<GitHub>
[solvespace] whitequark commented on issue #263: > If you wanted to pull the "render data" out of solvespace, and render it yourself, I do know that there's a solvespace as a library thing. I don't know how up to date it is.... https://github.com/solvespace/solvespace/issues/263#issuecomment-315501409
traverseda_ has joined #solvespace
traverseda_lapto has quit [Ping timeout: 258 seconds]
traverseda has quit [Ping timeout: 260 seconds]
traverseda has joined #solvespace
<GitHub>
[solvespace] whitequark commented on issue #263: @traverseda So, I looked at imgui. I think *the* problem is going to be font rendering. It really should be using freetype instead of stb_truetype. Maybe we could do with a fork of imgui... https://github.com/solvespace/solvespace/issues/263#issuecomment-315506020
<GitHub>
[solvespace] whitequark commented on issue #265: This is not necessary. The limiting resource is not money (SolveSpace has one full-time programmer funded by M-Labs, though right now they are on hiatus) but time, specifically, time I spend on code review, merges and testing. I've had a lot of personal stuff going on recently and was unavailable to do much, which is why the project has stalled since April; that should change soon.... https://git
<GitHub>
[solvespace] whitequark commented on issue #263: I do not subscribe to the philosophy of open-source that essentially just ignores non-Latin non-Cyrillic scripts and hopes they go away. People using Latin scripts are a minority. Why *shouldn't* we support Arabic, especially given that we *already ship with Freetype*? Sure, also needs HarfBuzz and bidi, but those are easy to integrate. https://github.com/solvespace/solvespace/issues/263#issuecom
<GitHub>
[solvespace] whitequark commented on issue #263: I do not subscribe to the philosophy of open-source that essentially just ignores non-Latin non-Cyrillic scripts and hopes they go away. People using Latin scripts are a minority. Why *shouldn't* we support Arabic, especially given that we *already ship with Freetype*? Sure, also needs HarfBuzz and bidi, but those are easy to integrate.... https://github.com/solvespace/solvespace/issues/263#issue
<GitHub>
[solvespace] whitequark commented on issue #263: Side note: you seem to be thinking that groups are "layers" (they are not), and the biggest problem with hierarchical sketches is the UI. That's not the case. The biggest problem with hierarchical sketches is that SolveSpace is currently unable to load more than one file at once because of pervasively used global data, and it will take literally *months* of refactoring before we are even near the poi
<whitequark>
traverseda_: i'm not interested in inane ramblings of an american nerd
<whitequark>
his psychiatry writing is, at least, informed
<whitequark>
the rest is about as good as linus pauling's views on vitamin c
<whitequark>
if something, the *east* has won. i'm going to specifically target a chinese audience with solvespace at a later point
<whitequark>
i mean people already use it to teach in some uni in sz, for one
<whitequark>
that's beside the point though, because i *don't have* to pick and choose languages to support
<whitequark>
developers of freetype and harfbuzz did that for me
<traverseda_>
whitequark, how the west *was* won. It's arguing that western culture was simply the first to fall to what he terms "universalist culture". Take sushi as an example. In what way is sushi "western"?
<traverseda_>
whitequark, Alright, first, a bit of my background. I grew up dirt poor in the local ghetto, mulgrave park. I was part of the maybe 10% that was white, and there was definitly prejudice against me because of it.
<whitequark>
i'm not interested in an argument
<whitequark>
i've stated my policy as a maintainer
<traverseda_>
I've seen first hand how cultural norms can do real and concrete damage to people. Something as simple as how a family chooses to deal with mental illness has a huge effect on their quality of life, and is definitely part of their culture. It's my opinion that if a culture discourages people getting help, they need to *get that fixed*, and this is something I've been fighting my own family on for a while now
<whitequark>
it's not like i'm asking you personally to integrate harfbuzz, anyway; all the existing i18n work is done by me
<whitequark>
when i picked up solvespace it didn't even support umlauts
<traverseda_>
I know whitequark, but you've talked about how I need to have empathy for other cultures and it's pissed me off a bit. You're a university student right? I dropped out of highschool due to my family being in an incredibly shitty situation. And I'm getting pretty tired of privileged university kids who've never experienced any real poverty lecturing me about empathy. And better to rant here then in public.
<traverseda_>
And that's the thing, so damn certain that they're right that they refuse to engage with anyone else.
<whitequark>
I'm not an university student
<traverseda_>
"We are never going to agree on this" That's a pretty strong statement. It either means that you think that you're incapable of changing your mind or that I am. I think that's a pretty shitty attitude.
<whitequark>
I dropped out of university because of mental health issues and had to escape an abusive family
<whitequark>
you make a lot of assumptions.
<whitequark>
English isn't my first language either, by the way
<traverseda_>
whitequark, well, that's where I've been getting most of that kind of "I am so certain I am right I won't engage with people" stuff from lately. I've also been doing university contracts lately.
<whitequark>
I do not owe you "engagement"
<traverseda_>
But I do apologize for the assumptions. I'm just pretty pissed off about that whole set of things in particular.
<traverseda_>
whitequark, you don't owe me anything, but I think it's part of not being shitty.
<whitequark>
I think this argument is unproductive and becomes even less so as time goes on
<whitequark>
I would *much* rather discuss technical aspects of making a CAD than cultural, especially when the cultural aspects do not lead to technical problems
<traverseda_>
That's a pretty reasonable opinion. I think that culture/game-theory is probably one of the biggest parts of making an open-source project successful beyond the first handful of contributors. Culture is a technical problem.
<whitequark>
yes, I know about Meredith Patterson's perspective, and I also know that it represents a relatively small subset of the US
<whitequark>
say, none of that is relevant in Hong Kong, where I live
* whitequark
shrugs
<whitequark>
US has the problems of a decaying empire, and while I sympathize the population that was inevitable and self-inflicted, in the grand scheme of things
<whitequark>
I do not come from the same place as US fourth wave feminism, we barely even intersect
<traverseda_>
So why don't you think some languages are a pain in the ass? Japan was doing *pretty* well with computers during the early-80's/late-90's. A lot of people though they were well positioned to occupy the place intel/microsoft is now. Looking at the history, it certainly *looks* like the overhead of supporting the japanese character set was a big handicap in their ability to take over the computer/software market.
<whitequark>
oh, hanzi and kanji are a linguistically fascinating but outstandingly inefficient writing system
<whitequark>
first, that's not a reason to not support hangul, devanagari or arabic
<whitequark>
(we already support all kanji and most hanzi, I think, with no real work to be done)
<whitequark>
second, they're not a pain in the ass because other people have done the work for me.
<whitequark>
freetype knows how to work with thousands of characters efficiently. unifont has drawn them for me to put into the UI. the OS has provided an IME interface.
<whitequark>
*none* of that involved any work on my part except a bit of code for composing the font textures in the text window
<whitequark>
which would have to be written anyway
<traverseda_>
Support freetype better, hopefully get upstream support glyphs better. Certainly noble enough. Would be nice if it was available upstream.
<whitequark>
our difference is that you consider this noble (read: optional). I consider this baseline
<whitequark>
it's not rocket science, it's a few calls into a text layout engine, come on
<whitequark>
integrating imgui with stb_truetype would be a regression
<whitequark>
(also, what about opentype?)
<whitequark>
(macOS uses that for a system font I think)
<traverseda_>
I don't know how many open-source opentype fonts there are. I think you might be underestimating the work involved in getting freeType working cleanly with imgui, but I can't really say.
<traverseda_>
Each ligature glyph would be taking up texture memory, with how the freetype fork works now. But it might just be implemented in a particularly crappy way
<whitequark>
sure it will, is that an issue?
<traverseda_>
Could be. Depends on if you can dynamically add ligatures, or if you'd need to render all possible ligatures ahead of time. I think the way truetype is implemented on that fork isn't very efficient.
<traverseda_>
Might be an issue on low-end android phablets/tablets and the like.
<whitequark>
sure, dynamically adding ligatures is fine
<whitequark>
low-end android tablets will be unusable for CAD because of mesh generation time
<whitequark>
heck, i have a top-of-the-line core i7 laptop here and it still struggles with that occasionally
<whitequark>
i mean not all CAD, just parametric CAD
<whitequark>
mid to high end should be okayish
<GitHub>
[solvespace] traverseda commented on issue #263: Huh. Well that's something where our intuitions differ vastly. Imgui looks clean, minimalist, and professional. At least from where I'm standing. Hard to get concrete data on that though. https://github.com/solvespace/solvespace/issues/263#issuecomment-315541509
<GitHub>
[solvespace] whitequark commented on issue #263: I consider both SolveSpace's current UI and dear imgui clean and minimalist, and I don't think that migrating to dear imgui will either help or hurt SolveSpace in that regard.... https://github.com/solvespace/solvespace/issues/263#issuecomment-315541619
<whitequark>
I'm finally discussing the solution to the CLA matter
<whitequark>
I consider it a make-or-break question at this point, so it's not going to drag on
<GitHub>
[solvespace] whitequark commented on issue #68: So far the majority of people have been satisfied with the default bindings. I think I'll first implement the OS-native gestures, and see how it goes from there. https://github.com/solvespace/solvespace/issues/68#issuecomment-315555412
<cr1901_modern>
CLA?
<whitequark>
contributor level agreement
<whitequark>
copyright assignment basically
traverseda has quit [Ping timeout: 260 seconds]
traverseda has joined #solvespace
<wpwrak>
whitequark: whee ! ;-) not sure what "make-or-break" means in this context, but progress can only be good :)