<beach>
Thanks. I had no recollection of it, but now I can figure out more about it.
ello has quit [Remote host closed the connection]
ello has joined #commonlisp
pillton has quit [Quit: ERC 5.5.0.29.1 (IRC client for GNU Emacs 29.4)]
<jackdaniel>
I need to keep a mental note to not announce my new common lisp implementation on #commonlisp when I make one; after reading that log I feel discouraged by proxy!
<jackdaniel>
it's the vibe of this channel I think; someone comes and says "look at this something cool I've created" and what they get is "did you know about/ why not this particular way/ etc", while nobody looked at this from the angle of "wow, cool!"
<jackdaniel>
perhaps that's because we (as channel members) are too staurated with clever ideas?
* jackdaniel
shrugs
King_julian has quit [Read error: Connection reset by peer]
amb007 has quit [Read error: Connection reset by peer]
amb007 has joined #commonlisp
<beach>
I think those were mostly question to understand detailed design and motivation.
<beach>
... and to help with information that the person may have missed.
bendersteed has joined #commonlisp
<jackdaniel>
don't get me wrong, I don't think that questiosn were bad at all
<beach>
And it is hard to get excited when there are many design decisions that one finds suboptimal or otherwise strange. Though perhaps one should keep that to oneself and express excitement anyway.
<jackdaniel>
but I imagine that someone announcing a new common lisp implementation would expect at least a little excitement; that's not directed at you - rather at discussion surrounding the topic (from logs, I'm also present there)
<beach>
I see, yes.
<jackdaniel>
and nothing wrong with not feeling excitement, but it may be discouraging; I tend to notice more these kind of things as the little one grows
<beach>
I understand.
rainthree has joined #commonlisp
King_julian has joined #commonlisp
mishoo has quit [Remote host closed the connection]
mishoo has joined #commonlisp
treflip has joined #commonlisp
King_julian has quit [Read error: Connection reset by peer]
attila_lendvai_ has quit [Ping timeout: 244 seconds]
<rainthree>
'it's the vibe of this channel I think; someone comes and says "look at this something cool I've created" and what they get is "did you know about/ why not this particular way/ etc", while nobody looked at this from the angle of "wow, cool!"' yes, giving encouragement and then very slowly moving towards improving/correcting the thing, is what gives better results (psychologically speaking) in most cases.
<rainthree>
Another more extreme approach, is to even support wrong things. For a certain percentage of the population, it's the only way to help them (i.e. attract them towards the right path). Otherwise that part of the population will remain in complete darkness, attracted by even worse things.
<rainthree>
The biggest downside of this approach is that very soon it will be forgotten that the only reason the wrong things were being talked about as if they are good, was to help those who can't be helped otherwise (the big majority). What's more: in order for this approach to be effective, you must never publicly mention or admit that what you said was only for this reason, and it was a lie. You have to lie 'till the very end. It
<rainthree>
is the approach that has been used by the religions.
FragmentedCurve has quit [Read error: Connection reset by peer]
FragmentedCurve has joined #commonlisp
<jackdaniel>
the fallout of being sceptically received on the community channel is usually a) moving to another place featuring the language, b) changing the language
<jackdaniel>
there are exceptions of course
<jackdaniel>
so that's probably the key difference between professional activity where you will keep doing /the thing/, and own-time activity where a certain threshold of anxiety makes you hop onto another topic
<pranav>
Dishonest appreciation is worse than scarce but genuine one IMHO, and will dilute the joy people get out of sharing. At least you can trust the criticism is meant. What works for me is to take the time to empathize with the intent and effort behind what's presented while exploring it, and avoiding skin-deep judgments of their work while commenting/questioning. Expressing appreciation first works, but ha
<pranav>
s to be genuine.
* jackdaniel
agrees
<jackdaniel>
that's why I've mentioned (on #commonlisp), that it is probably the problem of people being exposed to too many interesting ideas
<pranav>
I like how Taleb puts it: "No politeness, without warmth."
<jackdaniel>
i.e it is much harder to impress #commonlisp crowd than other crowds
<jackdaniel>
and that preasumbly impacts the amount of newcomers who are swiftly dealt with by redirecting to #clschool or #clnoobs (I don't remember which is the current one)
<jackdaniel>
to put it less politely, people on cl are often jaded and (at the same time) infatuated with their own stuff (me included! :)
<rainthree>
pranav: If a beginner comes to the realization that 1+1=2, I can appreciate his efforts, and push him forward in a way that he feels encouraged
<scymtym>
to be fair, criticism and pointers to prior work are often qualified with "if you are making your own lisp/compiler/gui/testing framework as a learning project, that's fine of course"
<jackdaniel>
that sounds a bit condescending to me; it is fine to work on any delusion at ones discretion without any conditions; working for someone is another story of course
<bjorkintosh>
sometimes, ignoring the helpful critics is best for creative work :-D
<jackdaniel>
but perhaps for native english it does not sound it that way
<jackdaniel>
s/sound it/sound/
<jackdaniel>
the prhas
<bjorkintosh>
Imagine Dijkstra talking to a beginner interested in programming.
<jackdaniel>
the phrase "if .. .then that's fine"
<bjorkintosh>
best to ignore him and stick with BASIC until one gets comfortable.
<jackdaniel>
I'm ignoring him to this day, because "goto is harmful" is harmful ,)
<bjorkintosh>
good. we're not all writing programs to control rocket ships.
<bjorkintosh>
hell, these days it might be easier to make a GPT do the hard work.
yisef has quit [Quit: Client closed]
Everything has joined #commonlisp
Fade has quit [Ping timeout: 245 seconds]
Fade has joined #commonlisp
Artea has joined #commonlisp
yitzi has joined #commonlisp
random-nick has joined #commonlisp
dinomug has quit [Remote host closed the connection]
Everything has quit [Ping timeout: 252 seconds]
Everything has joined #commonlisp
King_julian has quit [Read error: Connection reset by peer]
mgl has quit [Ping timeout: 255 seconds]
mange has quit [Remote host closed the connection]
triffid has quit [Remote host closed the connection]
triffid has joined #commonlisp
dstein64 has quit [Ping timeout: 252 seconds]
rainthree has quit [Ping timeout: 260 seconds]
vyrsh has joined #commonlisp
<vyrsh>
how to post cute code snippets here? I want to just post small code here
<beach>
plaster.tymoon.eu
wacki has quit [Ping timeout: 248 seconds]
wacki has joined #commonlisp
Everything has quit [Quit: leaving]
yitzi has quit [Remote host closed the connection]
Posterdati has quit [Ping timeout: 252 seconds]
Th30n` has quit [Quit: going off]
treflip has quit [Ping timeout: 260 seconds]
JuanDaugherty has joined #commonlisp
Posterdati has joined #commonlisp
FragmentedCurve has quit [Ping timeout: 248 seconds]
Fade has quit [Changing host]
Fade has joined #commonlisp
Posterdati has quit [Remote host closed the connection]
ndanilov has quit [Read error: Connection reset by peer]
<beach>
You are supposed to use a paste site for more than one line of code.
<vyrsh>
sorry
<vyrsh>
was just trying
<vyrsh>
I have a question
<beach>
What does ARR mean here?
<vyrsh>
list, its the argument in the flatten function
<vyrsh>
array
<beach>
You can't take the CAR and the CDR of an array.
<beach>
And (EQ ARR NIL) is better expressed as (NULL ARR).
<vyrsh>
oh thanks
<vyrsh>
I wrote arr instead of list
chiselfuse has quit [Remote host closed the connection]
<beach>
And if ARR is an atom other than NIL, this will fail, because you will attempt to take the CAR of that atom.
triffid has quit [Remote host closed the connection]
<vyrsh>
you are correct. I only was thinking about lists as input
zwr has quit [Read error: Connection reset by peer]
<beach>
Do you mean proper lists? Because a list can be a dotted list.
<vyrsh>
list like '(a b c (d e) f)
<vyrsh>
few months ago, I wrote a small automatic back prop library in lisp. but it was for singular values and I wanted to use it in a small neural network library. so I figured I have to write automatic backpropagtion library for matrices.
<beach>
But a list can be (A B . C), which is why I pointed out that you were not thinking only of lists, but you were thinking only of proper lists.
<vyrsh>
oh ok
<vyrsh>
what are some important pointers I should have in my design of the automatic back prop matrices code?
<vyrsh>
I want to have it so when I do matrix operations like mult matrices, then I can do (backward matrixX) and all the matrices that result in the matrixX get a grad value on it
alcor has quit [Remote host closed the connection]
alcor has joined #commonlisp
zwr has joined #commonlisp
gorignak has quit [Read error: Connection reset by peer]
cercopith has quit [Ping timeout: 260 seconds]
matt` has joined #commonlisp
mgl has joined #commonlisp
matt` has quit [Ping timeout: 248 seconds]
matt` has joined #commonlisp
_whitelogger has joined #commonlisp
rakka has quit [Remote host closed the connection]
rakka has joined #commonlisp
shmorg83 has joined #commonlisp
Akbar-Birbal has left #commonlisp [#commonlisp]
<vyrsh>
ok so I read it.. it doesnt seem to have automatic back propagation.. I once wrote a neural network library in C and this code seems similar to that, where you are just writing a loop to do the backpropagation one by one for the layer and then the layer after. not really automatic backpropagation
<vyrsh>
I want to have some operations. and the library should be able to compute grads for whatever parameters I have used with the operations
<vyrsh>
I think I should make a list of operations first.. because Im also looking at special operations for neural networks that preserve symmetries
<jackdaniel>
you seem to know what do you want from the program - why not just write it then?
vyrsh has quit [Quit: Konversation terminated!]
<jackdaniel>
there's a machine learning library mal if you look for some more involved and production ready code
<jackdaniel>
mgl * not mal
<aeth>
what you want to do is design a DSL (domain-specific-language) that has the semantics that you want, and then implement that.
<aeth>
there probably isn't a library that does it exactly the way you want
eeeoo2 has quit [Quit: Leaving]
eeeoo has joined #commonlisp
amb007 has quit [Ping timeout: 260 seconds]
amb007 has joined #commonlisp
_whitelogger_ has quit [Remote host closed the connection]
chiselfuse has quit [Remote host closed the connection]
chiselfuse has joined #commonlisp
spdegabrielle has joined #commonlisp
eeeoo has quit [Quit: Leaving]
eeeoo has joined #commonlisp
_whitelogger_ has joined #commonlisp
jadzi has joined #commonlisp
amb007 has quit [Read error: Connection reset by peer]