00:18 <chadmed> is there a specific reason for targetting GLES other than theres just less stuff to implement to get a pass w/ khronos?

00:41 <alyssa> chadmed: Less stuff, and what /is/ there is a lot more ... sane .. reasonable ... testable ...

00:42 <alyssa> Desktop OpenGL has a metric tonne of obscure things nobody actually uses and are broken in a lot of real drivers but are still specced

00:42 <alyssa> the Arm drivers in Mesa generally take the policy of "aim for conformant GLES, which gives good enough GL"

00:43 <alyssa> (and support the few actually useful GL-only things, e.g alpha testing)

01:54 <chadmed> alyssa: yeah that makes sense, i'd imagine theres not much point in getting anything else working right now anyway and things like full vk support at this stage would just be wasted time

02:28 <alyssa> yeah

02:29 <alyssa> relatedly, GLES2 is the /smallest/ useful graphics API

02:29 <alyssa> but conformant GLES2 alone is enough for running an accelerated desktop and some WebGL demoes

02:31 <chadmed> yeah i mean even so, its a massive massive feat to have nearly full GLES2 support starting from nothing in only a few months, and for what 90% of people are gonna want from these machines its more than enough

02:32 <alyssa> not my frist rodeo 🤷

02:35 <krbtgt> at least you don't have to implement PHIGS

02:36 <alyssa> krbtgt: I had to Wikipedia that. Oh boy.

02:37 <alyssa> > An ad hoc ANSI committee was formed around these proposed extensions to PHIGS, changing its name to the more descriptive and (optimistically) extensible name "PHIGS PLUS" -- "PLUS" being a slightly tongue-in-cheek acronym for "Plus Lumière Und Surfaces" (the two major areas of advancement over the base PHIGS standard).

02:37 <alyssa> 😆

02:38 <chadmed> now imagine doing so with old nvidia quadrilateral shaders

02:38 <krbtgt> as someone who studied ancient graphics crap as a hobby, it is ludicrously baroque

02:39 <krbtgt> i looked into SGI GPUs once with a friend, that's an interesting rabbit hole

02:39 <krbtgt> they basically applied every combination in the early days

02:40 <krbtgt> like, various hardcoded µcode command processors, 68k based ones, etc., as well as custom chips for geometry, DSPs, actually using dedicated FPU chips in parallel (thing like x87), etc.

02:40 <krbtgt> actually kind of a mess, lol

02:40 <chadmed> even with hindsight, what engineers and developers managed to achieve in such a no mans land with so many hard limitations is still extremely impressive to me. on some levels, even more impressive than a lot of software and hardware development happening even now

02:41 <krbtgt> i have a voodoo! i got it for 20$ which is a steal compared to the insane prices you see on like ebay

02:41 <krbtgt> because there's an amiga like cult around them. oh, and glide

02:41 <chadmed> outboard were once the norm. intel only started integrating real floating point hardware with the pentiums iirc

02:41 <chadmed> one of the reasons the fast inverse sqrt function exists

02:41 <Yuzu> 486, not pentium

02:41 <chadmed> ah right

02:42 <krbtgt> 486/68040 integrated it... the 487 was basically another 486DX to replace your existing 486SX

02:42 <krbtgt> kind of funny

02:42 <krbtgt> also as someone who fixed emulation bugs in 68k FPUs they are hell and motorola just kept adding to the tire fire

02:42 <krbtgt> never let the amiga fanboys tell you differently, 68k sucks

02:43 <chadmed> 68k around then reminds me of whats going on with x86 now only marginally less bad