<beneroth>
"Most dependencies in npm, Maven and Ruby are indirect dependencies, requested by the few libraries explicitly defined. Vulnerabilities in indirect dependencies account for 78% of overall vulnerabilities"
<beneroth>
Too many turtles. More dependencies = less control.
jibanes has quit [Ping timeout: 268 seconds]
jibanes has joined #picolisp
anddam has joined #picolisp
<anddam>
howdy
razzy has joined #picolisp
<Regenaxer>
Hi anddam!
<razzy>
hi all
<Regenaxer>
Hi razzy!
<beneroth>
heya
<beneroth>
Welcome anddam :)
<beneroth>
[OT] Certificate Authority Digicert does willingly sign Certs for private LAN IPs: https://groups.google.com/forum/#!msg/mozilla.dev.security.policy/JFwqZx7RLL0/My83uQA5AAAJ
<beneroth>
(use case: the smart light bulb or insecure printer in your LAN can MitM-attack your local routers admin-website)
<beneroth>
I hope Google & Mozilla kill Digicert for that failure. Incompetent companies going bankrupt is a positive and essential feature of capitalism (<> 'capitalism neo-capitalism)
freemint has joined #picolisp
<razzy>
i always felt that symbols should have properties, and those properties should be "live" and "inherit-able". never found language that do that.
<freemint>
beneroth, only one part of your inequality is coded
<freemint>
*forget that
<beneroth>
freemint, yeah was a typo :)
<beneroth>
pseudo-code anyways
<Regenaxer>
razzy, look at 'var'
<Regenaxer>
is inherited
<beneroth>
picolisp can do it.
<beneroth>
T
<razzy>
i know picolisp has OOP as inheretable properties. but in a way i would not do, so it is hard for me to internalise. i did not found a way to think in it yet.
<beneroth>
why do you never listen?
<beneroth>
<Regenaxer> razzy, look at 'var'
<beneroth>
var has nothing to do with the OOP system
<beneroth>
look it up in the ref. try it out. then you can make clever comments.
<beneroth>
<razzy> i always felt that symbols should have properties, and those properties should be "live" and "inherit-able" AND "but in a way i would not do, so it is hard for me to internalise. i did not found a way to think in it yet." - HOW does that fit together? You have feelings how something ideally should be (without communication your reasoning for believing so), but additionally you write yourself you don't know how to deal with such a system? WUT?
<beneroth>
think more specific, dude. and try your assumptions out instead of just endlessly theorizing about it in your head using incomplete information (as you haven't run any experiments to test your hypotheses, intuitions and ideas, I presume).
<beneroth>
this unspecific "hey I have the most awesome ideas but no actual arguments or insights"-hipsterism is really annoying me.
<beneroth>
in my view that is the antithesis of picolisp spirit. and scientific method anyway.
<beneroth>
we IT people should a lot more view ourselves as computer "scientists" (!). we must dare to do stuff properly.
<beneroth>
I know it's Don Quixote's battle against the windmills, but its still "the right thing to do" (TM) I believe.
<razzy>
beneroth: i think you misunderstood what i said. i said "i know picolisp has something like that, i invest too little to understand it yet"
<beneroth>
yes that I understood perfectly fine.
<beneroth>
I demand you to properly invest yourself according to your possibilities before asking questions here.
<razzy>
it was not a question. it was acknowledge "i like everything as symbol with properties"
<beneroth>
I demand you do something with the responses you get and show that you do something wit it, instead of reacting with stupid comments to the generosity that someone took time to answer you.
<beneroth>
I would not tell you that if this would not happen again and again with you. It really annoys people here, not just me.
<beneroth>
You can ignore my recommendations to you. But people here will eventually put you on /ignore so they don't see your writings anymore.
<beneroth>
just fyi.
<razzy>
people should warn with /ignore themselfs.
xkapastel has joined #picolisp
orivej has quit [Ping timeout: 250 seconds]
freemint has quit [Remote host closed the connection]
freemint has joined #picolisp
<freemint>
Do you an estimate what overhead OO brings in PicoLisp speed wise?
<freemint>
I have a side project with lot's of symbolic shuffeling around which implements existing algorithms which are mostly specified at "array level".
razzy has quit [Ping timeout: 264 seconds]
<freemint>
It is not like the project is performance critical and i see the benefits of extensiability and readability very cleary but i do not want to make a dog slow by accident ;)
jibanes has quit [Ping timeout: 246 seconds]
jibanes has joined #picolisp
<beneroth>
freemint, you mean using picolisp with OOP vs. using picolisp without OOP (e.g. with custom symbol or list structures) ?
<beneroth>
I would estimate that properly using the OOP system speeds things up generally. OOP in picolisp is a rather lightweight extension on top of lisp symbols.
<freemint>
beneroth, yes
<beneroth>
It surely speeds up development time and improves maintainability greatly - unless one builds an overcomplicated over-engineered OOP bloat structure - but that is always the risk with OOP.
<Regenaxer>
ret
<beneroth>
wb Regenaxer
<Regenaxer>
:)
<beneroth>
freemint, I would recommend to first try to design your program nicely (as reduced/simple but readable / human-manageable as possible) and then taking measurements if the performance is lower than desired.
<Regenaxer>
However, there is some overhead in Pil OOP, dynamically looking up methods
<Regenaxer>
beneroth, exactly
<beneroth>
then individual steps (e.g. method implementations) could always be offloaded into another program or picolisp asm lib if required, I think.
<freemint>
the problem is boolean satisfaction. I have a expression made out of logical "verbs" like and, not, implies ... and "symbols" which maybe true or false and the problem is to choose values for the symbols so the expression becomes false
<freemint>
*True
<beneroth>
Regenaxer, T, but I suspect using some other advanced list/symbol structure (or whatever) is probably having that lookup overhead too, just in a worse looking packaging.
<beneroth>
freemint, sounds like pilog!
<Regenaxer>
I suspect so too
<Regenaxer>
yeah, logic programming
<freemint>
beneroth, the difference is that it uses cleverer algorithms than pilog
<beneroth>
freemint, sounds like a task which can be solved entirely with pilog (without OOP) or with a mix of OOP and pilog.
<freemint>
and i want to understand the clever algorithms
<beneroth>
uh, bold assumptions. And I'm sure you could put those clever algorithms into pilog expressions, not?
<freemint>
These cleverer algorithms would not work on pilog statements since they are "to powerfull" => i would need to restrict myself too a limited input language.
<Regenaxer>
Sorry, must go again
<Regenaxer>
bbl
<beneroth>
check out the pilog tutorial in the official ref. and the blog at prodevtipps has some good examples.
<freemint>
Furthermore i will need to write my own backtracking for this, since the clever algorithms work by interacting stronlgy with it.
<beneroth>
I don't see how you would need to restrict yourself, but I cannot tell as I don't know your actual stuff.
<beneroth>
ok
<beneroth>
well the question is if you like to solve the problems, or if you mainly intend to re-implement the clever algorithms yourself for understanding them.
<beneroth>
:)
<freemint>
the latter
<beneroth>
then speed doesn't matter.
<freemint>
speed does matter, since i want to test it :)
<beneroth>
you could even do multiple implementations (using different picolisp features) to compare the speed relatively (that would be correct speed comparison)
<freemint>
i thought about that too
<beneroth>
it is unlikely that speed is that bad that it makes your testing infeasible ;-) if so, you can probably reduce the size of the test data set without sacrificing test value.
<freemint>
beneroth, not that easy.
<beneroth>
spend 2 hours trying to implement a reduced, simpler use case instead of thinking on how to do it :)
<freemint>
I did that once. It scaled poorly.
<freemint>
beneroth, the problem is already very minimal. You can not reduce it without making it trivial
<beneroth>
it is not meant for scaling. you want to gain insights, no?
<beneroth>
hm
<beneroth>
okay
<beneroth>
well I cannot tell. I wrote you my opinion and estimates. you must decide yourself which approach you follow. when you did actual experiments I would like to hear your insights :)
<freemint>
I got a strucure which looks like this: an and of many or terms which contain negated or non negated variables.
<freemint>
(AND (OR (not C) (not A) ) (OR A (not B) C ) ... )
<freemint>
If i have results it will tell you.
<freemint>
afk but reading logs
freemint has quit [Ping timeout: 250 seconds]
<Regenaxer>
ret
<beneroth>
wb :)
<Regenaxer>
:)
<Regenaxer>
Hmm, why implement one's own backtracking?
<beneroth>
<freemint> These cleverer algorithms would not work on pilog statements since they are "to powerfull" => i would need to restrict myself too a limited input language.
<beneroth>
*facepalm*
<Regenaxer>
hehe
<beneroth>
I cannot rule out that this is indeed the case, but I think its rather unlikely ;-)
<Regenaxer>
He could write his clever (tm) algorithms *in* Pilog
<beneroth>
that I tried to say. but apparently he only sees pilog as an "limited input language" :P
<Regenaxer>
oh
<Regenaxer>
input of what?
<beneroth>
no idea. facts? :D
<Regenaxer>
Anyway, no problem
<beneroth>
T
<beneroth>
he will find something out and tell us then about it :)