Executor 2.0/executorv2.0.iso/pc/linux/extra/docs/maillist/text/archive.96/text0976.txt

Text File | 1996-07-25 | 16.9 KB | 349 lines

>>>>> "Joe" == Joe Ragosta <[e-mail]> writes:
In article <[e-mail]> [e-mail] (Joe Ragosta) writes:


    Joe> In article <[e-mail]>, Clifford
    Joe> T. Matthews <[e-mail]> wrote:

    >> NOTE: This is not a thread I normally read.  It's quite
    >> possible that I will miss follow-ups.
    >>
    >> >>>>> "Joe" == Joe Ragosta <[e-mail]> writes: In
    >> article <[e-mail]>
    >> [e-mail]
    Joe> (Joe Ragosta) writes:
    >>
    Joe> In article <4gjfid$[e-mail]>, [e-mail]
    Joe> (John Goerzen) wrote:
    >> >> Joe Ragosta ([e-mail]) wrote: >>
    Joe> Wrong. PCs can't emulate Macs worth beans. The best you can
    Joe> do is Executor which runs only circa 1987 version (System 6)
    Joe> Mac programs. . .
    >>  This is not correct.  Many, if not most, Mac apps that require
    >> System 7.0, but do not *require* post-System 7.0 features, run
    >> under Executor 1.99q6, the latest (available to the public)
    >> "Bleeding Edge" release of Executor.  Don't take my word for
    >> it, download Executor from http://www.ardi.com/ and try it
    >> yourself.

    Joe> Why should I bother? First, I don't waste my time with x86
    Joe> hardware.  Second, every report I've ever seen lists about 6
    Joe> or 8 apps which run under Executor. Then, there's the little
    Joe> matter of AppleShare and other features which don't work.

I encourage anyone reading this thread to download Executor to see
whose claims are correct.  My purpose in responding to your post
wasn't to try to change your mind, but to point out that you were
making claims that are not correct.  The claim was:

    "... Executor which runs only circa 1987 version (System 6)
     Mac programs. . ."

It is simple for anyone to download Executor and see that the above
quote is incorrect.  People are using Executor under Linux to play
games like Swoop, which requires System 7.  Scientists are using
NIH-Image 1.58 under Executor on PCs.  Although we only recently got
it running, Quicken 6 is already being used under Executor (again
under Linux).  Some find GraphicConverter 2.3 very handy under
Executor/DOS.

As to the "every report you've seen", I'd think you've seen enough
people who could claim that "every report I've seen says that Apple is
going out of business in a year".  Hearsay isn't particularly
compelling, but for anyone who doubts my claims, http://www.ardi.com/
has a demo version that can be downloaded.

As to only 7 or 8 apps running under Executor, you clipped the portion
of my post that said:

   "For those interested in a workable programs for your 199m get
    Wayzata Best of Macintosh Shareware for 19.95 from Tiger
    Software. It has 1500 programs all areas. I've tried about 500 and
    over 60% work on my comp..."

That was an unsolicited comment from an Executor user.  So right there
you've seen a claim that over 200 apps run under Executor.

You're right.  AppleShare doesn't work under Executor.  In fact,
currently you can't even use serial ports under Executor.  I wasn't
objecting to that.  I was specifically objecting to the incorrect
claim that I have reproduced in quotes above.

    Joe> Look Executor is a nice start, but you're going to have a
    Joe> hard time convincing people that it's a useful business
    Joe> tool. If you want to claim that it's a useful business tool,
    Joe> I'm sure you can give us sites where it is being used
    Joe> significantly?

That would violate our clients confidentiality.  Right now our license
database has about 3,500 entrys in it.  Probably about 200 of those
were given away free (users groups, press copies, etc.), the others
are paying customers.  That's not very many users, but Executor 2,
of which Executor 1.99q7 is a pre-beta version is not released yet and
our only advertising is currently in Linux Journal.  However, the
purpose of my post was to point out that some of your claims were
wrong, not to claim that we had a large number of users.

    Joe> Finally, I'm glad you're the one comparing Win3.1 to Mac
    Joe> System 6. The difference is that System 6 was phased out 4
    Joe> years ago and hardly anyone uses it. Win3.1 is still the most
    Joe> widely used OS in the PC world.  Considering that almost
    Joe> every app you want to run in the Windows world runs under
    Joe> Win3.1 and SoftWindows, I'd say that's not much of a
    Joe> limitation. Executor is severely limited--it runs only a half
    Joe> a dozen apps. Not much of a productivity machine.
    >>  Again, Joe's information is incorrect.  Although Executor does
    >> have many limitations (see below), the claim that Executor only
    >> runs half a dozen apps is ludicrious.  From
    >> ftp://ftp.ardi.com/pub/executor-faq.ascii

    Joe> Gee, that's funny. I checked your own site:
    Joe> http://www.ardi.com/AppNotes.19q and there are absolutely NO
    Joe> (that's right, zero) apps which have your "green" (or "OK for
    Joe> most users") rating. (Granted, you claim that you are waiting
    Joe> until the version reaches 2.0, but that's a pretty feeble
    Joe> excuse).  Based on you own standards (an app gets a yellow
    Joe> rating if it even opens and reads a file in its own file
    Joe> type), the number is only a couple of dozen. Of these, most
    Joe> of them are trashware.

For people who haven't looked at AppNotes, here is an excerpt:

       "The purpose of this list currently is to give you an idea of
    what we're working on and the general class that various
    applications fit into.  The possible classes are:

        Green -- thoroughly tested; no bugs that would prevent
             the vast majority of users from using this
             application

    [Currently nothing is in Green, because we will not do thorough
    testing until we've added all the 2.0 functionality, since things
     break in between experimental releases]

        Yellow -- less well tested or known bugs that prevent some
              functionality from being usable.  At a minimum, the
              application must be able to read and display its
              own document types

        Orange -- Will take mouse and keyboard events, but has enough
              major bugs to make the application unusable

    ..."

Joe is correct.  Our listings do not contain a single application in
Green.  That is a conscious decision that we won't start moving apps
into Green until we go beta (a few weeks away).  That doesn't mean
that no apps work, and anyone who would like to see for themselves can
download a copy of Executor and test.

Our bug-list is also available for public consumption.  There are over
500 open bugs and suggestions (out of a total of over 1500 reported).
The fact that we're open and share as much information as we can with
our customers and our potential customers shouldn't be held against
us.  On the Semper Fi mailing list, people have suggested that Apple
make their bug-list publicly available, but some people worry that by
allowing people to read about Mac bugs will cause people to lose
respect for the Mac.  In my opinion, that's sad.

    Joe> Even if some of the yellow applications turn into green ones
    Joe> at some point, you're claiming that the applications work
    Joe> well when you haven't even completed the testing. To me,
    Joe> _that_ is ludicrous.

That's fine.  That's an opinion.  You don't think we should make
partial testing information available to our customers and potential
customers, but we feel otherwise.  Everyone has their own opinion of
how they would run companies, from small ones like ours to large ones
like Apple.  That's fine.

However, just because we have consciously decided not to label
anything Green yet doesn't mean that there aren't applications that
aren't working.

    Joe> To make matters worse, even the apps which might be somewhat
    Joe> interesting are consistently one or two versions behind. At
    Joe> least SoftWindows runs the majority of Win3.1 current apps.

Some apps are a rev or two behind because they suddenly *require*
features that we don't support, e.g. Prince of Persia 2 *requires*
QuickTime, Microsoft Word 6 and Excel 5 require sub-process launching
in order to support OLE (there's no way to turn it off).  Other
programs aren't on the list just because we haven't expended the money
to buy the latest version yet or taken the time to request a free copy
for testing purposes.

But there's still a major difference between not having the latest
versions of all apps running and your quote of "...only circa 1987
(System 6)", which is what I objected to.

You also cavaliarly ignore the fact that Win 3.1 is not the latest
version of Windows and that SoftWindows as available to the general
public also runs apps that are consistently a version behind the
current Windows '95 apps.

    >> However, Joe neglected to mention that within Executor's
    >> limitations, an N MHz Pentium will run 68k Mac programs
    >> *faster* than an N MHz PPC601.  In fact, an entry level Pentium
    >> running Executor will run 68k code faster than the vast
    >> majority of 68k based Macs Apple sold, including most Quadras
    >> (ftp://ftp.ardi.com/pub/SynPaper explains how we get such great
    >> performance).

    Joe> ROTFL. I'll tell you what. Try running a PowerMac 7500 with
    Joe> Speed Doubler and System 7.5.3 and I doubt very much that
    Joe> you'll be even in the same ball park.

I will try to do that.  I stopped by EggHead Software last night in
order to buy a copy of Speed Doubler, but it wasn't on the shelves.  I
asked the clerk about "Speed Doubler for the Macintosh" and he said
"You mean RAM doubler, right?"  I said "No, Speed Doubler", to which
he replied "Haven't heard of it".  I pointed out that they had a place
for it, but that it was empty and I was just wondering if they had a
copy in the back room.  Ouch.

Anyway, all I could dig up on short notice was a PowerMac 8500/120 w/o
Speed Doubler.  I ran Speedometer 3.23 on it and got a CPU rating of
54.06, where our 90 MHz Pentium gets 28.83.  If you scale our P90
scores by a factor of 120/90, you get 38.44, so the PPC604 based
machine w/o Speed Double beats us by 40%.  So that leaves the question
of how much faster a PPC604 is than a PPC601 and how much faster Speed
Doubler is than Apple's 2nd generation 680x0 emulator.  If the values
are, say, 3 times and 2 times, then my claim is correct, even with
Speed Doubler.  If it's, say 2 and 2, then I'm wrong.  However, these
are back of the envelope estimates.  I'll order a copy of Speed
Doubler from MacConnection (was too busy today and I forgot about it)
and then see who I can find with a PowerMac 7500 locally.


    Joe> It's convenient that you make this claim without data. Your
    Joe> data shows that a 486DX4 75 nearly matches a Q610. Even if we
    Joe> assume that your results are accurate, my 100 MHz PowerMac
    Joe> 7500 is approximately 3-4 times faster than a Quadra 610 when
    Joe> running 68K code. Unless you want to claim that a 100 MHz
    Joe> Pentium is 3-4 times the 486DX4 75, you lose.

Can you pick up Speedometer 3.23 and run it?  I can hqx a copy and
e-mail it to you if you want.  I'd prefer to use Speedometer 3.23
rather than 4.x because we used a stopwatch and multiple iterations
under Speedometer 3.23 to verify that the reported times we were
getting were indeed real (in emulation you can make a mistake in how
you handle the clock and get get benchmarks that make incorrect
claims, so when we wrote SynPaper (//ftp:ftp.ardi.com:/pub/SynPaper)
we used a stopwatch and also ran the our tests immediately adjacent to
our reference Mac to make sure we got the numbers correct.  I know
that in at least one place Speedo 4.x reports numbers that are far too
high for Executor due to Scott Berfield using a different timing
mechanism.

    Joe> Besides, running 68K code faster is no big deal when the
    Joe> amount of code you can run is so low.

Back to those 8 programs again, heh.

*Our* customers think it's important because many of them had/have 68k
based Macs and are very pleased that they can run NIH Image, Stella
and other programs faster on their Pentiums than on their Macs.  Some
people say "The amount of code you can run under SoftWindows is no big
deal when the speed at which it runs is so slow", but hey, we're up
front about Executor's current limitations.  It's clearly not the
solution to everyone.  I just wanted to point out that the claim that
Executor only runs circa 1987 software was spurious and that although
Executor is less capable than SoftWindows when judged by percentage of
apps that run, it's more capable when judged by speed.  I'll let our
potential customers decide whether or not Executor 2 is useful for them.

    >> Different people have different needs.  Many people can make do
    >> with just one machine and no compatibility solutions, for
    >> others a Mac and SoftWindows or a Reply card make sense.
    >> However, Executor's true capabilities should be examined by
    >> anyone who needs (or just wants) to work in both the
    >> DOS/Windows (or OS/2, or Linux, or NEXTSTEP) and Mac worlds.
    >> Telling people that Executor only runs a half dozen circa

    Joe> True. I was being generous. Your own web site says that NO
    Joe> applications meet your green standards.

No, it does not say that.  It says

    "[Currently nothing is in Green, because we will not do thorough
      testing until we've added all the 2.0 functionality, since things
      break in between experimental releases]"

which is completely different.  You were not being generous, you were
being ignorant.  Executor runs hundreds of Mac applications, perhaps
thousands.  We are very frugal with our engineering dollars, so we're
not concentrating on maintaining a big database of what runs and what
doesn't until we go beta.  We have a nice bug-tracking system (gnats
-- Thanks Cygnus!) and use that to prioritize what we work on from day
to day and experimental release to experimental release.  We're about
to hit code freeze and when we do we'll then start officially putting
apps in our "Green" designation.

    >> 1987 programs initially makes Executor look bad, at least until
    >> someone downloads a fully-functional 10-minute time limited
    >> demo from http://www.ardi.com/ and discovers what Executor
    >> really can do.  Then the specious claims reflect poorly on the
    >> claimant.
    >>
    >> --Cliff [e-mail]

    Joe> Sure. But marketing hype reflects poorly on the claimant.

That can be true.  That's why I encourage everyone who is reading this
who doubts my claims to download a copy, or get a PC friend to
download a copy, or hang out on comp.emulators.mac.executor for a
while (I've added c.e.m.e. to the Newsgroups line and deleted many
others).

    Joe> I agree. You've got a good piece of technology. However, it's
    Joe> not something that is of much use to very many people.

Is Mac-n-Dos of much use to very many people?  It's carried by all the
major PC mail order catalogs.  It allows people to read and write Mac
formatted media on PCs.  Executor will have a street price that will
be just slightly more than Mac-n-DOS's.  Executor allows people to
read and write Mac formatted media and also run much Macintosh
software at speeds significantly faster than 68k based Macs.

    Joe> If you
    Joe> were to approach Amelio (I know Spindler wasn't interested),
    Joe> perhaps with Guy Kawasaki at your side, you might be able to
    Joe> interest Apple in working with you to make this a real
    Joe> product. I'm a believer that reasonable Mac emulation on a PC
    Joe> would be a big seller. Until you get System 7.5 support,
    Joe> however, it will not reach its potential.

I am glad we can agree 100%.  I do not like to be argumentative, but I
also do not want people mislead about Executor's current capabilities.
I do think that with System 7.5 support Executor will be significantly
more valuable, and we hope to have that done by the end of '96 with or
without Apple's help.  So far we've done everything using strictly
clean-room techniques, but with the revenue from Executor 2 fueling us
we should be able to afford separated clean-room and dirty-room teams
so we can apply the same legal reverse engineering strategy that
Phoenix used to clone the IBM-PC Bios.  Of course if Apple were
willing to license System 7.5 to us, the work would be much easier
although the profits would (probably) be less.

I'm serious about wanting to run Speedometer 3.23 on a PowerMac 7500.
Feel free to ping me if I forget.  Other than reporting those results,
unless new issues are raised I probably won't respond to further
follow-ups to this newsgroup because people who would like to test my
claims can do so without further commentary from me.

I've set followups to c.e.m.e. and c.s.m.a.  There are some very
passionate Executor users out there.  I hope that each of them will
avoid flaming.

    Joe> -- Regards, Joe Ragosta

    Joe> Copyright Joseph M. Ragosta, 1996. Non-exclusive, royalty
    Joe> free license to distribute this post granted to any service
    Joe> provider except Microsoft. By posting this, Microsoft agrees
    Joe> to pay $1,000 per posting.

--Cliff
[e-mail]