<Virindi>
if you use file->quit, the window size is not remembered the next time openscad is started. If I click the x in the corner to close it, it works
othx has quit [Remote host closed the connection]
<Virindi>
world's stupidest bug :)
othx has joined #openscad
<teepee>
it even scared the bot ;-)
<InPhase>
Virindi: Well I guess that's not surprising. Clicking the button calls MainWindow::closeEvent() which saves the setting, while clicking File->Quit calls MainWindow::quit() which does not save the settings.
<InPhase>
Virindi: Pretty straight forwardly the code is doing what it's told. It's just a little bit of a silly design.
<InPhase>
There should probably be a centralized close cleanup function that gets called by both.
<teepee>
yeah, the crash at exit fix seems to have missed some opportunities
<InPhase>
Looks like this bug was inserted in July of 2011. :)
<teepee>
maybe the other issue masked it
<InPhase>
Perhaps.
<InPhase>
Or just you have to really be paying attention to notice it.
<teepee>
right, printing a smaller thread with PETG and no fan did not work
<teepee>
so a bit more tests are needed, just switched back to PLA for the cable guides
gunnbr has joined #openscad
gunnbr has quit [Ping timeout: 240 seconds]
<lf94>
InPhase: do you think we could build a house with OpenSCAD
<lf94>
"it's been done"
<InPhase>
lf94: Certainly.
<InPhase>
I don't know if any of the 3D printed houses have used it though.
<InPhase>
FreeCAD does not fit my purposes well, but one thing about FreeCAD is it was initially written by an architect.
<InPhase>
So it could conceivably fit the ambitions and workflow of an architect much better. :)
<JakeSays>
i tried using freecad a while back and it was painfully slow
<lf94>
Code CAD is the future
<JakeSays>
of what?///
<JakeSays>
damn kbd
<lf94>
Of computer aided design
<JakeSays>
ah no.
<lf94>
It is
<JakeSays>
it's very much a niche
<lf94>
Until it's not
<lf94>
math was very much a niche
<JakeSays>
it'll never not be
<lf94>
How can you say this
<JakeSays>
because it's an inferior way to design something.
<lf94>
How can you say this
<JakeSays>
i would much much much rather use a traditional cad system, but my mind doesn't work that way
<JakeSays>
i say this because the D in CAD is all about what people see.
<lf94>
Code CAD does not need to be exclusively text.
<lf94>
Case in point: OpenSCAD has a viewer right beside code.
<lf94>
There are further ways to augment the viewer too.
<JakeSays>
if anything the exact opposite will happen. as CAD systems incorporate more AI, etc. the need for less visual techniques will diminish even more
<lf94>
On the "code" side we've pretty much explored the space. Now we need a beautiful marriage between the viewer and code.
<JakeSays>
code cad is all about text - it's code after all
<lf94>
Code can use non-textual symbols
<lf94>
You can use code interfaces such as Scratch
<lf94>
etc
<lf94>
There are many benefits which you are not discussing
<JakeSays>
that's still 'text' in the context of code.
<lf94>
You've used OpenAI's Codex?
<JakeSays>
there's visual design and there's non-visual design. non-visual design is code
<lf94>
I don't know how the hell AI will help do any CAD.
<JakeSays>
really?
<lf94>
And that's bleeding edge AI
<JakeSays>
i can see many ways
<JakeSays>
but ai is just one example
<JakeSays>
add in VR/AR and things get even more visual
<JakeSays>
with AI the machine can understand what we're trying to design and automate a large portion of it
<lf94>
Could you explain to me how that would be used? Would you say what you want? Could you give me an example of expressing to AI what you would like? :)
<JakeSays>
well, voice is definitely one way
<JakeSays>
another way would be using crude sketches to design a house, for example.
<JakeSays>
it's hard to explain, but i can visualize it in my head
<JakeSays>
or i should say it's hard to explain using pure text
<JakeSays>
give me a whiteboard and i'd do a much better job
<lf94>
CAD involves precise figures. Giving a crude sketch will still leave the engineer to have to verify and adjust /everything/.
<JakeSays>
lf94: he will only have to adjust what the machine can't
<lf94>
if you are handing it crude sketches you will get crude results.
<lf94>
there is no way around it
<JakeSays>
that's where ai comes in
<lf94>
no, even if the ai was a human, the human can only do its best to guess the dimensions
<lf94>
now I'm really curious what you will say about "voice"
<JakeSays>
that depends on the sketch
<lf94>
explain to me how voice can be used / what expressions you would say to the ai
<JakeSays>
"design a bridge over Foo Canyon to handle X traffic."
<JakeSays>
the machine studies the geology of Foo Canyon, figures out the best place for it
<JakeSays>
it analyzes weather patterns over the past N years
<JakeSays>
determines the best bridge design, etc
<lf94>
I can see so many issues with that
<JakeSays>
that's because you're trying to see issues
<lf94>
"Handle X traffic" can be interpreted as optimizing for absolute throughput, which would maximize risk
<lf94>
hold on
<lf94>
so you have to refine further
<lf94>
and further
<JakeSays>
yeah obviously this is an example.
<lf94>
eventually you end up with what you wouldve started with
<lf94>
a full specification of what you need.
<lf94>
said specification could be done with code, in a DSL
<JakeSays>
and you would have gotten there by providing a 10th of the information in a 10th of the time
<lf94>
Doubtful, everything the AI generates has to be read and tested
<InPhase>
A proper use of AI for CAD would actually produce something like OpenSCAD source code.
<lf94>
Just as a human would do
<lf94>
InPhase: yeah, and could probably be formally verified.
<InPhase>
It woudl be necessary, because a sufficiently AI could guess structure for your design, but you would need parametric inputs then that you could adjust, which the AI could generate for you if sufficiently smart.
<JakeSays>
lf94: well of course everything would be verified.
<InPhase>
I know this, because people come in here with sketches of designs, and I make examples for them, and those examples almost always have the wrong dimensions, but just the right structure. :)
<JakeSays>
right
<lf94>
good bot
<lf94>
B)
<InPhase>
:)
<JakeSays>
it's an iterative, collaborative process
<InPhase>
So code CAD is still the future, even if AI helps.
<JakeSays>
all the more reason why code will be less and less involved from the users perspective
<lf94>
You're going to see code CAD become more popular in the next few years - there are lots of developments happening.
<InPhase>
Because fully parametric design with abstraction is just too powerful.
<lf94>
I plan to contribute back to OpenSCAD too - it is useful in its own ways.
<JakeSays>
InPhase: code cad won't be how the user interacts with the system.
<lf94>
you've been to the future?
<lf94>
:)
<lf94>
I have, everyone knows how to code in some form.
<lf94>
excel spreadsheets, form creators, even stock trading.
<lf94>
constructs of logic invade every crevice of our lives.
<JakeSays>
perhaps we have different definitions of code
<teepee>
must be a parallel universe without Apple
<lf94>
multiverse is real
<lf94>
etc
<JakeSays>
to me, even openscad is way too low level
<lf94>
I cannot get myself to finish this damn arc drawing code.
<lf94>
JakeSays: ha. there are even lower level constructs which a lot of people which it exposed.
<lf94>
JakeSays: there are abstractions which can be used to make it even simpler
<lf94>
and higher level
<lf94>
that's the beauty of code. it's a morphing machine.
<JakeSays>
but none of those people are architects. they are code oriented people.
<lf94>
just give me time
<lf94>
I will show you how architects will be ok with this
<lf94>
I need like 20 years to show you
<lf94>
so please dont leave the planet before then
<lf94>
I promise you
<lf94>
otherwise you can have ... my computer
<lf94>
B)
<lf94>
and i will live in the woods forever
<JakeSays>
you'll have to work faster than that. i'd be surprised if i make it 20 years
<teepee>
how can you promise woods in 20 years?
<lf94>
lmao
<JakeSays>
but for sure i definitely won't give a shit about any of this if i'm still around
<lf94>
you're right we may have less time
<lf94>
if we have less than 20 years cad wont matter
<JakeSays>
not "we" - me
<lf94>
teepee was saying we wont have trees in 20 years lol
<lf94>
i took that as the planet is fucked :)
<lf94>
but just for you jake I will try to make it happen sooner
<lf94>
B)
<teepee>
well, trees, I sure hope so, but actual woods are really on a bad path
<lf94>
I'm going to go full time with this stuff I think in 2024
<lf94>
Like it will be my job to design products with code cad
<lf94>
It will be a testament to its power
<JakeSays>
oh i fully grasp it's power
<lf94>
Imagine I'll be a one man design team with full production capacity
<lf94>
that's fucked
<lf94>
in terms of productivity
<lf94>
Four 3D printers all going - even the _packaging_ is done with code cad.
<JakeSays>
i have two printers and i could easily keep two more going, and that's just for my hobby shit
<InPhase>
JakeSays: I think programming is poised become a fundamental skill of all people upon graduating high school.
<InPhase>
s/poised/poised to/
<lf94>
Yeah pretty much.
<lf94>
It is
<lf94>
Thanks mostly to the web
<lf94>
and excel
<lf94>
It's amazing actually how Excel has introduced array programming to people
<InPhase>
It's only the current generations that struggle with code CAD. And, well, they're going to get left behind by the younger generations if they can't do these sorts of things.
<JakeSays>
InPhase: i think you WAY over-estimate, like way, the motivation of the average joe to be more technical
<lf94>
Average Joe, for sure
<InPhase>
JakeSays: I don't think it has anything to do with motivation. I think they will need to program the same way everyone needs to learn math in school.
<lf94>
Joe is on his way out of the world, fortunately
<InPhase>
Also at one point reading and writing was not common education. It's pretty much a mandatory life skill now in most of the world.
<JakeSays>
InPhase: i disagree. programming is hard, and most people can't do it, even at a basic level
<JakeSays>
just like most people aren't good at math
<InPhase>
And by "at one point" I really mean not all that long ago.
<lf94>
I'm actually afraid of all the young people who will have a logical thinking advantage over me
<InPhase>
lf94: Spend time on twitter. You'll be less afraid of that, and more afraid of the opopsite. ;)
<JakeSays>
InPhase: that's EXACTLY my point.
<lf94>
That's a small minority
<JakeSays>
ahhh no
<lf94>
There are actual smart young people
<JakeSays>
lol so not
<JakeSays>
of course there are
<lf94>
Are we really that fucked
<JakeSays>
but the majority of young people are not
<JakeSays>
lf94: YES we are
<JakeSays>
fortunately i'll be long dead. lol
<lf94>
code cad will make me rich then just because i can do what a full team of 10 brain cells will need to do
<JakeSays>
lf94: that's entirely possible
<lf94>
like...wouldnt the incentives just line up? more power = more money
<lf94>
companies will hire more cad coders
<lf94>
etc
<lf94>
game theory applies
<lf94>
InPhase: thoughts
Jack2276 has joined #openscad
<JakeSays>
dude have you noticed how desperately companies want to get out of the programming business altogether? that's why the "low/no code" movement refuses do die
<JakeSays>
*to
<lf94>
what companies
<lf94>
low/no code is pointless because it's just cookie cutter stuff
<JakeSays>
of course it is, but it's the hope of any company who has to maintain a development staff to be competitive
<JakeSays>
it's the nature of, well, life - do more with less
Jack22 has quit [Ping timeout: 256 seconds]
<JakeSays>
manufacturing automation is a perfect example
<JakeSays>
lf94: also average joe is not on his way out. the pool of average is constantly expanding.
califax- has joined #openscad
<lf94>
true.
<lf94>
why cant average joe incorporate if...else into his vocabulary though
<lf94>
and sphere(22
<lf94>
sphere(2)
<JakeSays>
average joe isn't even allowed to assemble cars - it's all automated
<lf94>
I wonder if there'll be an intellectual renaissance.
<JakeSays>
no
califax has quit [Ping timeout: 276 seconds]
califax- is now known as califax
<teepee>
I'd say possible, but we do seem to be on a different trajectory
<JakeSays>
yup
<teepee>
and a pretty scary one that is watching the news the last couple of month
<teepee>
not so much for single people or specific groups, but as a whole
<JakeSays>
right
<JakeSays>
i honestly believe the wealth gap will become the single greatest issue. it will overshadow racism, etc
<lf94>
Agreed
<lf94>
I think "universal income" will be necessary to keep people happy
<lf94>
Eventually not even this will be enough
<JakeSays>
*keep people alive
<lf94>
yea.
<JakeSays>
for my generation the idea of a universal income is absurd, but it's going to be necessary
lostapathy has quit [Read error: Connection reset by peer]
gunnbr has joined #openscad
gunnbr_ has joined #openscad
gunnbr has quit [Ping timeout: 240 seconds]
<lf94>
Anyone have anything which can just take points and draw the lines quickly?
<teepee>
a pen?
<teepee>
or in other words you may need to be a bit more specific to get some useful answer ;-)
<teepee>
there's no mind reading plugin for irc yet (which is good)
<lf94>
Actually you're right I can use my graph paper
<lf94>
Heh, they're points in a textual format of [[x,y],...] X)
<InPhase>
lf94: Yeah, I agree with that game theory principle. That's basically my reasoning. If it can do substantially more in a more efficient manner, that approach eventually wins.
<InPhase>
lf94: On the points, matplotlib
<InPhase>
Once you get the syntax down, you can make basic plots from points in seconds.
<JakeSays>
Scopeuk: so turns out i ordered the wrong size sticks. i have .28" instead of the thicker ones. not sure if i like them as much as the thicker ones
<JakeSays>
they are perfect for ice sickles though
othx has quit [Ping timeout: 256 seconds]
othx has joined #openscad
<Scopeuk>
that can happen
<JakeSays>
"read twice, order once"
<Scopeuk>
mine are the thinner size I'm use to
<JakeSays>
Scopeuk: how well does your clock in daylight?
<JakeSays>
*work
<Scopeuk>
its readable, its not stunning
<JakeSays>
oh they are? they looked fatter
<JakeSays>
*thicker
<Scopeuk>
7.5mm
<JakeSays>
ah ok. cool.
<JakeSays>
well then i accidentally ordered the right ones. lol
miique has joined #openscad
<JakeSays>
Scopeuk: i just discovered that if you heat the sticks slightly with a hot air gun they become clearer
<InPhase>
If you pick out certain lignin extracts, it can add some flexibility and thus a little toughness on impact to the filament.
<InPhase>
Wood of course has some good toughness on impact, so there are definitely things that would be nice to copy about its components for printing purposes. lignin, cellulose fibers, and so on seem like obvious choices, mixed with other binding agents like PLA to assist the thermal printing process.