[Prev][Next][Index][Thread]
Re: OS/2 Porting
We are doing everything in our power to get Executor 2 out the door
ASAP. We can not work on a native OS/2 port until after Executor 2 is
shipping. It is true that we have slipped significantly and for this
I apologize and am working hard to see that we don't slip much more.
Today we decided that there will be no 1.99r; after 1.99q12 is
released we will finish beta 0 and begin work on beta 1.
>>>>> "Doug" == Doug Wing <doug_wing@il.us.swissbank.com> writes:
In article <9603230019.AA02168@ch1d92iwk> Doug Wing <doug_wing@il.us.swissbank.com> writes:
>> I just have to respond to all of the OS/2 people saying that
>> an OS/2 port would be the best thing for ARDI to do...
>>
>> While I like OS/2, own it (though I'm not currently running
>> it...) and think that it should be much more widespread than it
>> is. I don't agree that ARDI should put a priority to porting
>> Executor to OS/2.
Doug> If you add up the total Linux users and Nextstep users they
Doug> still total less than OS/2 users.
We do all our development under Linux, so the incremental cost of a
Linux port is nil. We have had to spend a little more time on packing
our internal Linux port, but even that was a very small amount of
time. The bug reports we receive from our Linux users have more than
made up for this amount of time already because, since Linux is our
normal development environment, bugs reported under E/L are easier to
track down and 9 times out of 10 the same bugs occur in E/L and E/NS,
anyway.
Our first commercial release was Executor for NEXTSTEP. Due to this,
a large portion of our customer base is NEXTSTEP users. When we added
color support, our NEXTSTEP port broke, so we did actually have to
spend a little bit of time reporting Executor to NEXTSTEP, but that
time was less than what a totally new port would take, since Mat,
Cotton and I are already proficient NEXTSTEP programmers.
Doug> Win95 still runs on top
Doug> of DOS, so a Win95 does not seem as critical as an OS/2
Doug> port. E/D runs on DOS as does Win95. There is only one
Doug> DOS/Win program I run and that is Executor. I had to turn
Doug> DOS support on to run E/D and will be glad to turn support
Doug> off as soon as I can.
Win 95 / Win NT users will profit from having a native port, just like
OS/2 users will profit from having a native port. We do not have the
resources to do either port yet. All future porting decisions will be
made after Executor 2 ships.
>> Unfortunately my take on situation is IBM is not willing to do
>> the right job in getting OS/2 out there, and are likely to cut
>> their losses and drop support of the system altogether.
Doug> OS/2 Merlin is currently in testing and nears completion.
Doug> This is a major upgrade to Warp and should carry a 4.0
Doug> badge. Does this appear to be dropping support? OS/2 for
Doug> the PPC seems near death, but OS/2 for x86 continues to
Doug> evolve.
I agree with Doug; I don't think IBM is planning on dropping OS/2.
>> IF IBM were to somehow give ARDI some financial incentive to
>> develop an Executor/2, I think that would be great.
Doug> How much money does Bill Gates give Ardi to devolpe E/D,
Doug> E/NT and E/Win95? How much money does Steve Jobs give Ardi
Doug> to devolpe E/NS? How much money does Linux Whatshisname
Doug> give Ardi to develope E/Linux? I am sure IBM matches their
Doug> total contributrion dollar for dollar.
When strictly limited to dollar contributions, you are correct neither
Bill, nor Steve, nor Linus has given us a penny of cash. Now please
realize that we're already planning on doing an OS/2 port and I
wouldn't have brought the subject of contributions up had you not
asked about Bill, Steve and Linus, but you did, so here's some
information:
Bill Gates, via Microsoft gave us a very large potential customer
base, the largest in the PC industry, by far. E/D can run under DOS,
Windows 3, Windows '95, Windows NT and OS/2. That is a very
significant contribution.
Steve Jobs, via NeXT, lent us equipment and provided us with with
access to NEXTSTEP engineers who could help us with our port. There
are other things that they have done that are covered under
non-disclosure statements and hence can't be mentioned. In addition,
because Executor originally didn't have a mc680x0 emulator, NeXT
provided a platform on which Executor could be sold, since the Sun3
was not viable by the time Executor was first ready for sale.
Linus, via Linux, has given us a tremendous development environment,
with free source code and access to incredibly talented engineers. I
have been programming professionally since 1978 and have frequently
found myself at mercy of Operating System bugs, misdocumentations and
non-documentations. Under Linux when I found that a particular system
call (setitimer) didn't do what I thought it should, I looked at the
kernel source and fixed it. I sent a bug report and got a
confirmation from Linus that my fix was correct and within a day or
two there was a new experimental version of Linux with my fix present
in it. Similarly when I found out that only some mouse drivers
supported a feature that Executor needed (FASYNC), I wrote the new
code myself and sent it off, again getting a nice response from Linus
and finding my meager contribution made it quickly into the next
experimental version.
We are not trying to hold IBM hostage -- you give us money or we won't
port Executor -- even though that is the industry standard practice
(Insignia gets good money up front for their ports of SoftPC and even
Apple gets money for their ports of MAE). It is true that if IBM gave
us money earmarked for an OS/2 port that we'd have that much more
money to put into such a port and we could even start the port today,
instead of waiting until E2 is shipping, but we're not holding out for
such a contribution.
>> And I do indeed agree that OS/2 would be a much more stable
>> platform for a native port of Executor than Win 95, except that
>> there are much less people using OS/2 than there are Windows,
>> Windows NT and Windows 95. It's unfortunate but true...
Doug> See my first comment.
Doug> The only reason I made the purchase of E/D (Dec 95), was
Doug> that Cliff told me that they hoped to have an OS/2 port out
Doug> around April. In Feb, Melissa told me that they hoped to
Doug> have the OS/2 port out before summer. I know that 2.0 is
Doug> the priority, but I have no intentsion of holding out for
Doug> the OS/2 port indefinetly.
I have always tried to make it clear that our targets may slip.
Again, in the software industry this is all too common. We do not
deliberately set unrealistic goals, but we are running at full
capacity, so when we do slip behind it is impossible to catch up by
coming in during the weekends to work since we're already working
during the weekends.
Believe it or not, we're looking forward to doing the OS/2 port. OS/2
has some nice features that we'd like to see incorporated into
Executor and porting is much more fun than doing the drudge work
associated with getting E2 out the door. However, *everyone* profits
by the release of E2, since it brings in money which we can spend on
more engineers, which allows us to port faster, add features faster
and fix bugs faster.
Thank you for your concern.
--Cliff
ctm@ardi.com
References: