[gmpi] Re: Decision Time: 7.1.2
- From: Steve Harris <S.W.Harris@xxxxxxxxxxxxxxx>
- To: gmpi@xxxxxxxxxxxxx
- Date: Fri, 13 Jun 2003 11:35:49 +0100
On Fri, Jun 13, 2003 at 12:58:15 -0700, Tim Hockin wrote:
> > Not ant once, I think its better to specify a profile where the
>
> What if I _WANT_ to support Float32 and Float64?
Then support both versions of the same profile.
> > Those would not interoperate (why would you want them to) and could be two
>
> You're making policy decisions in an API - bad bad bad news.
Um, since when. Good APIs are full of (good) policy decisions.
> > completly seperate SDKs, though they might only vary in the typedef
> > gmpi_sample line of gmpi.h.
>
> ACK!! Now we have two SDKs? This is getting really far from what I
> perceive as sane. Maybe I need to go have a sleep on it.
That was an extreme example, it doesn't have to be, the point is that they
are two seperate things that share a design viewpoint. Plugins written for
paltops are very unlikly to run accepaably on desktops whithout extensive
porting effort.
> We can make THAt decision (well, maybe we are actually incapable, but..).
> We SHOULD NOT make the decision as to "All x86 uses Float32 and all Arm uses
> Int24". First of all, we then need to enumerate EVERY platform we expect to
> run on. Second of all, there are x86 systems that SUCK at FPU, just as an
> example of how architecture CAN'T have anything to do with profile.
> Architecture SHOULD NOT come into this discussion.
Your misrepreseting or misunderstanding, the point was "desktop machines"
use float, noone writes realtime plugins for genuine 80x86 for x < 5,
you're splitting hairs.
A machine that cant use IEEE floats is not binary compatible with a
plausible desktop PC based DAW anyway, by definition.
> > But we dont need to support more than one _at the same time_, and there is
>
> I think you're really saying we don't need to supoprt more than one
> datatype/profile on any single platform. Is that a correct interpretation?
Your conflating profiles and datataypes, but yes. We know (well, strongly
suspect) this is true for our current target platforms, paltops and desktop
machines.
> > But in the case of mobile and desktop use there are exact fits, int24 and
> > float.
>
> And what of x86 palmtops or x86 webpads? What of other CPUs that are used
> in (for example) game systems and palmtops and other? MIPS, for example,
> traditionally has an FPU that is optional. Do we now have to decide
> the best profiles per chipset?
What x86 palmtops? Webpads are just laptops in a different form factor and
are prefectly capable of running "desktop" profile plugins. Maybe the
example name is a bad one, whatever.
All current game systems (ie. ones fast enough to run RT plugins on) have
hardware FPU support.
> I'd like to understand, and profiles didn't seem too far gone, until
> architectures came into play.
OK, ignore the architecture thing, it was an offhand comment and probably
doesn't make sense.
- Steve
----------------------------------------------------------------------
Generalized Music Plugin Interface (GMPI) public discussion list
Participation in this list is contingent upon your abiding by the
following rules: Please stay on topic. You are responsible for your own
words. Please respect your fellow subscribers. Please do not
redistribute anyone else's words without their permission.
Archive: //www.freelists.org/archives/gmpi
Email gmpi-request@xxxxxxxxxxxxx w/ subject "unsubscribe" to unsubscribe
Other related posts:
