jonathanzuniga has quit [Quit: Connection closed for inactivity]
<alexgordon>
cool
<ja>
hi
<purr>
ja: hi!
<ja>
’sup, purr?
<brixen>
ELLIOTTCABLE: what was your impression of _why's Potion?
<ja>
Oh my, I haven't heard of Potion in half a decade
<ja>
> I avoid @, #, $, %, {}. Stick with ., |, (), [], =, !, ?. Easier on the eyes. These are common punctuations in English.
<ja>
That's pretty awesome. | is common English punctuation though?
<creationix>
so redirection in shell is backwards
<creationix>
< and | are like function application, but > is like assignment
<creationix>
shouldn't < and > be like application and | like assignment?
<creationix>
ls > grep | results.log
<alexgordon>
yeah I'm not really sure where you're going with that
<alexgordon>
in haskell it is $
<alexgordon>
foo $ bar $ baz foo
alexgordon has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
alexgordon has joined #elliottcable
alexgordon has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<ja>
dddddddddddddd
eligrey has quit [Quit: Leaving]
<pikajude>
| is common english punctuation?
<ja>
¯\_(ツ)_/¯
<ja>
_why thinks so
Rurik has joined #elliottcable
Rurik has quit [Quit: Leaving]
<jfhbrook>
| is not particulary common in englilsh imo
fujisan has joined #elliottcable
alexgordon has joined #elliottcable
alexgordon has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
alexgordon has joined #elliottcable
alexgordon has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
alexgordon has joined #elliottcable
alexgordon has quit [Client Quit]
<ljharb>
um, how is | english punctuation
lispykarma has joined #elliottcable
<lispykarma>
know what i got once told by a series of bad dreams? about uniform AI...
<lispykarma>
heLL no, will that scene in Top Gun, "What you should've done is land that plane" ever be comprehensible, like...
<lispykarma>
is he using should in the shoulder/badge sense, is he talking about it in the 3rd person like he'll always think about it differently, etc etc etc... like, it worked.. ah, u gotta feel me by now eh?
lispykarma has left #elliottcable ["bashfu1"]
alexgordon has joined #elliottcable
<fujisan>
ELLIOTTCABLE: not the mama
eligrey has joined #elliottcable
alexgordon has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<ELLIOTTCABLE>
hi all
<ELLIOTTCABLE>
who the fuck was lispykarma, wat
<ELLIOTTCABLE>
brixen: I don't remember much, if at all, except “This looks really boring” and then that precise feeling inspired me to start making languages
<brixen>
heh
<ELLIOTTCABLE>
brixen: a sort of “Wait, not every worthwhile language is Lisp or Haskell: sometimes you can just splash a slightly different syntax over JavaScript, or address some inconsistencies in Ruby, or add immutability to Lisp, and ... you actually might *have something useful to real people. O_O” moment
<brixen>
I remember being all like "OH C'MON _WHY, WHY?! We have Ruby", then I looked at the syntax and was like cool, then he disappeared :'(
<ELLIOTTCABLE>
where ‘language design’ went from a high plateau of influence reserved for the cognoscente, to a real thing I could really do?
<ELLIOTTCABLE>
I heavily and strongly support _why's departure, for a multitude of reasons.
<brixen>
yeah, that's the attitude!
<brixen>
I meant the previous attitude
<ELLIOTTCABLE>
He was a close-friend of a very-close-friend of mine, so I had tertiary interest in his well-being;
<ELLIOTTCABLE>
but moreso, he provided a *great example* for all the hordes of women in tech I *currently* care about:
<brixen>
I certainly hope he's well
<ELLIOTTCABLE>
“You can ... leave. If it's making you unhappy, don't fucking fight the good fight, don't stick around because your skills are needed on some important project; do what fucking makes you *feel okay.*”
<brixen>
yeah, that's very good
<ELLIOTTCABLE>
Also, he really helped bring mental-illness-concerns to the fore as something that *it's okay to share*, if you want to share, with the community and your compatriots.
<brixen>
yep
<brixen>
he left a lot of good stuff
<ELLIOTTCABLE>
ugh @microSFF is the besttttt
<ELLIOTTCABLE>
bbl shower and dogs and food and etc
<ELLIOTTCABLE>
oh wait, lispykarma might have been a dude I met on campus ... or possibly lucien / @8bit_ghost?
<ELLIOTTCABLE>
oh wait nope he's in here as LucIenn I'm dumb
LucIenn has quit [Quit: Connection closed for inactivity]
<ljharb>
yeah microsff is great
alexgordon has joined #elliottcable
LucIenn has joined #elliottcable
<LucIenn>
Elliottcable: I haven't met you anywhere irl yet
<ELLIOTTCABLE>
LucIenn: two different people!
<ELLIOTTCABLE>
I met a guy named Kyle who's older / returning-student as well, at IIT
<LucIenn>
ELLIOTTCABLE: ah
<ELLIOTTCABLE>
turns out his area of interest is computational linguistics
<ELLIOTTCABLE>
so I invited him in here
<ELLIOTTCABLE>
ljharb: hi!
<ELLIOTTCABLE>
also lol hi LucIenn
<purr>
lolllllll
<ELLIOTTCABLE>
(=
<ljharb>
hi!
<ELLIOTTCABLE>
man what the fuck purr ಠ_ಠ
<brixen>
ELLIOTTCABLE: I'm experimenting with a "specification" system again, forewarned :p
<brixen>
ELLIOTTCABLE: I've reached the conclusion that there is no difference between "docs", "specification" and "tests" and there's no going back
fujisan has quit [Quit: Connection closed for inactivity]
<ELLIOTTCABLE>
brixen: hm?
<ELLIOTTCABLE>
I believe there's a subtle difference
<ELLIOTTCABLE>
but I also have Opinions on this so yay let's discuss them some say
<ELLIOTTCABLE>
… when I'm not about to leave for Boystown
<ELLIOTTCABLE>
:P
<brixen>
well, "no difference" is the wrong way to phrase it, I mean, "inseparable"
<ELLIOTTCABLE>
strongly agree
<ELLIOTTCABLE>
well
<ELLIOTTCABLE>
no
<ELLIOTTCABLE>
also still disagree
<brixen>
but yes, I'm very interested in your Opinions
<ELLIOTTCABLE>
would posit: ‘separation had a high cost’ and ‘separation has no value’
<ELLIOTTCABLE>
but not ‘not separable’
<brixen>
yeah, agree
<brixen>
well, if you separate them, you pay a high enough cost that you diminish the value of each and significantly diminish the overall system value
<ELLIOTTCABLE>
I really love dynamic, integrated stuff like the one that runs tiny JS one-line tests inside doc-blocks and verifies their output against //=
<ELLIOTTCABLE>
and I also have Strong Feelings about “self-documenting code”
<ELLIOTTCABLE>
my feelinngs on that will Shock and Amaze you
<ELLIOTTCABLE>
admission is 1 nickel
<brixen>
here are the 5 parts of my system: earth, wind, fire, water, metal
<ELLIOTTCABLE>
there are beards involved and also maybe tigers
<brixen>
oh wait, that's something else...
<ELLIOTTCABLE>
(the tiger is just me in a tiger suit)
<brixen>
heh
<ELLIOTTCABLE>
(but it's still p. cool)
<whitequark>
ELLIOTTCABLE: are you a furry now
<ELLIOTTCABLE>
5:48 PM <brixen> well, if you separate them, you pay a high enough cost that you diminish the value of each and significantly diminish the overall system value
<ELLIOTTCABLE>
whitequark: was I ever not? ;)
<whitequark>
TIL
<brixen>
5 parts of my system are: docs, specification language, source language, "coordination / run system" and the system itself
<ELLIOTTCABLE>
brixen: can actually summarize easily:
<ELLIOTTCABLE>
don't have time but that is okay:
<brixen>
these 5 are irreducible
<brixen>
ELLIOTTCABLE: run run, I'll keep you forever :p
<ELLIOTTCABLE>
I think documentation should be code, but I don't think documentation should be *the same code* as the runtime code. #boom
<brixen>
anyway, most systems try to reduce those 5 to 1 and then it sucks and people have pain and sadness
<ELLIOTTCABLE>
I also think specifications/tests should be code, but they should not be the same code as the documentation-code nor the hotpath-code
<brixen>
ELLIOTTCABLE: we're very close on this
<whitequark>
"code" is such a vague notion
<whitequark>
rspec-style tests are code
<ELLIOTTCABLE>
I think the very best way to document a system is in nearly-equivalent, but simplified code: unoptimized, without the security concerns or verification steps … possibly not even in the same language (and there's a special insight, there.)
<whitequark>
contracts are code
<brixen>
rspec-style is exactly the reducing 5 to 1 that I'm talking about
<whitequark>
constructive coq proofs are code
<ELLIOTTCABLE>
whitequark: yes, that's what brixen means by specifications
<ELLIOTTCABLE>
brixen: idk 5, but I definitely see the three I listed.
<ELLIOTTCABLE>
here's what I'm picturing for Paws projects:
<whitequark>
ELLIOTTCABLE: "without the security concerns" lol
<purr>
lolllll
<ELLIOTTCABLE>
a “documentation block” implementing the *purpose* of a chunk of code, in the simplest way possible, with clear naming, and an omission of real-world concerns
<brixen>
ELLIOTTCABLE: "specification" above in my 5 is a specific, non-code-code for an operational semantics of the code-code
<ELLIOTTCABLE>
an optional "implementation block", with the same thing, with verification, optimization, and whatever real-world messiness is necessary,
<whitequark>
brixen: i have no idea why rubyists like rspec that much, anyway
<brixen>
"specification" is something with precise logic that can be checked
<ELLIOTTCABLE>
and a "verification block" that *checks both*.
<brixen>
whitequark: because everyone likes to reduce everything to their favorite thing
<ELLIOTTCABLE>
in the sense that the documentation-block should be the simplest possible code that implements the tested/exposed API.
<brixen>
whitequark: 2nd reason, tooling is hard and it sucks, so why not do it all in the language
<ELLIOTTCABLE>
imho, all three should even be different *languages*
<brixen>
ELLIOTTCABLE: we are really, really close on this
<whitequark>
brixen: yeah, i guess it's all rooted in the thirst for eDSLs
<brixen>
exciting!
<whitequark>
which i also do not understand anymore
<ELLIOTTCABLE>
e == executable?
<brixen>
ELLIOTTCABLE: I'll try to put this into coherent prose
<whitequark>
embedded
<whitequark>
as in, abuse unrelated syntax to make something pretty
<ELLIOTTCABLE>
will read, but am now afk
<ELLIOTTCABLE>
whitequark: oh lol yes
<purr>
undefined
<ELLIOTTCABLE>
purr …
<ELLIOTTCABLE>
no
<brixen>
whitequark: (e,e)DSLs are the last resort of people without proper language tools
<whitequark>
brixen: yes!
<brixen>
(embedded, external)
<ELLIOTTCABLE>
brixen: where do you work?
<whitequark>
i really wish everything used OMeta
<brixen>
ELLIOTTCABLE: Rubinius, Inc
<whitequark>
except with a decent parser instead of PEG
<whitequark>
but even with PEG it could be good
<brixen>
whitequark: I'm building PEG into Rubinius a la LPEG
<whitequark>
brixen: have you seen the shift-resolve parsing paper?
<brixen>
it may not be the best, but it's the best bottom line, and someone can build up from it
<brixen>
whitequark: dunno, link?
<whitequark>
unlimited lookahead in linear time *and* space
<whitequark>
well, unlimited in the sense that it is fixed for any given grammar but arbitrarily large