home *** CD-ROM | disk | FTP | other *** search
Wrap
Received: from sloth.swcp.com (sloth.swcp.com [198.59.115.25]) by nacm.com (8.6.10/8.6.9) with ESMTP id MAA28780 for <executor@nacm.com>; Mon, 16 Oct 1995 12:11:24 -0700 Received: from iclone.UUCP (uucp@localhost) by sloth.swcp.com (8.6.9/8.6.9) with UUCP id NAA19882; Mon, 16 Oct 1995 13:11:23 -0600 Received: from beaut.ardi.com by mailhost with smtp (nextstep Smail3.1.29.0 #11) id m0t4utN-000YcBC; Mon, 16 Oct 95 13:08 MDT Received: by beaut.ardi.com (linux Smail3.1.28.1 #5) id m0t4utM-00002QC; Mon, 16 Oct 95 13:08 MDT Message-Id: <m0t4utM-00002QC@beaut.ardi.com> Date: Mon, 16 Oct 95 13:08 MDT From: ctm@ardi.com (Clifford T. Matthews) To: Kurt R Glaesemann <kurtg@iastate.edu> Cc: executor@nacm.com Subject: Future of ARDI in general (was Re: BeBox Port??????) In-Reply-To: <9510161811.AA15659@las3.iastate.edu> References: <9510161811.AA15659@las3.iastate.edu> Sender: owner-paper@nacm.com Precedence: bulk >>>>> "Kurt" == Kurt R Glaesemann <kurtg@iastate.edu> writes: Kurt> I think I will jump the gun (Ok, how about a whole army of Kurt> guns) and ask. Kurt> A new dual CPU PowerPC computer has popped up called the Kurt> BeBox. It has very little software since it was designed Kurt> from the ground up with no backwards compatability (see Kurt> http://www.be.com/). This would be a good market for Kurt> executor since the'll be a dearth of software in the Kurt> begining (not to mention that the box's look pretty Kurt> powerful. I could just see it now "I own a PowerPC computer Kurt> that run's mac apps, but not the native ones!!": proof that Kurt> truth can be stranger than fiction. We've read about it and we're fairly interested, even though future BeBox machines will be CHRP compliant and as such will theoretically be able to run MacOS once Copland is released ('97 at best -- we're not the only one who slips release dates of major new versions). It turns out that Mat has been working on a successor to syn68k for a while. The successor will make writing different backends much easier and should also be able to use different front ends. Additionally, the code that the successor generates will be much faster than what syn68k generates. A PPC backend has already been written for this new synthetic CPU, so we've planned on putting Executor on the PPC for a while now. However, when we get Executor to run on the PPC, we'll also support native PPC apps since adding that support should actually be fairly easy, so the "but not the native ones" disclaimer will probably not apply for any amount of time greater than three or four months, if ever. But Mat's syn68k successor can also have different front-ends, so our plan is to allow Executor to run PPC Mac apps on non-PPC machines as well. So, when 2.0 is shipping and we have some extra dollars, we'll *probably* pick up a Be Box and do an internal proof of concept port. We don't expect Be to be particularly interested in our work because eventually they'll have Copland and in the meantime, using ARDI's technology is overkill, since they could also just as easily get some clever engineers like Jim Drew, the Quix engineers, the author of ShapeShifter, or half a dozen other people/teams who could get Mac programs to run under the BeOS -- a simple task when compared to what we do. There would be a couple slight differences though. Executor would of course be running totally native on the PPC, while Copland will still have portions running under emulation and with any Apple based port of the Mac OS, Apple would play a significant role in determining the price of the ported OS, while we could set the Executor/Be price to be fairly low if Be decided they liked our work enough to bundle it with every machine they shipped. Right now people who think of SoftWindows when they think emulation technology in general, or who think of Executor 1.2 or even 1.99<x> where <x> < 'p' probably don't have a particularly good opinion of emulators. We will try very hard to change people's mind with 2.0's release. If we convince only a tiny fraction of our potential customer base that we're worth taking a chance on, then the resultant revenues will enable us to make the next release of Executor after 2.0 have at least as much improvement as you've seen between 1.99k and 1.99p3. --Cliff ctm@ardi.com