optimizations in mind in terms of quality of generated asm
17:29:12
aeth
MichaelRaskin: technically speaking, every math formula is readable, like (to write some mathematical nonsense) "(a + b) / c = 42 ⇒ ε ∊ (-∞, 7]" is just "a + b over c equals 42 implies epsilon is in the range -∞ to 7 inclusive of 7" or something
17:29:51
aeth
Now, whether or not any tool is that smart, idk, but if you provide something like MathML (or LaTeX!) it could in theory be done
17:32:43
MichaelRaskin
In complicated LaTeX formula structure is sometimes sacrificed for output readability and aesthetics
17:33:29
aeth
Yes, which is why MathML would probably be better. Technically it's still "presentation MathML" but it's probably easier to parse
17:33:35
MichaelRaskin
Of course, one hopes that the maths you use _in documentation_ should not be _that_ complicated and even LaTeX might be readable
17:51:43
aeth
If there isn't a formal description of Common Lisp, you'd have to write it to cite it, of course.
19:48:36
Bike
stop me if you've heard this one before: The world is everything that is the case. The world is the totality of facts, n