00:06 <Guest9> still kind of confused. I have an stl imported, added areas of text with linear extrusion. I can't get them to subtract from the STL though

00:12 <Guest9> I may need someone to look at my code and explain how to do this

02:53 <InPhase> caveman: I want to note for you that Prusa's code that you linked is from the 2013 era. The simple reality is that what a lot of people are writing now for OpenSCAD tends to look a lot better than what people were writing back then. The language has advanced, the community has shared many best practices through collective accumulated experience, and things as a whole just overall seem to look better

02:53 <InPhase> now.

02:54 <InPhase> caveman: My first instinct when glancing at it is that it felt like "old code" to me.

02:54 <InPhase> And in fact, it is. :)

02:55 <InPhase> Good chunks of my oldest code don't look much different from that.

04:01 <caveman> InPhase: how much code shrinkage are we talking about?

04:24 <J23k73> caveman  i have seen people coming with 150lines of code and asked for help and i gave them 50 lines doing the same.  It is bit of a question how creative you are.

04:25 <InPhase> caveman: If you wanted a solid estimate of scope I'd have to try reimplementing some of the designs and see what they come out to with what I'd consider a good modern approach. But, not tonight. It's late.

04:27 <InPhase> caveman: Note that shorter is not always better though, but I note there are many deviations from good abstraction in that code, and it has a lot of raw sprawled out linear layout.

04:27 <J23k73> Just look at the demoscene to understand what you can do with 65k

04:29 <J23k73> and a sequence hull is less code than a polyhedron but it is not better

04:30 <InPhase> caveman: I'd second the notion that the calendar is a good illustration of compact code, and often elegant code. Although it is of a particular subset of code not using libraries. I do rely on a lot more libraries for current designs.

04:38 <caveman> my problem is that, i went to the calendar, and i saw two things: (1) stuff seeming useful, (2) stuff seeming like little tests. (1) was large, (2) was ok but useless.

04:40 <J23k73> they are demonstrations of how something can be done - so you need to apply them

04:41 <InPhase> Perhaps I can illustrate what I picture as the gradual evolution of things with two random selections of mine. I don't know if you will see what I see. Here is a design of mine from 2018, selected because it's halfway between now and 10 years ago. It works! It was a great design mechanically speaking. I kind of don't like the code, and wouldn't do it that way today. https://bpa.st/UUTSS

04:41 <J23k73> best thing you have something you need to design and talk about the approach

04:44 <InPhase> caveman: In comparison, here is a design of mine from last year. I have no lingering regrets about that design other than a subtle personal objection to the vertical rather than angular spacing that caused that extra wide second-last layer at the top of the knob curve. I might fix that if I redid it, but honestly that's super subtle and not much to do with the core structural differences.

04:44 <InPhase> https://bpa.st/32RK4 https://i.imgur.com/nF2hu2Q.png

04:45 <InPhase> caveman: Now the way I evolved in how I do things is not going to be the same as how everyone else has changed. We all have our personal code styles. But there's a trend toward an increased use of smaller functions and modules, and parameterization seems to on average have improved over the years, with fewer magic numbers.

04:46 <InPhase> caveman: People in general have more tricks for abstraction and reuse, which compactifies and simplifies things.

05:36 <J23k73> caveman you also need to develop a different mindset of thinking compared to the usual cad software - that requires some time

05:38 <caveman> i heard that openscad, due to its mesh nature, will become too slow compared to, say, freecad.

05:41 <J23k73> depends on how you create them, and the new manifold (2023) is much faster

05:42 <J23k73> but yes if you have a preview with high detail it will slow down, but you can choose to use high detail only for render

05:43 <caveman> anyone who uses openscad for making buildings? steel structures?

05:44 <J23k73> from time to time we have seen here people to use it for concrete molds, or block houses, carpentry

05:45 <J23k73> but as there is no build in function for counting parts and create lists of parts (BOM) other CAD software may be more useful

05:47 <J23k73> you get a nice design but if you need the area of your building floors you may get headache, also there is no static and you have no materials to export (or even colors)

05:47 <J23k73> so there are clear limitations

05:50 <J23k73> and viewing a whole detailed building with all hidden constructions, plumbing, electrical etc.  will be too complex - you need to make switches to simplify details. ( i assume other cad software have that build in)

05:54 <J23k73> but some design their room like this https://www.reddit.com/r/openscad/comments/13f2yx5/our_makerspace_debates_new_kitchen_so_i_made_a/

06:29 <guso78[m]> Yes, my flow IS to Print 3d busts and use Them for Green Sand Metal Casting Then

06:30 <guso78[m]> But ITS quite difficult to find Models which easily demold.

07:43 <Scopeuk> guso78[m] I use a program called "3d tool" https://www.3d-tool.com/en-cad-viewer-download.htm (just the free viewer) which can take a model and report the draft angles for you throughout to work out how well things will de mold

07:44 <Scopeuk> that's not something I use much but I'm aware it is in there

07:46 <Scopeuk> I use the measurement tools in there a lot (it is windows only and it's not open if those are deal breakers)

08:04 <guso78> Scopeuk, thank you for the hint.i  will watch that out

12:27 <guso78> the 3d tool installation files are incredible big(500 MB) . suppose they also include libfive

12:34 <Scopeuk> I thought the free viewer was smaller than that

12:35 <Scopeuk> yeh, viewer is 30 meg

12:35 <Scopeuk> the full tool does a whole load of format conversion stuff (but it is commercial software)

12:37 <Scopeuk> I think https://www.3d-tool.com/downloads/setup_3dtool_freeviewer.exe is a direct download ldink

14:01 <lf94> guso78[m]: nice :)

14:02 <lf94> guso78[m]: let's see using RNG SDF :p

14:30 <guso78> RNG SDF ? what is RNG ?

14:31 <InPhase> Random Number Generator.

14:33 <lf94> Like what I tried the other day

14:33 <lf94> Noise with sphere SDF

14:33 <lf94> Except I failed because I'm not really sure

14:33 <lf94> I think it has to do with my libfive translation of the original SDF algorithms

14:37 <guso78> lf94, but it would work better with geodesic spheres instead

16:14 <joseph_> kintel: To me it seems like a strange property of GitHub repository permissions that after I made a pull request from my fork to the OpenSCAD repo, you were able to force push to my branch. It's not a problem in this case because I still have the original authored commits locally. But as a general rule I would not have expected that making a PR implicitly gives the other repo's maintainer write access to my fork.

16:31 <lf94> it may only be write access to a particular branch?

16:40 <teepee> joseph_: yes, that's the default. I think there's a flag to disable pushes, not sure if there's something separate for force-push

16:46 <gbruno> [github] njsokalski opened issue #4639 (Blank Lines Not Visible When Module Is Collapsed) https://github.com/openscad/openscad/issues/4639

17:49 <peepsalot> joseph_: iirc the checkbox is available at the time of creating the PR. "allow maintainers to write" or something like that

18:09 <joseph_> That makes sense. I remember seeing that option before but I had forgotten about it

19:01 <guso78> tada!!! https://imgpile.com/i/9h3GYG

19:01 <guso78> lf94, ups, meant  for you

19:09 <lf94> woo

19:09 <lf94> almost a smooth union haha

19:14 <guso78> we need some sort of directed randomness to become  manifold, but starting the "smooth"version now.