[Prev][Next][Index][Thread]
ANNOUNCE: Executor 2 and beyond Policy
There has been a bit of discussion regarding what ARDI will be doing
once Executor 2 is released. For people who haven't followed all our
responses, here's an explanation of our plans for Executor 2 and
beyond.
Upgrades:
All Executor 1.x licensees will receive an Executor 2 CD-ROM
sent to them free of charge. This includes Executor/NEXTSTEP
users. The rumor that we would drop NEXTSTEP support before
Executor 2 came out is FALSE.
In addition, all Executor 1.99<x> licensees (i.e. people who
licensed Executor within the last year or two, but excluding
people who licensed Executor 1.0, 1.1 or 1.2) will also get
whatever free upgrades that we offer to Executor 2 licensees
during the first two months that Executor 2 is released. We
don't yet know what our upgrade policy will be for Executor 2,
but whatever it is, 1.99<x> licensees will automatically get
it.
Development:
All the material in this ("Development:") section is subject
to change. ARDI is currently a small company but we do not
plan to remain small. Some of what is described here will
require more engineers than ARDI currently employees. We're
assuming that we'll sell more copies of Executor 2 when it's a
finished product than we've been selling to date, although
that is not a given. If we can't support ourselves through
Executor 2 sales we'll seek funding elsewhere, but that could
cause some delays.
After Executor 2 ships, ARDI will spend approximately one
month restructuring our existing code. Executor is much
cleaner than most commercial software I've worked on (I was a
freelance software contractor in the early days of Executor as
well as before then), but to be maximally efficient when we
hire new engineers we need to clean up and restructure our
code. This is no big deal, but it will allow us ARDI
engineers to catch our breath and even do some brainstorming
before going back to the implementation/bug-fix grind.
Beyond the restructuring we'll be working in many different
areas in parallel (in alphabetical order):
Compatibility with apps -- Many programs still die for mysterious
reasons. We'll continue to look into and solve mystery crashes.
Compatibility with extensions -- Our goal will be to support
INITs and CDEVs, *especially* ATM and QuickTime. Some of the
work done in this area will be the beginnings of support for
System 7.5 drop-in capability.
Intermachine communication -- Support for serial ports and
networking will be added. We plan to leverage off of existing
OS support to do this, instead of writing our own low-level
drivers.
Multitasking -- We plan to add both support for multi-tasking
within a single address space, at least enough to allow Word 6
and Excel 5 to run and also piggy-back on top of existing OS
support for multi-tasking under separate address spaces
(i.e. set things up so that when one app dies others will
continue to run).
Ports to other OSes -- Tentatively it looks like we will port
Executor to Win32 (for Windows '95 and Windows NT) and OS/2.
PowerPC emulation -- Mat has already done significant work on
a successor to Syn68k which will allow us to emulate the PPC
as well as the 68k and do so faster than Syn68k does. The
successor is internally known as VCPU and it should also allow
us to port to non-x86 architectures more easily so that if
we're ever financially motivated we could (re)port Executor to
the Alpha or take advantage of the "native instruction set"
the P7 is rumored to sport.
Working with ISVs -- A few Macintosh Independent Software
Vendors have expressed interest in using a bundled
"wrapperized" Executor to allow them to sell their Mac apps in
the Windows world. We will pursue this to the degree that the
ISVs are willing to pay for the wrapperization work. With
luck we can get extra funding for core Executor work from
royalties coming from bundled versions of Executor.
Supported OSes:
All the material in this ("Supported OSes:") section is subject
to change for the same reasons why the material in the
previous paragraph is subject to change.
NEXTSTEP has not been a viable operating system for
Independent Software Vendors for a long time now. AppSoft,
ITS, Millenium, Pages, RightBrain, Xanthus and many other ISVs
have all bowed out of the NEXTSTEP market. We will release
Executor 2 for NEXTSTEP but that very well may be our last
major release for NEXTSTEP. Everything I know about the
NEXTSTEP marketplace suggests that it's time for us to leave,
but that decision won't be officially made until after
Executor 2 has been released so that the remaining powers in
the NEXTSTEP community can see what we've done so far and see
where we're going and point out market possibilities that we
may have missed.
After Linux 2.0 is released we may consider dropping a.out
support in Executor. Clearly a.out is on its way out although
we don't know whether we'll need to continue supporting it for
a year or five years. I believe support for Linux ELF is
being worked into NetBSD, FreeBSD and BSDI, but if I'm
mistaken and they remain only able to run Linux a.out files
then we won't drop a.out support. Unlike the NEXTSTEP
situation where we're likely to drop support for NEXTSTEP
soon, we're *not* planning on doing that with a.out (Linux 2.0
is still a ways away, I believe). I am mentioning a.out
support here because it is something that will probably be
dropped at some point.
DOS support will continue after Executor 2 is released and
Executor/DOS will probably directly benefit from our
"Compatibility with apps", "PowerPC emulation" and some of our
"Compatibility with extensions", "Intermachine communication"
and "Multitasking" work. However, some of the work in these
areas is less likely to appear in Executor/DOS due to DOS's
limitations. QuickTime support, Networking support, and
support for Multi-tasking in separate address spaces are
unlikely to be incorporated to Executor/DOS; DOS is just too
primitive to easily support those features.
In all liklihood our major development efforts will be put in
Executor/Win32 (that's where the money appears to be) and
Executor/Linux (that's the OS we like). On the other hand, if
Executor/OS-2 winds up being commercially successful we
definitely won't rob its revenue stream; we'll pump some
fraction of the profits into OS/2 specific enhancements and no
fraction of the OS/2 profits into Win32 or Linux specific
enhancements. Most of the profits from all versions will go
into core work because Executor is > 90% core.
Reiterating one more time, to make a point, all the information above
except for our Upgrade policy is subject to change. Anyone who has
tracked ARDI for a considerable amount of time will have noticed that
we *haven't* been able to meet the timetables that we've set for
ourselves in the past. Almost all of this has been due to a lack of
money -- it was *very* hard to get where we were with the budget we
had.
When the bottom fell out of the NEXTSTEP market, it hit us hard. When
we thought we had lined up external financing and decided to rely on
that rather than Executor 1.x sales to fund Executor 2, only to find
out that the company that was to fund us wasn't serious, that too hit
us hard.
Hopefully after Executor 2 ships we will have a larger budget to work
with, either due to Executor 2 sales or due to the ability to acquire
financing based on Executor 2's promise. Although we haven't been
able to do things in the timeframe that we've tried in the past, we
*have* done most of the things we claimed we would (color, sound, our
own Finder, dramatic speed improvements, increased compatibility), so
there really is a good reason to believe we can do the above, at least
if the market says we should have money.
We're very honest at ARDI. I am serious when I say that all the major
decisions will officially be made after Executor 2 ships. For that
reason I suggest that anyone wanting to provide us with suggestions
for our post Executor 2 activity should wait until *AFTER* Executor 2
is shipping. We'll be much less harried and much more open to
suggestion after Executor 2 is shipping.
Thank you for your interest.
--Cliff
ctm@ardi.com
[This was cross-posted to c.s.n.a. since there has been a little
traffic there concerning our declining support for E/NS. For people
not familiar with Executor, http://www.ardi.com provides more
information]