discuss@lists.openscad.org

OpenSCAD general discussion Mailing-list

View all threads

Collecting a list of good and well documented OpenSCAD libraries

TP
Torsten Paul
Sat, Jun 1, 2019 2:00 PM

I'm starting with some of the more widely known ones:

BOSL

The Belfry OpenScad Library - A library of tools, shapes,
and helpers to make OpenScad easier to use.

Link: https://github.com/revarbat/BOSL
License: BSD-2-Clause

dotSCAD

Helpful modules and functions when playing OpenSCAD

Link: https://openhome.cc/eGossip/OpenSCAD/index.html
Link: https://github.com/JustinSDK/dotSCAD
License: LGPL-3.0-only

BOLTS

BOLTS is an Open Library for Technical Specifications.

Link: https://github.com/boltsparts/BOLTS
License: LGPL-2.1-or-later ??

I'm starting with some of the more widely known ones: BOSL ---- The Belfry OpenScad Library - A library of tools, shapes, and helpers to make OpenScad easier to use. Link: https://github.com/revarbat/BOSL License: BSD-2-Clause dotSCAD ------- Helpful modules and functions when playing OpenSCAD Link: https://openhome.cc/eGossip/OpenSCAD/index.html Link: https://github.com/JustinSDK/dotSCAD License: LGPL-3.0-only BOLTS ----- BOLTS is an Open Library for Technical Specifications. Link: https://github.com/boltsparts/BOLTS License: LGPL-2.1-or-later ??
TP
Torsten Paul
Sun, Jun 2, 2019 7:45 PM

And this is how it looks on the web site:

https://www.openscad.org/libraries.html

Now hoping for some more interesting libraries to add..

ciao,
Torsten.

And this is how it looks on the web site: https://www.openscad.org/libraries.html Now hoping for some more interesting libraries to add.. ciao, Torsten.
S
shadowwynd
Sun, Jun 2, 2019 7:56 PM

tp3, apologies for hijacking but I really like this idea and the way you have
started.  Moderators, would it be possible to make this post (or something
similar) sticky so that is always on top, not lost and buried,  and at the
same time, still editable?

I have sat through all the arguments about libraries, and namespaces, and
built-in package management, and handling derivative branches and how lack
of better libraries is driving away new users, how real users should write
everything from scratch, etc.  Without adding package management or touching
OpenSCAD itself (thus letting the vitriolic fighting over purism of language
and namespace collisions and rounded polygons continue) can we use this
forum that we already have as an aid/springboard to discovering libraries
and code?  I have been using OpenSCAD for six years and routinely encounter
useful links to libraries and other code in the forums.

At the very least - there is not a single, cohesive place to find good
libraries, or even just good snippets of code (for example, I have the
occasional useful piece of code that is shareable, but not library worthy).
Some libraries are in github, some are in thingiverse, some are linked to
random dropbox or google drive accounts or other third party sites, some are
coded by gods and others by monkeys, some don't appear to work or are so
badly documented that even getting started is hard.  While most of the
information is in the forum, searching it and finding the useful information
is difficult; many users here do not have English as a first language and so
may not search for the right word, or the useful information about a library
is five pages into another discussion.

*I envision two sticky posts:  one for libraries, one for "short useful
code".  (perhaps a 3rd for useful trainings, youtube videos, etc.)  *

*Sticky Rules: *
1) No discussion / help requests.  If code becomes broken or unlinked,
there should not be a big discussion of what needs to happen to get it to
build correctly, no back-and-forth, it either works or it it doesn't (and
the mods erase it); Discussion should take place in the rest of the forums.
Keep this thread clean.

2) No theoretical codes; everything should be tested and working (perhaps
a "Tested On" version x, y,...).

3) One library or useful code per post.

4) Picture(s) of what the code snippet / library produces.  For example,
the BOSL library has no pictures in the main page (the manuals have lots),
but just a few example photos would help a new user understand why they
might want to try it out.

5) Short description, 1 paragraph or less.  In tp3's example, dotSCAD is
listed as "Helpful modules and functions when playing OpenSCAD".  I would
like another three or four sentences describing how it is "helpful", but not
another ten.

6) Link(s)

7) License(s)

--
Sent from: http://forum.openscad.org/

tp3, apologies for hijacking but I really like this idea and the way you have started. Moderators, would it be possible to make this post (or something similar) sticky so that is always on top, not lost and buried, and at the same time, still editable? I have sat through all the arguments about libraries, and namespaces, and built-in package management, and handling derivative branches and how lack of better libraries is driving away new users, how real users should write everything from scratch, etc. Without adding package management or touching OpenSCAD itself (thus letting the vitriolic fighting over purism of language and namespace collisions and rounded polygons continue) can we use this forum that we already have as an aid/springboard to discovering libraries and code? I have been using OpenSCAD for six years and routinely encounter useful links to libraries and other code in the forums. At the very least - there is not a single, cohesive place to find good libraries, or even just good snippets of code (for example, I have the occasional useful piece of code that is shareable, but not library worthy). Some libraries are in github, some are in thingiverse, some are linked to random dropbox or google drive accounts or other third party sites, some are coded by gods and others by monkeys, some don't appear to work or are so badly documented that even getting started is hard. While most of the information is in the forum, searching it and finding the useful information is difficult; many users here do not have English as a first language and so may not search for the right word, or the useful information about a library is five pages into another discussion. *I envision two sticky posts: one for libraries, one for "short useful code". (perhaps a 3rd for useful trainings, youtube videos, etc.) * *Sticky Rules: * *1) No discussion / help requests.* If code becomes broken or unlinked, there should not be a big discussion of what needs to happen to get it to build correctly, no back-and-forth, it either works or it it doesn't (and the mods erase it); Discussion should take place in the rest of the forums. Keep this thread clean. *2) No theoretical codes*; everything should be tested and working (perhaps a "Tested On" version x, y,...). *3) One library or useful code per post.* *4) Picture(s) of what the code snippet / library produces.* For example, the BOSL library has no pictures in the main page (the manuals have lots), but just a few example photos would help a new user understand why they might want to try it out. *5) Short description, 1 paragraph or less.* In tp3's example, dotSCAD is listed as "Helpful modules and functions when playing OpenSCAD". I would like another three or four sentences describing how it is "helpful", but not another ten. *6) Link(s)* *7) License(s)* -- Sent from: http://forum.openscad.org/
TP
Torsten Paul
Sun, Jun 2, 2019 9:51 PM

On 02.06.19 21:56, shadowwynd wrote:

At the very least - there is not a single, cohesive place to find good
libraries, or even just good snippets of code (for example, I have the
occasional useful piece of code that is shareable, but not library worthy).

Snippets could go on the Tips&Tricks page (if they match the CC0
requirement which I feel is very important for that kind of things).
In case it grows too big we can just split it into multiple pages
later.

https://en.wikibooks.org/wiki/OpenSCAD_User_Manual/Tips_and_Tricks

*Sticky Rules: *
1) No discussion / help requests.

Having a sticky post could be nice so it's not buried after two weeks.

But I don't know if the "no discussion" will work without lots of
moderation, there's always some discussion going on and at some point
it just gets long and cumbersome to go through all the posts.

All it needs is 2 or 3 people who care enough to just copy&paste
those links into one of the wiki pages when something interesting
gets mentioned.

E.g. the "Libraries" page on wikibooks:

https://en.wikibooks.org/wiki/OpenSCAD_User_Manual/Libraries

That makes it much easier for new people to go though and has a stable
link to pass around. Not as good as a package manager but probably
still very helpful.

4) Picture(s) of what the code snippet / library produces.  For example,
the BOSL library has no pictures in the main page (the manuals have lots),
but just a few example photos would help a new user understand why they
might want to try it out.

Yep, that bothered me too :-). The result can be seen on github:

https://github.com/openscad/image-generator

This simply uses 4 example scripts from each of the libraries and
uses OpenSCAD + ImageMagick to convert that into a simple animated
gif like:

https://www.openscad.org/images/bosl.gif

(10 Seconds delay between transitions, so be patient... better
examples welcome...)

5) Short description, 1 paragraph or less.  In tp3's example, dotSCAD is
listed as "Helpful modules and functions when playing OpenSCAD".  I would
like another three or four sentences describing how it is "helpful", but not
another ten.

True, for now it's just the subtitle posted on the library pages.
A short single paragraph introduction would be great. Maybe also
listing the main topics the library covers in case of multi-topic
ones like those 3 already mentioned.
Ideally we could ask the library owners to help with that, at least
for the bigger and more comprehensive ones where it's very useful
to have a summary of what the focus of the library is.

ciao,
Torsten.

On 02.06.19 21:56, shadowwynd wrote: > At the very least - there is not a single, cohesive place to find good > libraries, or even just good snippets of code (for example, I have the > occasional useful piece of code that is shareable, but not library worthy). Snippets could go on the Tips&Tricks page (if they match the CC0 requirement which I feel is very important for that kind of things). In case it grows too big we can just split it into multiple pages later. https://en.wikibooks.org/wiki/OpenSCAD_User_Manual/Tips_and_Tricks > *Sticky Rules: * > *1) No discussion / help requests.* Having a sticky post could be nice so it's not buried after two weeks. But I don't know if the "no discussion" will work without lots of moderation, there's always some discussion going on and at some point it just gets long and cumbersome to go through all the posts. All it needs is 2 or 3 people who care enough to just copy&paste those links into one of the wiki pages when something interesting gets mentioned. E.g. the "Libraries" page on wikibooks: https://en.wikibooks.org/wiki/OpenSCAD_User_Manual/Libraries That makes it much easier for new people to go though and has a stable link to pass around. Not as good as a package manager but probably still very helpful. > *4) Picture(s) of what the code snippet / library produces.* For example, > the BOSL library has no pictures in the main page (the manuals have lots), > but just a few example photos would help a new user understand why they > might want to try it out. Yep, that bothered me too :-). The result can be seen on github: https://github.com/openscad/image-generator This simply uses 4 example scripts from each of the libraries and uses OpenSCAD + ImageMagick to convert that into a simple animated gif like: https://www.openscad.org/images/bosl.gif (10 Seconds delay between transitions, so be patient... better examples welcome...) > *5) Short description, 1 paragraph or less.* In tp3's example, dotSCAD is > listed as "Helpful modules and functions when playing OpenSCAD". I would > like another three or four sentences describing how it is "helpful", but not > another ten. True, for now it's just the subtitle posted on the library pages. A short single paragraph introduction would be great. Maybe also listing the main topics the library covers in case of multi-topic ones like those 3 already mentioned. Ideally we could ask the library owners to help with that, at least for the bigger and more comprehensive ones where it's very useful to have a summary of what the focus of the library is. ciao, Torsten.
A
adrianv
Sun, Jun 2, 2019 10:30 PM

shadowwynd wrote

4) Picture(s) of what the code snippet / library produces.  For example,
the BOSL library has no pictures in the main page (the manuals have lots),
but just a few example photos would help a new user understand why they
might want to try it out.

5) Short description, 1 paragraph or less.  In tp3's example, dotSCAD is
listed as "Helpful modules and functions when playing OpenSCAD".  I would
like another three or four sentences describing how it is "helpful", but
not
another ten.

I can divide libraries for OpenSCAD into perhaps three basic categories.
One type of library is that which provides some standard part, gears,
screws, other sorts of specific things that one might want to model.  For
this type of library, perhaps a few pictures can identify what the library
does.  So for example there is a knurling library on thingiverse.  One
picture suffices.  For screws and gears just a few pictures capture the
essence of the matter.  A second category are libraries which provide some
new but more general CAD function.  Perhaps a library for making rounded
cubes, or other roundings such as the fillet library, or the sweep() and
skin() modules from list-comprehension-demos and re-implemented elsewhere by
everybody in typical wheel-reinvention fashion.  The round_corners and
rounded_extrude functions I wrote would also be in this category.  Beziers.
Then the final category are libraries that extend the underlying
computations in OpenSCAD.  Things like this can't be documented in pictures:
computation of line intersection, determinant, sinh, cosh, sorting,
functions for converting lists in various ways.  These are the basic utility
functions that make it easier to write the other stuff.

The dotSCAD library is about 4000 lines of code; BOSL is about 16000 lines
of code.  I feel like it's pretty hard to capture what dotSCAD does in 4
sentences and probably hopeless to describe BOSL usefully in 4 sentences.
I'm not sure what the answer is for how to describe libraries in a way that
is truly useful.

What makes a snippet of code useful, shareable, but not library worthy?  I
would suggest that the main difference I would imagine is the code interface
and generality, but you might be treating yourself too harshly.  I mean, if
you find a particular capability useful, other people probably would too.
If the snippet isn't library-worthy by itself it probably is could be
developed into a function suitable for a library.  Ideally people can read
the libraries to learn code coding practice.  I think if you want to
introduce short code snippets in a useful way some sort of tutorial would
make more sense, though I'm not sure how to organize and write it.

I searched around for libraries, trying to figure out what capabilities I
thought a good general OpenSCAD library should have.  I was personally
specifically trying to figure out of BOSL had any "holes", so I could
contribute "plugs" to BOSL, since this seemed to me like the best way
forward (as opposed to trying to use a disparate set of unmaintained (?)
libraries).

In this hunt I found quite a few small libraries that in many cases did
interesting things, but rarely was anything well documented.  It's not
really clear whether things are maintained or not.  Most of the things I
found did not show signs of recent development.  Unfortunately, I didn't
keep a list of everything.  Mainly I did not find a lot of things that
seemed interesting.  When I found dotSCAD it appeared to be abandoned,
though we have seen a recent update, so it is not.  That is probably the
best and also most broad library I encountered after BOSL.  There is
list-comprehension-demos, which has the sweep and skin functions.  A
capability that I think looks very important is some kind of
attachment/relative positioning, since in my design experience I feel that I
spend a bunch of time fiddling with this (trial and error) instead of having
things just show up where I intend.  (Maybe others have better visualization
skills?)  This is provided by obiscad (which appears unmaintained, but is
documented), and also the relativity library which also looks abandoned, but
has documentation.  (And BOSL2 has implemented an attachment capability as
well.)

There is this list:  https://github.com/openscad/openscad/wiki/Libraries

I don't think my searching was particularly thorough, as I keep stumbling
upon other things.  Especially thingiverse is difficult to search.  (If you
search for openscad you find 11400 hits, but of course most are just things,
not libraries---it's difficult to imagine a worse system for disseminating
libraries.)  It seems like I mainly find out about interesting new stuff on
thingiverse by seeing someone report it on the forum.  Someone mentioned
puzzlecut a while ago, which seems interesting, for example.  I was just
trying to find the fillet library (which is listed at the above site) and
looked on thingiverse and instead found a bunch of other fillet libraries.

https://www.thingiverse.com/thing:35834
https://www.thingiverse.com/thing:32122

Things like this don't seem in the same category as dotSCAD or BOSL.
Perhaps it makes sense to have a list of "one trick" libraries vs
multi-capability librarires.  The former can have brief descriptions and
the latter can have long descriptions.

I think another thing which is a pretty big annoyance is suppose I want to
make gears, as the recent poster.  So I look for libraries to make gears.
What can I find?  Well, so there's MCAD.  There's BOSL.    Then there is
more stuff to be found in thingiverse and perhaps elsewhere.  So how do I
decide which gear library to use?  It seems that there are some basic
possibilities:  library A is simply better than B.  It provides more
function, is more correct, or some such.  Another possibility is that the
libraries provide different capabilities, where A provides gears of one type
and B another.  And so both libraries serve a purpose.  (One might ask if
they can be merged.)  And the last possibility is that the libraries are
equivalent, but have different interfaces, and perhaps no compelling
argument can be made that one is superior.  But for a user, it is annoying
to be presented with this decision.  It is better if there is just ONE gear
library, or perhaps two but with a clear way to pick which one to use.  I
suppose this gets back to probably of consensus.  But having lots of choices
is not necessarily an asset, especially to the new person who just wants to
make some gears.  And then when that person tries to share code the
recipient must download yet another gear library, even if he already has
four of them at hand.

So it appears that if we are listing libraries and we list ones that do the
same thing, we need to provide some hint to the user about what the
difference is---how to pick, or which one to try first.  And this does seem
ultimately to mean making a judgement call about which libraries are better.

--
Sent from: http://forum.openscad.org/

shadowwynd wrote > *4) Picture(s) of what the code snippet / library produces.* For example, > the BOSL library has no pictures in the main page (the manuals have lots), > but just a few example photos would help a new user understand why they > might want to try it out. > > *5) Short description, 1 paragraph or less.* In tp3's example, dotSCAD is > listed as "Helpful modules and functions when playing OpenSCAD". I would > like another three or four sentences describing how it is "helpful", but > not > another ten. I can divide libraries for OpenSCAD into perhaps three basic categories. One type of library is that which provides some standard part, gears, screws, other sorts of specific things that one might want to model. For this type of library, perhaps a few pictures can identify what the library does. So for example there is a knurling library on thingiverse. One picture suffices. For screws and gears just a few pictures capture the essence of the matter. A second category are libraries which provide some new but more general CAD function. Perhaps a library for making rounded cubes, or other roundings such as the fillet library, or the sweep() and skin() modules from list-comprehension-demos and re-implemented elsewhere by everybody in typical wheel-reinvention fashion. The round_corners and rounded_extrude functions I wrote would also be in this category. Beziers. Then the final category are libraries that extend the underlying computations in OpenSCAD. Things like this can't be documented in pictures: computation of line intersection, determinant, sinh, cosh, sorting, functions for converting lists in various ways. These are the basic utility functions that make it easier to write the other stuff. The dotSCAD library is about 4000 lines of code; BOSL is about 16000 lines of code. I feel like it's pretty hard to capture what dotSCAD does in 4 sentences and probably hopeless to describe BOSL usefully in 4 sentences. I'm not sure what the answer is for how to describe libraries in a way that is truly useful. What makes a snippet of code useful, shareable, but not library worthy? I would suggest that the main difference I would imagine is the code interface and generality, but you might be treating yourself too harshly. I mean, if you find a particular capability useful, other people probably would too. If the snippet isn't library-worthy by itself it probably is could be developed into a function suitable for a library. Ideally people can read the libraries to learn code coding practice. I think if you want to introduce short code snippets in a useful way some sort of tutorial would make more sense, though I'm not sure how to organize and write it. I searched around for libraries, trying to figure out what capabilities I thought a good general OpenSCAD library should have. I was personally specifically trying to figure out of BOSL had any "holes", so I could contribute "plugs" to BOSL, since this seemed to me like the best way forward (as opposed to trying to use a disparate set of unmaintained (?) libraries). In this hunt I found quite a few small libraries that in many cases did interesting things, but rarely was anything well documented. It's not really clear whether things are maintained or not. Most of the things I found did not show signs of recent development. Unfortunately, I didn't keep a list of everything. Mainly I did not find a lot of things that seemed interesting. When I found dotSCAD it appeared to be abandoned, though we have seen a recent update, so it is not. That is probably the best and also most broad library I encountered after BOSL. There is list-comprehension-demos, which has the sweep and skin functions. A capability that I think looks very important is some kind of attachment/relative positioning, since in my design experience I feel that I spend a bunch of time fiddling with this (trial and error) instead of having things just show up where I intend. (Maybe others have better visualization skills?) This is provided by obiscad (which appears unmaintained, but is documented), and also the relativity library which also looks abandoned, but has documentation. (And BOSL2 has implemented an attachment capability as well.) There is this list: https://github.com/openscad/openscad/wiki/Libraries I don't think my searching was particularly thorough, as I keep stumbling upon other things. Especially thingiverse is difficult to search. (If you search for openscad you find 11400 hits, but of course most are just things, not libraries---it's difficult to imagine a worse system for disseminating libraries.) It seems like I mainly find out about interesting new stuff on thingiverse by seeing someone report it on the forum. Someone mentioned puzzlecut a while ago, which seems interesting, for example. I was just trying to find the fillet library (which is listed at the above site) and looked on thingiverse and instead found a bunch of other fillet libraries. https://www.thingiverse.com/thing:35834 https://www.thingiverse.com/thing:32122 Things like this don't seem in the same category as dotSCAD or BOSL. Perhaps it makes sense to have a list of "one trick" libraries vs multi-capability librarires. The former can have brief descriptions and the latter can have long descriptions. I think another thing which is a pretty big annoyance is suppose I want to make gears, as the recent poster. So I look for libraries to make gears. What can I find? Well, so there's MCAD. There's BOSL. Then there is more stuff to be found in thingiverse and perhaps elsewhere. So how do I decide which gear library to use? It seems that there are some basic possibilities: library A is simply better than B. It provides more function, is more correct, or some such. Another possibility is that the libraries provide different capabilities, where A provides gears of one type and B another. And so both libraries serve a purpose. (One might ask if they can be merged.) And the last possibility is that the libraries are equivalent, but have different interfaces, and perhaps no compelling argument can be made that one is superior. But for a user, it is annoying to be presented with this decision. It is better if there is just ONE gear library, or perhaps two but with a clear way to pick which one to use. I suppose this gets back to probably of consensus. But having lots of choices is not necessarily an asset, especially to the new person who just wants to make some gears. And then when that person tries to share code the recipient must download yet another gear library, even if he already has four of them at hand. So it appears that if we are listing libraries and we list ones that do the same thing, we need to provide some hint to the user about what the difference is---how to pick, or which one to try first. And this does seem ultimately to mean making a judgement call about which libraries are better. -- Sent from: http://forum.openscad.org/
A
adrianv
Sun, Jun 2, 2019 10:37 PM

tp3 wrote

4) Picture(s) of what the code snippet / library produces.  For
example,
the BOSL library has no pictures in the main page (the manuals have
lots),
but just a few example photos would help a new user understand why they
might want to try it out.

Yep, that bothered me too :-). The result can be seen on github:

https://github.com/openscad/image-generator

This simply uses 4 example scripts from each of the libraries and
uses OpenSCAD + ImageMagick to convert that into a simple animated
gif like:

https://www.openscad.org/images/bosl.gif

(10 Seconds delay between transitions, so be patient... better
examples welcome...)

The video form doesn't seem very good to me.  I would rather see 50 images
in a web page that I could quickly scan and then say HA!  That's kind of
like the thing I want to make!  Could it make more sense to have libraries
have a link to one page per library where that page has as many examples as
are appropriate for the library?  For some libraries the page could be very
short with just a couple examples, and for others longer, with many
examples.

This doesn't address the question about how to explain libraries whose
function is not visual, though.

--
Sent from: http://forum.openscad.org/

tp3 wrote >> *4) Picture(s) of what the code snippet / library produces.* For >> example, >> the BOSL library has no pictures in the main page (the manuals have >> lots), >> but just a few example photos would help a new user understand why they >> might want to try it out. > > Yep, that bothered me too :-). The result can be seen on github: > > https://github.com/openscad/image-generator > > This simply uses 4 example scripts from each of the libraries and > uses OpenSCAD + ImageMagick to convert that into a simple animated > gif like: > > https://www.openscad.org/images/bosl.gif > > (10 Seconds delay between transitions, so be patient... better > examples welcome...) The video form doesn't seem very good to me. I would rather see 50 images in a web page that I could quickly scan and then say HA! That's kind of like the thing I want to make! Could it make more sense to have libraries have a link to one page per library where that page has as many examples as are appropriate for the library? For some libraries the page could be very short with just a couple examples, and for others longer, with many examples. This doesn't address the question about how to explain libraries whose function is not visual, though. -- Sent from: http://forum.openscad.org/
TP
Torsten Paul
Sun, Jun 2, 2019 10:57 PM

On 03.06.19 00:37, adrianv wrote:

The video form doesn't seem very good to me.  I would rather
see 50 images in a web page that I could quickly scan and
then say HA!  That's kind of like the thing I want to make!

That would be a nice thing to see, but it's something which
belongs on the page of the library, not on an index at the
OpenSCAD site (or later a package manager listing). Also it's
certainly not something I'm going to maintain for multiple
libraries.

ciao,
Torsten.

On 03.06.19 00:37, adrianv wrote: > The video form doesn't seem very good to me. I would rather > see 50 images in a web page that I could quickly scan and > then say HA! That's kind of like the thing I want to make! That would be a nice thing to see, but it's something which belongs on the page of the library, not on an index at the OpenSCAD site (or later a package manager listing). Also it's certainly not something I'm going to maintain for multiple libraries. ciao, Torsten.
TP
Torsten Paul
Mon, Jun 3, 2019 12:00 AM

On 03.06.19 00:30, adrianv wrote:

It is better if there is just ONE gear library, or perhaps
two but with a clear way to pick which one to use.  I suppose
this gets back to probably of consensus.

Maybe it's better in theory, I don't know. In reality it's
just not going to happen for likely a wide variety of reasons
(want a different interface, don't agree on hosting provider,
just want to try new things, did not look for existing ones,
tabs vs. spaces ;-), ...).

There's one of those reasons that worries me though, and
that is something along the lines: "Oh, you can have libraries?"
or "Oh, I did not know that a gear library even existed!".

People should at least be aware that there are libraries
out there and have a chance to look for themselves. So
far the OpenSCAD web site did not do a good job at that.

So it appears that if we are listing libraries and we list
ones that do the same thing, we need to provide some hint
to the user about what the difference is---how to pick, or> which one to try first.  And this does seem ultimately to
mean making a judgement call about which libraries are better.

Yes, so maybe one of the plus points for adding a library
to the list would be the report by someone "I've used that
in a project of mine and it was great" vs. just "I found
that link in a 5 years old reddit thread.

But there's no need to like vote for a final list that will
have to stay for the next 10 years. We can see how to improve
and change as we go.

ciao,
Torsten.

On 03.06.19 00:30, adrianv wrote: > It is better if there is just ONE gear library, or perhaps > two but with a clear way to pick which one to use. I suppose > this gets back to probably of consensus. Maybe it's better in theory, I don't know. In reality it's just not going to happen for likely a wide variety of reasons (want a different interface, don't agree on hosting provider, just want to try new things, did not look for existing ones, tabs vs. spaces ;-), ...). There's one of those reasons that worries me though, and that is something along the lines: "Oh, you can have libraries?" or "Oh, I did not know that a gear library even existed!". People should at least be aware that there are libraries out there and have a chance to look for themselves. So far the OpenSCAD web site did not do a good job at that. > So it appears that if we are listing libraries and we list > ones that do the same thing, we need to provide some hint > to the user about what the difference is---how to pick, or> which one to try first. And this does seem ultimately to > mean making a judgement call about which libraries are better. Yes, so maybe one of the plus points for adding a library to the list would be the report by someone "I've used that in a project of mine and it was great" vs. just "I found that link in a 5 years old reddit thread. But there's no need to like vote for a final list that will have to stay for the next 10 years. We can see how to improve and change as we go. ciao, Torsten.
C
caterpillar
Mon, Jun 3, 2019 12:35 AM

This thread reminds me of one reason why I like to use OpenSCAD and what
dotSCAD recently focuses on. I rewrite the short description for it:

  • Reduce the burden of 3D modeling in mathematics.

http://openhome.cc

Sent from: http://forum.openscad.org/

This thread reminds me of one reason why I like to use OpenSCAD and what dotSCAD recently focuses on. I rewrite the short description for it: - Reduce the burden of 3D modeling in mathematics. ----- http://openhome.cc -- Sent from: http://forum.openscad.org/
S
shadowwynd
Mon, Jun 3, 2019 12:59 AM

In real life, the value of a library or a museum is not necessarily the
collection, but the curation.  It is less valuable to know that there are
twelve gear libraries than to know that here is a good (not necessarily
"best", whatever that means) gear library that many of us use have used.

I like the "Libraries" link on the main site, that is helpful and makes it
more visible.

The "rules" for the sticky page are simply suggestions for discussion.

I agree that Thingiverse is a bad platform for distribution.  The search is
clunky, and the site is often very slow (victim of its own success).

In terms of dividing libraries into parts (screws, gears), geometric
functions (sweep, bezier), and math(arcsin) - there are always going to be
things that don't represent pictorially - it is easy to represent a dog or
an apple in any language, harder but doable to show "run" in a picture (uses
learned conventions like motion lines), and impossible to show "was" or
"have".

--
Sent from: http://forum.openscad.org/

In real life, the value of a library or a museum is not necessarily the collection, but the curation. It is less valuable to know that there are twelve gear libraries than to know that here is a good (not necessarily "best", whatever that means) gear library that many of us use have used. I like the "Libraries" link on the main site, that is helpful and makes it more visible. The "rules" for the sticky page are simply suggestions for discussion. I agree that Thingiverse is a bad platform for distribution. The search is clunky, and the site is often very slow (victim of its own success). In terms of dividing libraries into parts (screws, gears), geometric functions (sweep, bezier), and math(arcsin) - there are always going to be things that don't represent pictorially - it is easy to represent a dog or an apple in any language, harder but doable to show "run" in a picture (uses learned conventions like motion lines), and impossible to show "was" or "have". -- Sent from: http://forum.openscad.org/