which is first, Intel or Microsoft
Hans Bus
>
> In comp.arch.embedded Douglas Beattie Jr. <beat...@whidbey.net> wrote:
> > first PC disk operating system. A strong rumor says that the first
> > MS-DOS was really a pirated, and slightly revised CP/M (from Digital
>
> Are you suggesting that Seattle Computer Products didn't write Q-DOS
> (the CP/M clone) from scratch? Do you have any evidence to support this?
>
> Doug listed a URL that has a good overall chronology, which is at:
>
> http://www.maxframe.com/HISZCOMP.HTM
>
> --
> =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
> Spehro Pefhany "The Journey is the reward"
> sp...@interlog.com
> Fax:(905) 271-9838 (small micro system devt hw/sw + mfg)
> =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Did you read the page you are referring to ???
literal quote :
- QDOS v0.10 (Quick and Dirty Operating System) is shipped by Seattle
Computer Products.
Even though it had been created in only two man-months as a quick copy
of Digital
Research CP/M, QDOS worked but had many serious bugs and was an
imperfect copy of CP/M.
A week later, the EDLIN line editor was created. EDLIN was supposed to
last only six
months, before being replaced. [2]
- Seattle Computer Products renames QDOS to 86-DOS Version 0.3.
Microsoft then bought
non-exclusive rights to market 86-DOS, which was a clone of Digital
Research CP/M in
virtually every respect. [2]
As I understood it, digital research forced ibm to sell cp/m-86 with
it's pc's
as well as msdos by either suing them or threatening to sue them over
ms-dos
being a copy of cp/m.
ibm did so (unwillingly), but at 4 to 5 times the price for ms-dos. Bye
bye cp/m-86 !
There are several articles about this, unfortunately I cannot locate
them right now.
Perhaps someone in comp.os.cpm can produce (a pointer to) an article. I
searched
my cp/m links, but cannot find one.
regards,
Hans Bus
bill_h
Microsoft Z80 card that was used to put CP/M capability onto
the Apple II would have *some* serious *intimate* knowledge
of the operating systems involved?
In the book ''Hard Drive'', it says Paul Allen hired Timothy
Paterson to work on that card. And that when he was unable to
complete the project another group was assigned the task.
Since this was AT LEAST two years before the IBM/DOS/CP/M
fiasco, clearly Bill Gates was telling a major LIE when he
said, in 'Triumph Of The Nerds', that Paterson was just 'some
guy that Paul knew' who just happened to have ''written''
this operating system so Microsoft could buy it.
This episode is layer upon layer of LIES. Once we get to SEE
the actual source codes involved, hopefully we'll find out what
really went on. Or, Paterson/Microsoft will be manufacturing
a piece of evidence that can land them all in prison.
Bear in mind that Paterson didn't actually rip off CP/M by copying
OBJECT CODE, he had access to the SOURCE. That means he could read
Kildall's comments about what needed to be done, and how it was done.
This is a level of theft that's almost unprecedented, since source
code is usually carefully guarded under lock and key (except Linux!).
In 1984 I worked at a software publishing house, and one of our
conditions for handling a product was to be given the SOURCE for
whatever we were publishing. And from personal experience I can
tell you authors would HOWL at that requirement. We kept their
source code in a safe. UNTIL the company went bankrupt and everything
was just dumped into a big pile and sold at a warehouse sale.
For those interested, that was Software Strategies, Inc (SSI).
In case you're wondering what ever happened to your program(s).
Bill
Tucson
Tim Shoppa
>
> Don't you suppose a person who worked on development of the
> Microsoft Z80 card that was used to put CP/M capability onto
> the Apple II would have *some* serious *intimate* knowledge
> of the operating systems involved?
In fact, all you need to know about CP/M to do this is to read
the _CP/M 2 Customization Guide_. Everyone who bought a full-blown
copy of CP/M got this booklet.
> Bear in mind that Paterson didn't actually rip off CP/M by copying
> OBJECT CODE, he had access to the SOURCE. That means he could read
> Kildall's comments about what needed to be done, and how it was done.
Sure, there's sample BIOS source code in Appendix C of the _CP/M 2
Customization Guide_. But again, *everyone* who bought a full-blown
(as opposed to dumbed-down) official copy of CP/M got this booklet
along with the 8" distribution floppy and a license signed by
Gary Kildall.
Writing a new BIOS for CP/M isn't the hardest thing in the world -
after all, everyone who didn't get it prepackaged onto their system
had to make their own BIOS from scratch - and doing
it for a coprocessor is indeed a bit more work, but you certainly
don't need the sources to the OS and CP/M utilities to do it. All
you need to know is right there in the _Customization Guide_.
--
Tim Shoppa Email: sho...@trailing-edge.com
Trailing Edge Technology WWW: http://www.trailing-edge.com/
7328 Bradley Blvd Voice: 301-767-5917
Bethesda, MD, USA 20817 Fax: 301-767-5927
dls2
Derrick Shearer
dl...@Lehigh.EDU
Tim Shoppa <sho...@trailing-edge.com> wrote in message
news:382BE89...@trailing-edge.com...
Roger Ivie
>So, has the copyright run out on the _CP/M 2 Customization Guide_?
No, but Tim Olmstead was able to convince Caldera (the current
owner of CP/M) to release the old CP/M stuff under a free-for-personal-use
license. The documentation is available from the Unofficial CP/M Web Site,
which may be found at http://cpm.interfun.net/. The documentation
available there includes the CP/M 2 Customization Guide.
--
Roger Ivie
TeraGlobal Communications Corporation
1770 North Research Park Way Suite 100
Logan, UT 84341
mailto:ri...@teraglobal.com
phoneto:(435)787-0555
faxto:(435)787-0516
-----------== Posted via Newsfeeds.Com, Uncensored Usenet News ==----------
http://www.newsfeeds.com The Largest Usenet Servers in the World!
------== Over 73,000 Newsgroups - Including Dedicated Binaries Servers ==-----
Chip Zempel
"CP/M Operating System" manual is called "CP/M Alteration." This may be
similar to what Tim was talking about. The ownership of the manual itself
isn't clear, but the owners of the source code have posted the manual on
the web. It's at:
http://www.mathcs.emory.edu/~cfs/cpm/
Appendix A of this guide is about modifying the BIOS and Appendix B has a
"skeletal" BIOS to start from. I think this might be the manual that Tim
spoke of - I don't recall seeing a "Customization Guide" and I don't see
one at the site above - but I could be wrong. Either way, there's a lot of
info at the link above. (Lot's of the old DR source code, too!)
In article <80hjan$16...@fidoii.cc.Lehigh.EDU>, "dls2" <dl...@Lehigh.EDU> wrote:
>So, has the copyright run out on the _CP/M 2 Customization Guide_?
>
-------------------------------------------------------------------
Chip Zempel
cze...@ns.net
-------------------------------------------------------------------
If a cluttered desk is a sign of a cluttered mind,
then what does an empty desk signify?
anon...@bogus_address.con
On 1999-11-12 bil...@sunsouthwest.com said:
>Don't you suppose a person who worked on development of the
>Microsoft Z80 card that was used to put CP/M capability onto
>the Apple II would have *some* serious *intimate* knowledge
>of the operating systems involved?
>
>In the book ''Hard Drive'', it says Paul Allen hired Timothy
>Paterson to work on that card. And that when he was unable to
>complete the project another group was assigned the task.
>
>Since this was AT LEAST two years before the IBM/DOS/CP/M
>fiasco, clearly Bill Gates was telling a major LIE when he
>said, in 'Triumph Of The Nerds', that Paterson was just 'some
>guy that Paul knew' who just happened to have ''written''
>this operating system so Microsoft could buy it.
>
>This episode is layer upon layer of LIES. Once we get to SEE
>the actual source codes involved, hopefully we'll find out what
>really went on. Or, Paterson/Microsoft will be manufacturing
>a piece of evidence that can land them all in prison.
[...snip...]
>Bill
>Tucson
Well, there you go AGAIN, Bill! <g> This was all =settled= years ago,
and you know it perfectly well.
Digital Research agreed NOT to sue over the theft of CP/M source code,
in return for IBM's agreement to make CP/M-86 available as an alternative
O.S. for the IBM PC.
IBM did this, even though they broke the =spirit= of the agreement by
essentially giving away PC-DOS for 'free' with the PC...while CP/M-86
cost the potential buyer an additional $US 495.00.
Digital Research got snookered; pure and simple. It's a doggone shame,
yes...but the reality is, nobody will ever go to prison over it.
While I share your animosity toward Mikro$loth and its Il Duce, this
is =not= the path by which they'll eventually be brought to account.
Get USED to it! :)
anon...@bogus_address.con
On 1999-11-12 sho...@trailing-edge.com said:
>bill_h wrote:
> > Bear in mind that Paterson didn't actually rip off CP/M by copying
> > OBJECT CODE, he had access to the SOURCE. That means he could read
> > Kildall's comments about what needed to be done, and how it was
> > done.
>
>Sure, there's sample BIOS source code in Appendix C of the _CP/M 2
>Customization Guide_. But again, *everyone* who bought a full-blown
>(as opposed to dumbed-down) official copy of CP/M got this booklet
>along with the 8" distribution floppy and a license signed by
>Gary Kildall.
You might have misunderstood, Tim; this wasn't just BIOS source code.
Seattle Computer Products, where Tim Paterson was ensconced, was an
=authorized= Digital Research reseller/developer...and as such, had
access to the original and full CP/M source code.
Back in those days, D.R.I. made the source for CP/M available to its
authorized outlets. A serious mistake, as it turned out.
anon...@bogus_address.con
On 1999-11-12 dl...@Lehigh.EDU said:
>So, has the copyright run out on the _CP/M 2 Customization Guide_?
Nope. It has over three more decades to run.
Limey
says...
>
> Back in those days, D.R.I. made the source for CP/M available to its
> authorized outlets. A serious mistake, as it turned out.
>
Well not really - the BDOS and CCP were not large pieces of code and were
quite easy to figure out - I'm sitting here looking at some line printer
output dated 1-dec-84 that I wrote and I have another copy by Clark
Calkins that's dated 1983... this isn't Rocket Science...
The ironic thing is that if you disassembled Windows today and then tried
to resell the code then you'd probably end up in court.
It's worth remembering that prior to the introduction of MSDOS most
utilities came with source code...
Limey
Barry Watzman
Bill Gates at the time, and I bought the Seattle Computer Products hardware
computer system, which I still have, including the 8" disks of 86-DOS version 0.3,
which I also still have, as well as all of the later versions through version 2.0
(which was, by then, simply SCP's licensed copy of the by-then-Microsoft MS-DOS
2.0).
DR threatened to sue IBM, but never actually did sue, I think.
I think that MS actually bought 86-DOS itself (e.g. SOLE ownership of the actual
copyright), and not a license, non-exclusive or otherwise. However, SCP was then
itself issued a special non-exclusive license by Microsoft, a license that was, I
think, quite unique in it's terms and not duplicated by any other licensee.
Barry Watzman
Barry Watzman
source code -- and by that, I mean to the operating system itself, not a
BIOS. But, to be honest about it, CP/M is a simple, almost trivial
operating system. And EVERYONE who was writing software in those days
knew, or had access to, the data structures and API used by CP/M -- they
were not proprietary information. And any programmer of Tim's caliber
could have created a CP/M clone from scratch in a relatively short time
(weeks, perhaps a month or two). Again, here, I am talking about the
operating system itself, e.g. the CCP and BDOS. Remember, both of these
combined don't amount to 6k of code.
If you look at the CP/M operating system, MOST of the code is in the
utilities, not the actual operating system. ASM is bigger, I think, than
the CCP and BDOS. Then there is DDT, STAT, ED, PIP, MOVCPM and so on.
These obviously were not copied in 86-DOS, as the functionality is
completely different, and the processor is also totally different (e.g. the
assembler in 86-dos MUST be a very different animal than the one in CP/M
due to the differences in the 8080 and the 8086 architecture -- and this
goes for DDT and DEBUG also). So the suggestion that Tim copied the SOURCE
code, to me, does not hold much water. In fact, he didn't copy the object
code either, rather, what he did was to copy the data structures and, to
some extent, the user interface (which was neither very sophisticated nor
very original anyway).
Barry Watzman
Tim Shoppa wrote:
> bill_h wrote:
> >
> > Don't you suppose a person who worked on development of the
> > Microsoft Z80 card that was used to put CP/M capability onto
> > the Apple II would have *some* serious *intimate* knowledge
> > of the operating systems involved?
>
> In fact, all you need to know about CP/M to do this is to read
> the _CP/M 2 Customization Guide_. Everyone who bought a full-blown
> copy of CP/M got this booklet.
>
> > Bear in mind that Paterson didn't actually rip off CP/M by copying
> > OBJECT CODE, he had access to the SOURCE. That means he could read
> > Kildall's comments about what needed to be done, and how it was done.
>
> Sure, there's sample BIOS source code in Appendix C of the _CP/M 2
> Customization Guide_. But again, *everyone* who bought a full-blown
> (as opposed to dumbed-down) official copy of CP/M got this booklet
> along with the 8" distribution floppy and a license signed by
> Gary Kildall.
>
bill_h
> Well, there you go AGAIN, Bill! <g> This was all =settled= years ago,
> and you know it perfectly well.
I don't think so. This is probably wishful thinking on Microsoft's part.
> Digital Research agreed NOT to sue over the theft of CP/M source code,
> in return for IBM's agreement to make CP/M-86 available as an alternative
> O.S. for the IBM PC.
The 'agreement' was with IBM. It didn't mention Microsoft. You have
something else to offer? Kildall was well aware of the difference.
> IBM did this, even though they broke the =spirit= of the agreement by
> essentially giving away PC-DOS for 'free' with the PC...while CP/M-86
> cost the potential buyer an additional $US 495.00.
At law, this is called ''failure of consideration''. It makes the
agreement
VOIDABLE. And is generally a very stupid thing to do. And which is why I
was asking about LOGO, as sold by IBM - perhaps there was further
agreement,
to satisfy Kildall, that involved offering LOGO in the IBM livery. Well
for whatever reason he failed to do much about it, except complain.
> Digital Research got snookered; pure and simple. It's a doggone shame,
> yes...but the reality is, nobody will ever go to prison over it.
Manufacturing evidence is a potential ticket to prison.
Differences of opinion make for horse races and lawsuits. It's FAR from
over yet. MS has chosen to make many many enemies in this business.
We'll
have to wait and see what happens in a couple months.....
> While I share your animosity toward Mikro$loth and its Il Duce, this
> is =not= the path by which they'll eventually be brought to account.
Animosity is the wrong word. I don't hate sharks. I just don't want one
in my backyard pool. Gates is a natural preditor. Sixty years ago we saw
a political Hitler. Gates is an ECONOMIC Hitler. Until recently he was
more or less free to conquer through uncivilized tactics. And accumulate
wealth beyond any possible constructive use. Now he's moving into
content.
He threatens to become a force in shaping opinion and influencing
politics.
The more traditional power structure, the one bought and paid for with
OLD
money, has decided it's time to rein this upstart in.
His choices are, back down and accept an attenuated role in his
remaining
years, or try to bluff and bully his way, as he appears hell bent on
doing,
and find his company carved up and out from under him.
The bigger they are, the harder they fall.
2000 should be a very interesting year.
Allison
>You might have misunderstood, Tim; this wasn't just BIOS source code.
>Seattle Computer Products, where Tim Paterson was ensconced, was an
>=authorized= Digital Research reseller/developer...and as such, had
>access to the original and full CP/M source code.
TIM P. reverse enginerred it from a copy of CPM-1.4 by dissassembly.
The BDOS is only 3.5k of 8080 code and it's not rocket science to
dissassemble it (Try the L option in DDT!).
>Back in those days, D.R.I. made the source for CP/M available to its
>authorized outlets. A serious mistake, as it turned out.
It was available but copying is still copying and he (Tim P) used the
8080->8086 code converter on it leading to the story that Q-dos
and early PCdos had the DRI copyright statement as it was passed
through.
Allison
Real address is: Allisonp @ world DOT std DOT com
++++BULK Email severely not wanted+++
Allison
>"skeletal" BIOS to start from. I think this might be the manual that Tim
>spoke of - I don't recall seeing a "Customization Guide" and I don't see
>one at the site above - but I could be wrong. Either way, there's a lot of
>info at the link above. (Lot's of the old DR source code, too!)
You're wrong. the Alteration Guide was orignally a seperate book and
later editions it was unchanged but bound in with the rest of the set.
I have the orginal 1.4 set, 2.0 set and the kaypro, morrow and DEC
bound version of the exact same information.
Tim is right. You don't need source for the CCP or BDOS to move it
from one platform to another, just the willingness to write the BIOS
which is very well documented by DRI and Several other authors
(Andy Johnson-Laird did a really fine job). I'll go further to say
the only part of CPM you cant easily hack is the BDOS (I know about
the improved clones like P2dos and ZRdos) as the CCP is a special
transient (gets loaded by warm and coldbooters). That made ZCPR
(a common improved CCP) a easily inserted mod due to CPM modulatory.
(MS)DOS was also portable in the same way at one time. You could get
the info to write your own BIOS and that lead to the PC-DOS on S100
systems we don't hear much about anymore.
Guy Macon
>
>So, has the copyright run out on the _CP/M 2 Customization Guide_?
Here is a list of when copyrights run out.
**************************************************
DATE OF WORK: Published before 1923
PROTECTED FROM: In public domain
TERM: None
**************************************************
DATE OF WORK: Published from 1923 - 63
PROTECTED FROM: When published with notice [3]
TERM: 28 years + could be renewed for 47 years, now extended
by 20 years for a total renewal of 67 years. If not so
renewed, now in public domain
**************************************************
DATE OF WORK: Published from 1964 - 77
PROTECTED FROM: When published with notice 28 years for first term;
TERM: now automatic extension of 67 years for second term
**************************************************
DATE OF WORK: Created before 1-1-78 but not published
PROTECTED FROM: 1-1-78 (Effective date of 1976 Copyright Act)
TERM: Life + 70 years or 12-31-2002, whichever is greater
**************************************************
DATE OF WORK: Created before 1-1-78 but published
between then and 12-31-2002
PROTECTED FROM: 1-1-78, (Effective date of 1976 Copyright Act)
TERM: Life + 70 years or 12-31-2047 whichever is greater
**************************************************
DATE OF WORK: Created 1-1-78 or after
PROTECTED FROM: When work is fixed in tangible medium of expression
TERM: Life + 70 years [1] (or if work of corporate authorship, the
shorter of 95 years from publication, or 120 years from creation [2]
**************************************************
[1] Term of joint works is measured by life of the longest-lived author.
[2] Works for hire, anonymous and pseudonymous works also have this term.
17 U.S.C. § 302(c).
[3] Under the 1909 Act, works published without notice went into the
public domain upon publication. Works published without notice
between 1-1-78 and 3-1-89, effective date of the Berne Convention
Implementation Act, retained copyright only if, e.g., registration
was made within five years. 17 U.S.C. § 405.
Source: Tom Field / Lolly Gasaway. Last updated 11-5-98
Charles E. Bortle, Jr.
Barry Watzman wrote in message <382CD337...@neo.rr.com>...
>I have never seen any evidence that Tim Patterson had access to the CP/M
>source code -- and by that, I mean to the operating system itself, not a
>BIOS. But, to be honest about it, CP/M is a simple, almost trivial
>operating system. And EVERYONE who was writing software in those days
>knew, or had access to, the data structures and API used by CP/M -- they
>were not proprietary information. And any programmer of Tim's caliber
>could have created a CP/M clone from scratch in a relatively short time
>(weeks, perhaps a month or two). Again, here, I am talking about the
>operating system itself, e.g. the CCP and BDOS. Remember, both of these
>combined don't amount to 6k of code.
This was also my take on it as I have followed this thread. FWIW
--
Charles cbr...@ix.netcom.com
"For God So Loved The World, That He Gave His
Only Begotten Son, That Whosoever Believeth
In Him Should Not Perish, But Have Everlasting
Life"John3:16 * http://pw2.netcom.com/~cbrtjr/wrdthing.html *
bill_h
It's worth noting that where Seattle Computer appears to have
registered nothing, and Microsoft very very little, DRI
in fact filed over a hundred copyright registrations.
And something people generally seem to be unaware of, in this
thread, is the FACT that Microsoft had signed a licensing agreement
for CP/M that quite explicitly forbade reverse engineering and
disassembly of CP/M, which was the only means POSSIBLE for Paterson
to 'discover' the internals of CP/M so that it COULD be cloned.
Put that fact; together with the fact that Paterson worked on the
softcard in 1978 (+/-); and the fact that Paul Allen hired him for
that work; and the fact that Paterson 'developed' 86-DOS which later
became PC- and MS- DOS; there's omly so much leeway before some
pretty inescapable conclusions begin to emerge.
A close inspection of known facts, as I've been trying to point
out, make clear both Microsoft's violation of their own signed
agreement, AND the clear theft of intellectual property that
represents.
Microsoft, and Paterson in particular, have been trying to spin
this story away for almost twenty years.
It IS NOT going to go away, because it is true.
What amazes me is that not one of those people charged with 'theft'
of Microsoft products, in particular DOS and/or Windows, has (as far
as I know) raised the defense that those products are based on the
prior work of someone else, challenging MS' right to recovery on
something they don't actually own.
A skilled defendant could pick apart such a claim, and actually get the
case dismissed if MS did NOT clearly exclude certain things in their
claim. Since (again, as far as I know) MS has NEVER limited it's claims
to any part of DOS or Windows, we can assume they're not about to start.
bill_h
> Barry Watzman wrote in message <382CD337...@neo.rr.com>...
> >I have never seen any evidence that Tim Patterson had access to the CP/M
> >source code -- and by that, I mean to the operating system itself, not a
> >BIOS. But, to be honest about it, CP/M is a simple, almost trivial
> >operating system. And EVERYONE who was writing software in those days
> >knew, or had access to, the data structures and API used by CP/M -- they
> >were not proprietary information........
IBM published the source code listing for the BIOS of their machines.
Can't hardly get any more 'open and accessable' than that!
Seems to me IBM sued a bunch of folks who tried to CLONE it,
though......
You are confusing ability to purchase and use something, with the
right to copy it, including look and feel, and sell copies for profit.
Have you ever taken a real good look at the Microsoft card? And at
the documentation, including the early stuff? There was a lot more
involved than simply re-writing the CBIOS that DRI normally supplied.
And Tim Paterson was unsuccessful in the task, remember. What does that
tell you about his so-called 'abilities'? You can't have it both ways.
And you also can't credit HIM with everything included with DOS,
including
the use of the FAT, because the official MS story is that this was
something
Tim 'borrowed' from M-DOS, after being shown that operating system.
Clearly then, there WAS some direct MS involvement in the writing of
PC-DOS, and THAT would be a CLEAR VIOLATION of the written DRI
agreement.
The successful 'cloning' of the IBM BIOS required those taking the
propritary product being 'cloned' apart have absolutely nothing to
do with producing the clone product itself; they only wrote the
'specifications' for it from examining the original.
Paterson was in BOTH camps, working on BOTH sides. That's the problem.
Ed Porter
>
> ... Sixty years ago we saw a political Hitler. Gates is an ECONOMIC Hitler.
Bzzzt! Godwin's Law says this thread is over, and bill_h loses. Shame really,
I was tending to agree with him.
-ed
--
"No, `Eureka' is Greek for `This bath is too hot.'" -- Dr. Who
Guy Macon
>And something people generally seem to be unaware of, in this
>thread, is the FACT that Microsoft had signed a licensing agreement
>for CP/M that quite explicitly forbade reverse engineering and
>disassembly of CP/M, which was the only means POSSIBLE for Paterson
>to 'discover' the internals of CP/M so that it COULD be cloned.
Not so. Pheonix cloned the IBM PC BIOS with another technique
known as "cleanroom coding" that involved no reverse engineering
or disassembly by the programming team that wrote the code.
I offer no opinion as to whether Paterson used such a technique
or on the validity of the case you are making. I only wish to
correct a slight error on an obscure technical point.
Grant Edwards
>In article <80i62d$5sr$2...@q.seanet.com>, anonymous@bogus_address.con
>says...
>>
>> authorized outlets. A serious mistake, as it turned out.
>
>quite easy to figure out - I'm sitting here looking at some line printer
>output dated 1-dec-84 that I wrote and I have another copy by Clark
>Calkins that's dated 1983... this isn't Rocket Science...
Not at all. I can think of a couple people who reverse
engineered CP/M (without access to DRI source code) enough to
allow them to impliment a CP/M emulator that would run CP/M
applications under other OSes: (RSX-11, and Unix). Not to take
anything away from their efforts, but the CP/M part wasn't hard
at all -- writing the Z80 CPU emulator took a lot more work.
--
Grant Edwards grante Yow! What's the MATTER
at Sid?... Is your BEVERAGE
visi.com unsatisfactory?
Mark Statzer
> In 1984 I worked at a software publishing house, and one of our
> conditions for handling a product was to be given the SOURCE for
> whatever we were publishing. And from personal experience I can
> tell you authors would HOWL at that requirement. We kept their
> source code in a safe. UNTIL the company went bankrupt and everything
> was just dumped into a big pile and sold at a warehouse sale.
>
> For those interested, that was Software Strategies, Inc (SSI).
>
> In case you're wondering what ever happened to your program(s).
So it's safe to make a backup copy of my TRS-80 cassette B-1 Bomber?
Mark "waiting for the Win9x port" Statzer
--
*NOTE* I am not responsible for equipment damage due to reeeealy
dumb children with no parental supervision, and access to a hammer.
Jesper Hansen
>
> The ironic thing is that if you disassembled Windows today and then tried
> to resell the code then you'd probably end up in court.
>
If you tried to disassemble the 40 million lines of code in Windows
2000,
you'd more likely end up in the loonie house !
/Jesper
Dave Mundt
And, even if you DID disassemble it, you would still not have
anything really useful, because, unless you had a VERY good sourcer,
you would have only millions of lines of assembler code. no symbolic
references...No Labels...NO DOCUMENTATION as to what the programmers
really intended it to do.
Truely, therein lies madness.
Dave Mundt
Jesper Hansen <qtx...@tn.etx.ericsson.se> wrote:
Remove the mapson. from the email address to get to me...
I hate Cullers who gather from newsgroups
Visit my home page at http://www.esper.com/xvart/index.html
Chris Fischer
> Greetings and Salutations...
> And, even if you DID disassemble it, you would still not have
> anything really useful, because, unless you had a VERY good sourcer,
> you would have only millions of lines of assembler code. no symbolic
> references...No Labels...NO DOCUMENTATION as to what the programmers
> really intended it to do.
So if I was in that situation, and suddenly had lots of money from
selling V1.0, my next step would be to re-write in V2.0.
But this is all hypothetical.
>
> Truely, therein lies madness.
> Dave Mundt
>
> Jesper Hansen <qtx...@tn.etx.ericsson.se> wrote:
>
> >Limey wrote:
> >>
> >> The ironic thing is that if you disassembled Windows today and then tried
> >> to resell the code then you'd probably end up in court.
> >>
> >
> >If you tried to disassemble the 40 million lines of code in Windows
> >2000,
> >you'd more likely end up in the loonie house !
> >
> >/Jesper
>
> Remove the mapson. from the email address to get to me...
> I hate Cullers who gather from newsgroups
>
> Visit my home page at http://www.esper.com/xvart/index.html
--
Chris Fischer cfis...@ieee.org
Coda Software, Limited
Roland van Straten
Roland
Dave Mundt <RemoveThis...@esper.com> wrote in message
news:38303a41...@news.esper.com...
> Greetings and Salutations...
> And, even if you DID disassemble it, you would still not have
> anything really useful, because, unless you had a VERY good sourcer,
> you would have only millions of lines of assembler code. no symbolic
> references...No Labels...NO DOCUMENTATION as to what the programmers
> really intended it to do.
Average Torvaldsian
> > In 1984 I worked at a software publishing house, and one of our
> >
<stupid question = yes>
Is this "SSI" the same company that put out the original WordPerfect, before
WordPerfect Corp was formed? I dimly remember an advert in one of the computing
magazines from the mid eighies for WP 3.?, and was struck by the company name
-not- being WPCorp, but iirc SSI.
</stupid question>
Max.
Barry Watzman
not the same company as Software Strategies, Inc.
Barry Watzman
bill_h
>
> > bill_h wrote:
> > > In 1984 I worked at a software publishing house, and one of our
> > >
> > > For those interested, that was Software Strategies, Inc (SSI).
> Is this "SSI" the same company that put out the original WordPerfect, before
> WordPerfect Corp was formed? I dimly remember an advert in one of the computing
> magazines from the mid eighies for WP 3.?, and was struck by the company name
> -not- being WPCorp, but iirc SSI.
Nope, not the Wordperfect people. I think THAT SSI sold softwares mostly
for the Data General Nova computers. (At least, I think I read that some
where). They were located in Utah. Wordperfect was written for the Nova
and later got 'moved' onto the IBM PC. I think the earliest WordPerfect
packages came from SSI, but they changed the name after a while.
The SSI I was 'affiliated' with was in a suburb of Minneapolis. It had
been
started by one Gerald Van Diver, who put out a book listing all the
softwares
(suppposedly) available for (mostly) the IBM PC, but seems to me
included
Apples and some others.
This SSI (Software Strategies, Inc.) 'packaged' software from all sorts
of people. Sort of maroon binders in light gray slipcases. Mostly very
poor quality printing job inside. Basically went belly up in late 1984.
All of this was more or less a scam to get free software - which wasn't
as
easy in those days. They spent several millions of dollars producing
3-ring
bindered softwares, I think about 100,000 packages, out of which they
sold
(well, were paid for ...) about 350 packages. Maybe 375.
It's almost hard to believe, years later, how badly the operation was
run.
Bill
Tucson
Gary Helbig
>
> I talked to a guy in Seattly once who mentioned that you could read Gary
> K. comment lines in the source for DOS 1.1.
>
> If somebody had the time, effort and teh money, it would be
> interesting...
>
> -don
Urban legend. Just ain't true.
I was a Seattle-DOS OEM; had access to the entire distribution.
I did a text search for 'Gary', 'Kildall', 'Digital' ad naseum and it
turned up NOTHING.
SC-DOS was a 'from the ground-up' project. They reverse engineered SOME
of the commands, but not a lot.
And it was written _entirely_ in ASM, where CP/M had some parts written
in PL/M.
Gosh, I wish that myth would die.
Gary.
Donato B. Masaoy III
>
> Differences of opinion make for horse races and lawsuits. It's FAR from
> over yet. MS has chosen to make many many enemies in this business.
> We'll
> have to wait and see what happens in a couple months.....
>
I don't know the ins and outs, and I practice family and criminal law,
not intellectual property. But it seems that if there really was
something (but the settlement with IBM wasn't just that -- a settlement)
that Caldera would have added that to their suit against M$. Although
that suit is under the antitrust laws for intentionally and wrongfully
undermining DR-DOS essentially through FUD. Goes to trial in January.
anon...@bogus_address.com
On 1999-11-21 "ghelbig"@mailcity(.)com said:
>Donato B. Masaoy III wrote:
> > I talked to a guy in Seattle once who mentioned that you could
> > read Gary K. comment lines in the source for DOS 1.1.
> [...snip...]
> > -don
>
>Urban legend. Just ain't true.
>I was a Seattle-DOS OEM; had access to the entire distribution.
>I did a text search for 'Gary', 'Kildall', 'Digital' ad naseum and
>it turned up NOTHING.
>SC-DOS was a 'from the ground-up' project. They reverse engineered
>SOME of the commands, but not a lot.
>And it was written _entirely_ in ASM, where CP/M had some parts
>written in PL/M.
>Gosh, I wish that myth would die.
Whether or not there were comments by Kildall in the QDOS or the M$-DOS
source code is pretty much immaterial. If someone were going to steal
code, he'd have to be a complete idiot to leave in such obvious evidence
of his theft.
The more cogent question is whether or not code was indeed "lifted" from
CP/M, and used in those other O.S.es.
From the public record, we could very well come to a personal conclusion
that it was.
After all, why else would mega-million-dollar giant IBM, in 1981, enter
into an agreement with Digital Research which specified that IBM would
make CP/M-86 available to buyers of the IBM PC as an alternative operating
system to DOS...in return for which Digital Research explicitly agreed
not to sue IBM for theft of intellectual property?
That's a precedent-setting, and rather convincing, bit of evidence.
But then, I suppose there are a few folks out there who still believe that
Billy-Bob Clinton "...never had sex with that woman."
Ian St. John
<anonymous@bogus_address.com> wrote in message
news:81d4o6$o0i$1...@q.seanet.com...
I have a copy of the "MS-DOS Bible" which goes in to some detail on
the history. Nothing is said about source code, but 86-DOS was
deliberately designed to mimic CPM/80 data structures and layout such
as executable files, file control blocks, etc. This was a deliberate
move to allow automatic translation from 8080 CPM code to the new OS.
There is an image of a SCP advertisement to back this up.
DR probably used this as a basis for the 'theft'. Not the code, but
the layout of the supporting storage areas. I can't say where Paterson
would have gotten the layout, but I suspect that technical references
documented most or all of it.
Guy Macon
>> But then, I suppose there are a few folks out there who
>> still believe that Billy-Bob Clinton "...never had sex
>> with that woman."
That depend on your definition of "had" and "that"...
Barry Smith
Guy Macon <guym...@deltanet.com> wrote in message
news:81docc$r...@chronicle.concentric.net...
then I could believe that! :)
bill_h
> I have a copy of the "MS-DOS Bible" which goes in to some detail on
> the history. Nothing is said about source code, but 86-DOS was
> deliberately designed to mimic CPM/80 data structures and layout such
> as executable files, file control blocks, etc. This was a deliberate
> move to allow automatic translation from 8080 CPM code to the new OS.
> There is an image of a SCP advertisement to back this up.
>
> DR probably used this as a basis for the 'theft'. Not the code, but
> the layout of the supporting storage areas. I can't say where Paterson
> would have gotten the layout, but I suspect that technical references
> documented most or all of it.
The answer to this is easy. Paterson was hired by Paul Allen to work on
the Z80 Apple II add-in board (the Microsoft board), some one to two
years prior to the genesis of the IBM PC. Paterson was not able to do
the work, so it was given over to others to finish. This is mentioned in
the book 'Hard Drive' by James Wallace and Jim Erickson (Page 158).
Since the Microsoft Board was designed and built to run CP/M, that is to
allow Apple II's to run CP/M programs, we might assume Paterson had more
than passing exposure to, and knowledge of, CP/M, inside and out.
The LEGAL PROBELM for Microsoft is EXACTLY this: in order to have
access as a developer working on a CP/M application, MICROSOFT
SIGNED A LICENSING AGREEMENT THAT EXPLICITLY FORBADE REVERSE
ENGINEERING OR DECOMPILING CP/M.
Paterson's work might have passed un-noticed, EXCEPT that in
''developing'' 86-DOS, he was 'contaminated' by prior exposure
to LICENSED DRI MATERIAL.
For some reason, the personal computer press has ignored or
overlooked this story, this monumental theft, for almost
twenty years. There is no Q-DOS, if you see that name you're
seeing a part of the later 'cover story'. It's an attempt to
minimize what really happened, to more or less blow the whole
thing off as insignificant. But it created such a pile of
wealth there's simply no way to hide it. In this morning's paper
it said MS has SIXTEEN BILLION DOLLARS CASH on hand to fight
ANY CONCEIVABLE LAWSUITS.
I think intelligent people, given the facts, are capable of
determining what's significant and what isn't. And, what's
right and what isn't.
Bill
Tucson, AZ
Grant Edwards
>
>On 1999-11-21 "ghelbig"@mailcity(.)com said:
>
> >Donato B. Masaoy III wrote:
> > > I talked to a guy in Seattle once who mentioned that you could
> > > read Gary K. comment lines in the source for DOS 1.1.
> > [...snip...]
> > > -don
> >
> >Urban legend. Just ain't true.
> >I was a Seattle-DOS OEM; had access to the entire distribution.
> >I did a text search for 'Gary', 'Kildall', 'Digital' ad naseum and
> >it turned up NOTHING.
> >SC-DOS was a 'from the ground-up' project. They reverse engineered
> >SOME of the commands, but not a lot.
> >And it was written _entirely_ in ASM, where CP/M had some parts
> >written in PL/M.
> >Gosh, I wish that myth would die.
>
>Whether or not there were comments by Kildall in the QDOS or
>the M$-DOS source code is pretty much immaterial. If someone
>were going to steal code, he'd have to be a complete idiot to
>leave in such obvious evidence of his theft.
It happens. I worked at company who caught a competitor
stealing source code for the embedded software in one of our
products (land-mobile fleet radios). During the copyright
infringement trial, the defendants were forced to bring in the
source code for their product. After searching through some
hardcopy listings, several of our comments (including copyright
notices) were found.
There are plenty of complete idiots out there...
--
Grant Edwards grante Yow! I hope the
at "Eurythmics" practice birth
visi.com control...
Paul E. Bennett
bil...@sunsouthwest.com "bill_h" writes:
> For some reason, the personal computer press has ignored or
> overlooked this story, this monumental theft, for almost
> twenty years.
Doesn't the USA have a "Statute of Limitations" which prevents lawsuits being
brought after a certain period of time? How long is that period?
--
Paul E. Bennett ................... <p...@amleth.demon.co.uk>
Forth based HIDECS Consultancy .... <www.amleth.demon.co.uk>
Tel: +44 (0)7971-620145
Going Forth Safely
phil
wrote:
>In article <81d4o6$o0i$1...@q.seanet.com>, anonymous@bogus_address.com wrote:
<snip>
>
>It happens. I worked at company who caught a competitor
>stealing source code for the embedded software in one of our
>products (land-mobile fleet radios). During the copyright
>infringement trial, the defendants were forced to bring in the
>source code for their product. After searching through some
>hardcopy listings, several of our comments (including copyright
>notices) were found.
There was an article a while ago in ESJ by PJP describing a copyright
suit where he used white space to help indicate theft. Still, showing
source copyright notifications to the judge must have been quite
bizarre.
phil.
--
The world is divided into two sorts of people: those that think the
world is divided into two sorts of people and those that don't.
George Neuner
wrote:
>The LEGAL PROBELM for Microsoft is EXACTLY this: in order to have
>access as a developer working on a CP/M application, MICROSOFT
>SIGNED A LICENSING AGREEMENT THAT EXPLICITLY FORBADE REVERSE
>ENGINEERING OR DECOMPILING CP/M.
>
>Paterson's work might have passed un-noticed, EXCEPT that in
>''developing'' 86-DOS, he was 'contaminated' by prior exposure
>to LICENSED DRI MATERIAL.
>
>For some reason, the personal computer press has ignored or
>overlooked this story, this monumental theft, for almost
>twenty years. There is no Q-DOS, if you see that name you're
>seeing a part of the later 'cover story'. It's an attempt to
>minimize what really happened, to more or less blow the whole
>thing off as insignificant. But it created such a pile of
>wealth there's simply no way to hide it. In this morning's paper
>it said MS has SIXTEEN BILLION DOLLARS CASH on hand to fight
>ANY CONCEIVABLE LAWSUITS.
Being 'contaminated' by prior exposure does not an infringement suit
make. Ideas are not protected under either patent or copyright law
law - patent law protects physical derivation of the idea and
copyright law protects expression of the idea.
Having been exposed to DRI's source, Paterson certainly may have
devloped ideas or preconceptions about how a new OS should [or could]
be structured. However, whatever he may have seen while legally
exposed to it is immaterial. To sue for infringement, DRI would have
to prove that either 1) Paterson illegally retained access to
non-public source and/or documentation (possibly by copying it) and
used it as a basis for his own work, or 2) that he deliberately
reverse engineered their product with the intent of producing a
work-alike. If he used personal notes taken while legally exposed to
the source, there may also be a contractual issue depending upon the
license agreement - but that would be separate from the infringement.
Based on everything I have heard, I don't think DRI could prove either
of those points. IBM's settlement with them could have simply been to
divert the bad publicity of a trial. Remember that, at the time, IBM
was still embroiled in the Justice Department's anti-trust suit and
didn't need more legal trouble of any kind.
George Neuner
Dynamic Resolutions, Inc.
===================================================
The opinions expressed herein are my own and do not
reflect the opinions or policies of my employer.
===================================================
bill_h
>
> In article <383AC6...@sunsouthwest.com>
> bil...@sunsouthwest.com "bill_h" writes:
>
> > overlooked this story, this monumental theft, for almost
> > twenty years.
>
> brought after a certain period of time? How long is that period?
The copyrights will run out in about fifty years. Give or take a few.
Curiously, they're designed to run many years beyond the death of an
author.
Richard Plinston
: The answer to this is easy. Paterson was hired by Paul Allen to work on
: the Z80 Apple II add-in board (the Microsoft board), some one to two
: years prior to the genesis of the IBM PC. Paterson was not able to do
: the work, so it was given over to others to finish. This is mentioned in
: the book 'Hard Drive' by James Wallace and Jim Erickson (Page 158).
: Since the Microsoft Board was designed and built to run CP/M, that is to
: allow Apple II's to run CP/M programs, we might assume Paterson had more
: than passing exposure to, and knowledge of, CP/M, inside and out.
Not only that, but SCP were themselves CP/M OEMs with access to all
source code and other material that DRI would provide. They
built Z80 based boxes (Zebra?). I have some ads somewhere in
my collection of old Bytes and Interfcae Age.
--
Richard Plinston
:> > > I talked to a guy in Seattle once who mentioned that you could
:> > > read Gary K. comment lines in the source for DOS 1.1.
:> > [...snip...]
:> > > -don
:> >
:> >Urban legend. Just ain't true.
:> >I was a Seattle-DOS OEM; had access to the entire distribution.
:> >I did a text search for 'Gary', 'Kildall', 'Digital' ad naseum and
:> >it turned up NOTHING.
:> >SC-DOS was a 'from the ground-up' project. They reverse engineered
:> >SOME of the commands, but not a lot.
:> >And it was written _entirely_ in ASM, where CP/M had some parts
:> >written in PL/M.
:> >Gosh, I wish that myth would die.
First of all the 'offending' DOS was 1.0. MS-DOS/PC-DOS 1.1 was
a rewrite by IBM in order to eliminate all DRI based code as well
as get rid of the bugs.
When PC-DOS 1.0 was released DRI threatened to sue IBM and SCP
and MS over Copyright infringement. There were at least two
items that showed reverse engineering at the least. One was that
CP/M had a bug relating to the close of a file. This existed
in PC-DOS 1.0 in exactly the same way in the code. The other
was that PC-DOS 1.0 could be made to show a DRI Copyright. This
was most likely in a utility program and may have been due
to particular sequences of code generating it rather than
it being in 'clear'.
IBM settled with DRI by: a) rewriting all DRI code to produce 1.1,
b) an undisclosed sum of money, c) offering CP/M-86 alongside
PC-DOS (and UCSD), d) othe 'considerations.
Of course they stuffed DRI by having CP/M-86 at $250 or so
while PC-DOS was $70 (or so).
As for PL/M, yes CP/M BDOS was originally in PL/M. However
the alegation was that the object code was put through a
disassembler then into Intel's 8 -> 16 bit code converter
to produce the inital 8086 ASM code. This is why it
was 'Quick' and 'Dirty'.
Of course there was no actual product called QDOS, just as
there was never a product from MS called Chicargo (it was
Windows 95 when released).
--
Guy Macon
>was that PC-DOS 1.0 could be made to show a DRI Copyright. This
>was most likely in a utility program and may have been due
>to particular sequences of code generating it rather than
>it being in 'clear'.
I have heard this claim, but never with instructions
on how to actually do it. Does anyone know how?
Gary Helbig
Not only did I to a text search in PC-DOS and MS-DOS for this, I even
'cracked' them both to look for easter eggs.
Gosh, I wish this myth would die.
Gary.
Gary Helbig
Having looked at the internals of both, I can state that no code was
'lifted'.
CP/M and DOS were both essentially "wrappers" for disk I/O routines.
The calling interface for these were _completely_ different.
And the DOS file system was elegant, when compared to CP/M's.
The only 'control structure' that was carried over was the first 256
bytes of the .COM format file. This was publicly documented, so that
other people coule write applications for CP/M. It was supported so
that CP/M-80 programs could be 'translated' for a first-cut release, but
the 'preferred interface' was the .EXE format.
CP/M was written in PL/M. It was Dr. Kildall's favorite language, and
Digital Research used it a lot.
SB-86 (or SC-DOS) was written in assembly. And assembled with a
not-very-good assembler, probably Sorcim's ACT.
The "public record" is full of hearsay and urban legends. It can be
ignored.
And why did IBM make DRI sign a no litigation pact? Because they
could. Not unprecedented, and rather mote.
Gary.
anonymous@bogus_address.com wrote:
>
> On 1999-11-21 "ghelbig"@mailcity(.)com said:
>
> >Donato B. Masaoy III wrote:
> > > I talked to a guy in Seattle once who mentioned that you could
> > > read Gary K. comment lines in the source for DOS 1.1.
> > [...snip...]
> > > -don
> >
> >Urban legend. Just ain't true.
> >I was a Seattle-DOS OEM; had access to the entire distribution.
> >I did a text search for 'Gary', 'Kildall', 'Digital' ad naseum and
> >it turned up NOTHING.
> >SC-DOS was a 'from the ground-up' project. They reverse engineered
> >SOME of the commands, but not a lot.
> >And it was written _entirely_ in ASM, where CP/M had some parts
> >written in PL/M.
> >Gosh, I wish that myth would die.
>
> Whether or not there were comments by Kildall in the QDOS or the M$-DOS
> source code is pretty much immaterial. If someone were going to steal
> code, he'd have to be a complete idiot to leave in such obvious evidence
> of his theft.
>
> The more cogent question is whether or not code was indeed "lifted" from
> CP/M, and used in those other O.S.es.
>
> From the public record, we could very well come to a personal conclusion
> that it was.
>
> After all, why else would mega-million-dollar giant IBM, in 1981, enter
> into an agreement with Digital Research which specified that IBM would
> make CP/M-86 available to buyers of the IBM PC as an alternative operating
> system to DOS...in return for which Digital Research explicitly agreed
> not to sue IBM for theft of intellectual property?
>
> That's a precedent-setting, and rather convincing, bit of evidence.
>
anon...@bogus_address.con
On 1999-11-27 "ghelbig"@mailcity(.)com said:
>Having looked at the internals of both, I can state that no code was
>'lifted'.
>CP/M and DOS were both essentially "wrappers" for disk I/O routines.
>The calling interface for these were _completely_ different.
>And the DOS file system was elegant, when compared to CP/M's.
>The only 'control structure' that was carried over was the first 256
>bytes of the .COM format file. This was publicly documented, so
>that other people coule write applications for CP/M. It was
>supported so that CP/M-80 programs could be 'translated' for a
>first-cut release, but the 'preferred interface' was the .EXE
>format.
>CP/M was written in PL/M. It was Dr. Kildall's favorite language,
>and Digital Research used it a lot.
>SB-86 (or SC-DOS) was written in assembly. And assembled with a
>not-very-good assembler, probably Sorcim's ACT.
>The "public record" is full of hearsay and urban legends. It can be
>ignored.
>And why did IBM make DRI sign a no litigation pact? Because they
>could. Not unprecedented, and rather mote.
That's all well and good. Nevertheless, a nagging little cricket of doubt
continues its cold chirping in the back recesses of my mind.
Why would arrogant, mega-multi-million-dollar IBM agree to make available,
for =their= PC, an alternative operating system from funky little Mom-and-
Pop D.R.I.? Would the elephant enter into contracts with the flies?
I'm not certain the question can be so glibly dismissed.
Paul Schlyter
Gary Helbig <"ghelbig"@mailcity(.)com> wrote:
> Having looked at the internals of both, I can state that no code was
> 'lifted'.
>
> CP/M and DOS were both essentially "wrappers" for disk I/O routines.
>
> The calling interface for these were _completely_ different.
with CP/M: all the FCB file functions calls where there, and
they even often had the same function numbers, and took the
same parameters in sometimes even the same CPU registers!
In DOS you could even, from a .COM program, call these functions with
a "Call 5" instead of an "Int 21h" -- and you could terminate
the program with a "Call 0", ir with a "Ret" from the main level,
instead of an "Int 20h" -- just like in CP/M !!!!
> And the DOS file system was elegant, when compared to CP/M's.
allowed that. And if you want to undelete a deleted file in
DOS, you must guess all allocation blocks except the first one:
if several files has been deleted from a fragmented disk, such
guesses will likely be erroneous. In CP/M there's no need
for this, since the allocation information still is intact
on deleted files, until the directory entry is overwritten.
No, you don't need to lecture to me about subdirectories, huge disk
volumes, and other advantages of the FAR file system compared to
CP/M's file system: I'm very well aware of this. I just wanted
to point out that in some cases, the FAT file system was inferior.
> The only 'control structure' that was carried over was the first 256
> bytes of the .COM format file.
same function number. This too is necessary if CP/M programs are to
be easily portable to DOS, of course.
--
----------------------------------------------------------------
Paul Schlyter, Swedish Amateur Astronomer's Society (SAAF)
Grev Turegatan 40, S-114 38 Stockholm, SWEDEN
e-mail: pau...@saaf.se paul.s...@ausys.se pa...@inorbit.com
WWW: http://hotel04.ausys.se/pausch http://welcome.to/pausch
Barry Watzman
thread.
First I liked the line "you don't need to lecture to me about
subdirectories" -- well, that's certainly a true statement, because not
only did CP/M not have them, NEITHER DID MS-DOS until version 2.0.
Then I like the line: "The only 'control structure' that was carried over
was the first 256
> bytes of the .COM format file" Huh ???? a .COM file in CP/M HAS NO
DEFINED STRUCTURE, AT ALL. It is simply loaded, arbitrarily at 100H and
control is transferred to 100H. Period. End of structure.
For what it's worth, I come down on the side that says that until 2.0, at
least, MS-DOS was neither significantly inferior to nor superior to CP/M.
The only real difference was that CP/M stored the allocation information
in directory entries, while MS-DOS stored the same information in a single
FAT table.
And finally, about the comments relative to the value of the 86-DOS disk,
and being able to retire to an island somewhere. I guess that makes me
wealthy also, because I have the original SCP 86-DOS disks from version
0.33 (which I think is before MS bought it) to 2.0 complete with the
original documentation. Obtained direct from SCP, not from Lifeboat, and
in SCP binders. And I have the hardware to actually run it, if it still
works, in an S-100 system with 8" floppy drives. Does anyone really think
that this is worth $10,000 or more ? I HAVE thought about offering it for
sale on E-Bay (although I would probably not sell the CPU support board,
as I use it in my CP/M systems also).
Barry Watzman
Richard Plinston
: Then I like the line: "The only 'control structure' that was carried over
: was the first 256
I think that he meant the PSP, the first 256 bytes (100h) of memory.
This included the FCB format, command line, jump table, ..
: For what it's worth, I come down on the side that says that until 2.0, at
: least, MS-DOS was neither significantly inferior to nor superior to CP/M.
: The only real difference was that CP/M stored the allocation information
: in directory entries, while MS-DOS stored the same information in a single
: FAT table.
I suspect that you are talking CP/M 2.2 here vs MS-DOS 1.x. This is
true. However, DRI also had MP/M and .REL relocatable programs and
this was multi-tasking and multi-user.
CP/M-86 did have some advantages over MS-DOS 1.x as it was somewhat
more flexible in configuration, but MS-DOS 2.x nullified this.
By the time DOS 2.x came out, DRI had brought out Concurrent-CP/M-86
which had proper multitasking (pre-emptive), but MS-DOS had
talked the market into DOS by pre-announcing all the things
that would be in the 'next version' of DOS creating the now
routine 'vapourware'. Some of the announcements about 'the
next version' made by Paul Allan in 1982 did not get implemented
until MS-DOS 5, some not at all.
--
Charles E. Bortle, Jr.
Richard Plinston wrote in message <81scjh$r02$1...@aklobs.kc.net.nz>...
>By the time DOS 2.x came out, DRI had brought out Concurrent-CP/M-86
>which had proper multitasking (pre-emptive), but MS-DOS had
>talked the market into DOS by pre-announcing all the things
>that would be in the 'next version' of DOS creating the now
>routine 'vapourware'. Some of the announcements about 'the
>next version' made by Paul Allan in 1982 did not get implemented
>until MS-DOS 5, some not at all.
>--
Of course, M$ did not invent the 'vapourware' technique...
Big Blue had been doing it for years.
--
Charles cbr...@ix.netcom.com
"For God So Loved The World, That He Gave His
Only Begotten Son, That Whosoever Believeth
In Him Should Not Perish, But Have Everlasting
Life"John3:16 * http://pw2.netcom.com/~cbrtjr/wrdthing.html *
Evandro Menezes
wrote in <38415E08...@neo.rr.com>:
>Huh ???? a .COM file in CP/M HAS NO
>DEFINED STRUCTURE, AT ALL. It is simply loaded, arbitrarily at 100H and
>control is transferred to 100H. Period. End of structure.
As a matter of fact, the first 256 bytes of a program memory space was
a structure known as the PSP and has always been fully documented,
with its FCB's and command line buffer, etc. It's still present in
Windows 3.11 and in the DOS box in Windows Millennium and 2000...
______________________________________________________
Evandro Menezes Austin, TX ICQ:7957253
eva...@geocities.com geocities.com/evandro
Evandro Menezes
wrote in <81qqs6$eig$1...@merope.saaf.se>:
>Not at all! DOS maintained a lot of backwards compatibility
>with CP/M: all the FCB file functions calls where there, and
>they even often had the same function numbers, and took the
>same parameters in sometimes even the same CPU registers!
Yes, almost all of the 20 or so 1st system calls mirror CP/M's, even
the register allocation for parameters (of course, in 8086's
equivalent ones for 8080's, as set by Intel for code translation).
>In DOS you could even, from a .COM program, call these functions with
>a "Call 5" instead of an "Int 21h" -- and you could terminate
>the program with a "Call 0", ir with a "Ret" from the main level,
>instead of an "Int 20h" -- just like in CP/M !!!!
Actually, a COM program can still to this day use the CALL 5 and
CALL 0 interfaces...
Richard Plinston
:>In DOS you could even, from a .COM program, call these functions with
:>a "Call 5" instead of an "Int 21h" -- and you could terminate
:>the program with a "Call 0", ir with a "Ret" from the main level,
:>instead of an "Int 20h" -- just like in CP/M !!!!
: Actually, a COM program can still to this day use the CALL and
: CALL055555rfaces...
I seem to recall that this stopped working in MS-DOS 6.2 and does
not work in Win95 or 98.
It may work in NT 4 because the DOS box is not a real DOS but an
emulator. It seems unlikely that 2000 will run .COMs at all.
-
-
John Elliott
>In comp.os.cpm Evandro Menezes <rin...@trbpvgvrf.pbz> wrote:
>: Actually, a COM program can still to this day use the CALL and
>: CALL055555rfaces...
>
>I seem to recall that this stopped working in MS-DOS 6.2 and does
>not work in Win95 or 98.
Just tried it in MSDOS 7.00 (the DOS from the original Win95) and it works
fine. It doesn't work in DEBUG because DEBUG is buggy (and has been since
MSDOS 2) but if you use SID to enter the program, it works. I don't know
if it works when the GUI's loaded.
>It may work in NT 4 because the DOS box is not a real DOS but an
>emulator. It seems unlikely that 2000 will run .COMs at all.
It does work in NT4, at least as far as printing on the screen is
concerned.
I had heard that the FCB calls you can get at (from CALL 5 or INT 0x21)
are slowly withering away - some don't work under the Win95 GUI, others
don't work on FAT32 drives...
------------- http://www.seasip.demon.co.uk/index.html --------------------
John Elliott |BLOODNOK: "But why have you got such a long face?"
|SEAGOON: "Heavy dentures, Sir!" - The Goon Show
:-------------------------------------------------------------------------)
Mark Zenier
>That's all well and good. Nevertheless, a nagging little cricket of doubt
>continues its cold chirping in the back recesses of my mind.
>
>Why would arrogant, mega-multi-million-dollar IBM agree to make available,
>for =their= PC, an alternative operating system from funky little Mom-and-
>Pop D.R.I.? Would the elephant enter into contracts with the flies?
>
>I'm not certain the question can be so glibly dismissed.
>
IBM offered several OS's with the PC: MS-DOS, CP/M-86 and The UCSD
P-System Pascal. CP/M-86 was sold for the going rate. As was the
P-System, which cost even more.
It was a matter of arrogance of the market leader DRI (or, perhaps,
they were trapped by their base of other customers, unable to drop the
price for just IBM or unwilling to drop the price for everybody), and
Microsoft using their lower price as a wedge to get the biggest share of
systems installed and, more importantly, the biggest share of software
development. As Judge Jackson's Finding of Fact report shows, Microsoft
follows the same strategy, (making sure they get the vast majority
of the software developers producing applications only for their OS),
to this day. (Executive summary: Netscape+Java was viewed as a threat
because it could be a platform for applications development that could
supercede Windows, not because a browser was a type of software that
Microsoft didn't offer at that time).
Mark Zenier mze...@eskimo.com mze...@netcom.com Washington State resident
bma...@iglou.com
On 1999-11-29 rin...@trbpvgvrf.pbz said:
>Barry Watzman <Wat...@neo.rr.com> wrote in <38415E08.267E7C48@neo.
>rr.com>:
>>Huh ???? a .COM file in CP/M HAS NO
>>DEFINED STRUCTURE, AT ALL. It is simply loaded, arbitrarily at
>>100H and control is transferred to 100H. Period. End of
>structure.
>As a matter of fact, the first 256 bytes of a program memory space
>was a structure known as the PSP and has always been fully
>documented, with its FCB's and command line buffer, etc. It's
>still present in Windows 3.11 and in the DOS box in Windows
>Millennium and 2000...
True, but the PSP is not part of the COM file, it's created by DOS when a
COM file is loaded.
Net-Tamer V 1.08X - Test Drive
Evandro Menezes
<3844...@news.iglou.com>:
>True, but the PSP is not part of the COM file, it's created by DOS when a
>COM file is loaded.
Well pointed out. Thanks for complementing my statements.
BTW, PSP's are still created for EXE as well, but in a separate,
256-byte long segment for this sole purpose. Alas, they're still
present under Windows 3.11 in low memory (1st 640K) for all PE
modules, DLLs included...
m...@here.com
> That's all well and good. Nevertheless, a nagging little cricket of doubt
> continues its cold chirping in the back recesses of my mind.
> Why would arrogant, mega-multi-million-dollar IBM agree to make available,
> for =their= PC, an alternative operating system from funky little Mom-and-
> Pop D.R.I.? Would the elephant enter into contracts with the flies?
They also offered a 3rd alternative initially....
And of questions, why would IBM require second sourcing of
the 8088 from Intel? (My true blue XT has ab AMD chip)
> I'm not certain the question can be so glibly dismissed.
There are simpler answers than conspiracy.
Chris Fischer
> In comp.arch.embedded anonymous@bogus_address.con wrote:
>
> > That's all well and good. Nevertheless, a nagging little cricket of doubt
> > continues its cold chirping in the back recesses of my mind.
>
> > Why would arrogant, mega-multi-million-dollar IBM agree to make available,
> > for =their= PC, an alternative operating system from funky little Mom-and-
> > Pop D.R.I.? Would the elephant enter into contracts with the flies?
>
> They also offered a 3rd alternative initially....
>
> And of questions, why would IBM require second sourcing of
> the 8088 from Intel? (My true blue XT has ab AMD chip)
>
To insure a supply even if Intel went out of business.
>
> > I'm not certain the question can be so glibly dismissed.
>
> There are simpler answers than conspiracy.
--
Chris Fischer cfis...@ieee.org
Coda Software, Limited
Tor Sjowall
>
>Not at all! DOS maintained a lot of backwards compatibility
>with CP/M: all the FCB file functions calls where there, and
>they even often had the same function numbers, and took the
>same parameters in sometimes even the same CPU registers!
>
>In DOS you could even, from a .COM program, call these functions with
>a "Call 5" instead of an "Int 21h" -- and you could terminate
>the program with a "Call 0", ir with a "Ret" from the main level,
>instead of an "Int 20h" -- just like in CP/M !!!!
>
The "call 5" convention is is still supported in NT4, I just checked.
If you look at location 5 in the psp, you will find "CALL F01D:FEF0".
This address is equivalent to the 'flat' address 1000C0, or 0:C0 since
the 8088 wrapped addresses at 1Mbyte. At address 0:c0 there is a jump
deep into msdos.
On the 80286, the addresses did NOT wrap at 1M, so they had to put
hardware into the motherboard to disable address line A20 to support
the "call 5" convention. This is still done on all motherboards today.
So the legacy of CP/M is not only found in msdos, but also in the PC
hardware.
Tor.
(remove the OBVIOUS from my email address)
Ing. Franz Glaser
> ...
> The "call 5" convention is is still supported in NT4, I just checked.
> If you look at location 5 in the psp, you will find "CALL F01D:FEF0".
> This address is equivalent to the 'flat' address 1000C0, or 0:C0 since
> the 8088 wrapped addresses at 1Mbyte. At address 0:c0 there is a jump
> deep into msdos.
> On the 80286, the addresses did NOT wrap at 1M, so they had to put
> hardware into the motherboard to disable address line A20 to support
> the "call 5" convention. This is still done on all motherboards today.
> So the legacy of CP/M is not only found in msdos, but also in the PC
> hardware.
Well, Tor,
the A20 feature is helpful for the CALL 5 too, but it has mainly
other reasons. It allows the use of the 65kB memory above the 1 MB
boundary.
Regards,
--
Franz Glaser, Glasau 3, A-4191 Vorderweissenbach Austria +43-7219-7035-0
Muehlviertler Elektronik Glaser. Industrial control and instrumentation
http://members.eunet.at/meg-glaser/ mailto:meg-g...@eunet.at
http://www.geocities.com/~franzglaser/ http://members.xoom.com/f_glaser/
John Elliott
>If you look at location 5 in the psp, you will find "CALL F01D:FEF0".
... which also means that the word at (psp+6) gives a TPA size of 64k (less
256 bytes of PSP and 16 bytes of FDOS). Quite elegant.
MS-DOS also allows you to get at INT 0x21 services by CALL 0050h, which
CP/M obviously doesn't.
Evandro Menezes
wrote in <3846bb80....@news.askertomra.no>:
>On the 80286, the addresses did NOT wrap at 1M, so they had to put
>hardware into the motherboard to disable address line A20
Alas, it DID in real mode, although it was not supposed to. Later
generations maintained this behavior (classical example of fixing bugs
by changing the docs... :^) )...
Paul Schlyter
Tor Sjowall <t...@tomra.OBVIOUS.no> wrote:
> pau...@merope.saaf.se (Paul Schlyter) wrote:
>
>>Not at all! DOS maintained a lot of backwards compatibility
>>with CP/M: all the FCB file functions calls where there, and
>>they even often had the same function numbers, and took the
>>same parameters in sometimes even the same CPU registers!
>>
>>In DOS you could even, from a .COM program, call these functions with
>>a "Call 5" instead of an "Int 21h" -- and you could terminate
>>the program with a "Call 0", ir with a "Ret" from the main level,
>>instead of an "Int 20h" -- just like in CP/M !!!!
>
Try checking out Windows-95 or -98 instead if you like.
> If you look at location 5 in the psp, you will find "CALL F01D:FEF0".
> the 8088 wrapped addresses at 1Mbyte. At address 0:c0 there is a jump
> deep into msdos.
>
> the "call 5" convention. This is still done on all motherboards today.
>
> So the legacy of CP/M is not only found in msdos, but also in the PC
> hardware.
instance know why the PC timer tick ticks 18.2 times per second? That's
a pretty awkward number really, so why did they choose that number?
Yes, there's a reason, and if you want to know, I can tell you.
Robert Cicconetti
> instance know why the PC timer tick ticks 18.2 times per second? That's
> a pretty awkward number really, so why did they choose that number?
> Yes, there's a reason, and if you want to know, I can tell you.
I'll bite.. why? I'd presume it'd be for serial port timing or something
Robert Cicconetti
Ralf Kern
Paul Schlyter wrote:
> Yup, there are lots of funny legacies in the PC hardware. Do you for
> instance know why the PC timer tick ticks 18.2 times per second? That's
> a pretty awkward number really, so why did they choose that number?
> Yes, there's a reason, and if you want to know, I can tell you.
I guess this number can be derived from some line frequency or the like
of the CRTs using the NTSC standard.
Ralf
map
cuntervalue was 2^16 = 65536 ==> 1,19318Mhz/65536 = 18.20675 ms. This
time was used to generatd the timerinterrupt.
Bye, Dieter.
Paul Schlyter schrieb:
>
> In article <3846bb80....@news.askertomra.no>,
> Tor Sjowall <t...@tomra.OBVIOUS.no> wrote:
>
> > pau...@merope.saaf.se (Paul Schlyter) wrote:
> >
> >>Not at all! DOS maintained a lot of backwards compatibility
> >>with CP/M: all the FCB file functions calls where there, and
> >>they even often had the same function numbers, and took the
> >>same parameters in sometimes even the same CPU registers!
> >>
> >>In DOS you could even, from a .COM program, call these functions with
> >>a "Call 5" instead of an "Int 21h" -- and you could terminate
> >>the program with a "Call 0", ir with a "Ret" from the main level,
> >>instead of an "Int 20h" -- just like in CP/M !!!!
> >
> > The "call 5" convention is is still supported in NT4, I just checked.
>
> NT4 emulates a fairly old version of DOS: DOS 5.0 to be more precise.
> Try checking out Windows-95 or -98 instead if you like.
>
> > If you look at location 5 in the psp, you will find "CALL F01D:FEF0".
> > This address is equivalent to the 'flat' address 1000C0, or 0:C0 since
> > the 8088 wrapped addresses at 1Mbyte. At address 0:c0 there is a jump
> > deep into msdos.
> >
> > On the 80286, the addresses did NOT wrap at 1M, so they had to put
> > hardware into the motherboard to disable address line A20 to support
> > the "call 5" convention. This is still done on all motherboards today.
> >
> > So the legacy of CP/M is not only found in msdos, but also in the PC
> > hardware.
>
> Yup, there are lots of funny legacies in the PC hardware. Do you for
> instance know why the PC timer tick ticks 18.2 times per second? That's
> a pretty awkward number really, so why did they choose that number?
> Yes, there's a reason, and if you want to know, I can tell you.
>
Tor Sjowall
>Tor Sjowall wrote:
>> ...
>> The "call 5" convention is is still supported in NT4, I just checked.
>> If you look at location 5 in the psp, you will find "CALL F01D:FEF0".
>> This address is equivalent to the 'flat' address 1000C0, or 0:C0 since
>> the 8088 wrapped addresses at 1Mbyte. At address 0:c0 there is a jump
>> deep into msdos.
>> On the 80286, the addresses did NOT wrap at 1M, so they had to put
>> hardware into the motherboard to disable address line A20 to support
>> the "call 5" convention. This is still done on all motherboards today.
>> So the legacy of CP/M is not only found in msdos, but also in the PC
>> hardware.
>
>Well, Tor,
>
>the A20 feature is helpful for the CALL 5 too, but it has mainly
>other reasons. It allows the use of the 65kB memory above the 1 MB
>boundary.
>
>Regards,
Not really, you have to ENABLE the A20 address line to access the
memory above 1Mbyte, not disable it. In DOS, A20 is normally disabled
so that old programs will work.
The 65K-16 bytes above 1Meg is accessed with a segment register at
FFFF and no wrap after the offset is added.
Tor.
Tor Sjowall
>
>Yup, there are lots of funny legacies in the PC hardware. Do you for
>instance know why the PC timer tick ticks 18.2 times per second? That's
>a pretty awkward number really, so why did they choose that number?
>Yes, there's a reason, and if you want to know, I can tell you.
>
Just guessing: Did it have anything to do with the dram refresh
cycles? They used a DMA channel and a timer to do dram refresh on the
original PC and AT.
Tor.
anon...@bogus_address.con
On 1999-12-03 pau...@merope.saaf.se(PaulSchlyter) said:
>Yup, there are lots of funny legacies in the PC hardware. Do you
>for instance know why the PC timer tick ticks 18.2 times per
>second? That's a pretty awkward number really, so why did they
>choose that number? Yes, there's a reason, and if you want to know,
>I can tell you.
Uff da! Tell us, Paul. :)
Steven R. Wheeler
> Yup, there are lots of funny legacies in the PC hardware. Do you for
> instance know why the PC timer tick ticks 18.2 times per second? That's
> a pretty awkward number really, so why did they choose that number?
> Yes, there's a reason, and if you want to know, I can tell you.
Having written a BIOS, I know this one, teacher!
It's because at that rate, you get 65536 interrupts in 24 hours for your
real-time clock. Actually, IIRC, there was something like a 10-second
discrepancy in the original PC BIOS, and if it detected a "midnight
crossing" when you made a BIOS call to read the time, it would adjust
the time by that amount.
- wheels
Peter Seed
Steven R. Wheeler wrote in message <3847CC...@xpert.net>...
>Paul Schlyter wrote:
>> Yup, there are lots of funny legacies in the PC hardware. Do you for
>> instance know why the PC timer tick ticks 18.2 times per second? That's
>> a pretty awkward number really, so why did they choose that number?
>> Yes, there's a reason, and if you want to know, I can tell you.
>
>Having written a BIOS, I know this one, teacher!
Wow, what BIOS is that, so I don't have to use it!
>It's because at that rate, you get 65536 interrupts in 24 hours for your
>real-time clock. Actually, IIRC, there was something like a 10-second
>discrepancy in the original PC BIOS, and if it detected a "midnight
>crossing" when you made a BIOS call to read the time, it would adjust
>the time by that amount.
>
Let's see, 65536 interrupts in 86400 seconds is one every 1.32 seconds. Try
"65536 interrupts per hour" ;-)
Piero Fraviga
Steven R. Wheeler <swhe...@xpert.net> wrote in message
3847CC...@xpert.net...
> Paul Schlyter wrote:
> > Yup, there are lots of funny legacies in the PC hardware. Do you for
> > instance know why the PC timer tick ticks 18.2 times per second? That's
> > a pretty awkward number really, so why did they choose that number?
> > Yes, there's a reason, and if you want to know, I can tell you.
>
> Having written a BIOS, I know this one, teacher!
>
> real-time clock. Actually, IIRC, there was something like a 10-second
> discrepancy in the original PC BIOS, and if it detected a "midnight
> crossing" when you made a BIOS call to read the time, it would adjust
> the time by that amount.
>
That actually gives 65536 interrupts in 1 hour, and an about 9.6 seconds
discrepancy
on a 24-hour basis. Anyway, while they were at it, why did they use a
1.19318 MHz
clock to get the 18.2 Hz rate, and not a 1.193046 MHz clock, which would
yield a
few hundreths of second error a day? Both frequencies seem rather odd to my,
admittedly
scarce, hardware design knowledge...
P.F.
Paul Schlyter
Ralf Kern <ke...@ai.fh-nuernberg.de> wrote:
>> Yup, there are lots of funny legacies in the PC hardware. Do you for
>> instance know why the PC timer tick ticks 18.2 times per second? That's
>> a pretty awkward number really, so why did they choose that number?
>> Yes, there's a reason, and if you want to know, I can tell you.
>
> of the CRTs using the NTSC standard.
with television.
PC's and NTSC-525 (NTSC-M) or "Why does the PC tick 18.2 times/second?"
=======================================================================
The timer tick count in PC's increments about 18.2 times/second,
which closely corresponds to 65536 (=10000hex) ticks per hour. Was
the intention to make this exactly 65536 ticks/hour? No, this is
just a concicdence.
The system clock is nominally 14.31818 MHz. This is divided by 12 to
give the 1.19318166666... MHz clock to the 8253/8254 (which is often
incorrectly said to be 1.193180 MHz). This frequency is divided by
65536 in channel zero of the 8253/8254 to give the timer tick rate of
18.206507364909 Hz nominal, with a cycle period of 54.92541649846559
ms nominal. There are 65543.4265 ticks in an hour.
There are 1573042.24 ticks in a day, but when the BIOS was written,
the 1.19318166666... MHz frequency was approximated to 1.193180 MHz,
so the BIOS writers used 1573040 (001800B0 hex) ticks per day. This
contributes a 1.42166 ppm error, which is insignificant compared to
the clock frequency inaccuracy (the absolute accuracy of the 14.31818
MHz clock depends mainly on the quality of the 14.31818 MHz crystal
or crystal oscillator module. Initial frequency errors of +/- 20 ppm
(0.002%) are common. Temperature and age will also cause frequency
drift.)
1573042.24 ticks/day corresponds to 65543.43 ticks/hour, which is
0.011% more than 65536 ticks/hour - if the intention was 65536
ticks/hour exactly, they could certainly have made the agreement
better than that!
The input frequency to the timer chip is 14.31818 MHz / 12 =
1.19318166666... MHz. If we instead divide by 4 we get:
14.31818 MHz / 4 = 3.579545 MHz
which is exactly the same as the NTSC-M (525 lines) colour subcarrier
frequency of 3579545 Hz.
Thus it appears that the frequency of the PC's timer tick was
selected in order to use the same quartz crystal to also generate an
NTSC-525 colour subcarrier frequency by simply feeding the timer chip
frequency into a x3 frequency multiplier. The good ol' CGA
video/graphics cards are, as you may remember, able to output not
just RGB video signals but also an NTSC-525 colour composite video
signal, which can be fed into the video input of any NTSC-525 colour
TV received.
So there we are -- the PC's somewhat odd timer tick frequency of
about 18.2 ticks/s is due to NTSC's choice of colour subcarrier
frequency for 525-lines, plus IBM's desire to keep down the number of
quartz crystals on the original IBM PC.
Paul Schlyter
> Paul Schlyter schrieb:
>
>> Yup, there are lots of funny legacies in the PC hardware. Do you for
>> instance know why the PC timer tick ticks 18.2 times per second? That's
>> a pretty awkward number really, so why did they choose that number?
>> Yes, there's a reason, and if you want to know, I can tell you.
>
> cuntervalue was 2^16 = 65536 ==> 1,19318Mhz/65536 = 18.20675 ms. This
> time was used to generatd the timerinterrupt.
some crystal of a different frequency, which would have produced, say,
10, or 100, timer ticks per second? Choosing a 0.65536 MHz crystal
would have produced 10 ticks/sec, and 6.5536 MHz would have produced
100 ticks/sec. Why wasn't this done? Why 1.19318 MHz?
Paul Schlyter
Steven R. Wheeler <swhe...@xpert.net> wrote:
> Paul Schlyter wrote:
>> Yup, there are lots of funny legacies in the PC hardware. Do you for
>> instance know why the PC timer tick ticks 18.2 times per second? That's
>> a pretty awkward number really, so why did they choose that number?
>> Yes, there's a reason, and if you want to know, I can tell you.
>
>
> It's because at that rate, you get 65536 interrupts in 24 hours for your
> real-time clock.
> in the original PC BIOS, and if it detected a "midnight crossing"
> when you made a BIOS call to read the time, it would adjust the
> time by that amount.
is exactly 65536 PC timer ticks/hour.
Evandro Menezes
<3847A94F...@mapag.de>:
>The PIT (8253) in the first PC had an 1,19318Mhz crystal.
Actually, there was the system clock crystal at 14.31818MHz, or 4x
NTSC chroma frequency for the color display, which was divided by 3 by
the 8284 clock chip that fed the CPU originally with 4.77MHz. This
signal was then divided by 4 with 2 flip-flops and was fed to the 8253
timer chip.
HTH
Mike Albaugh
: Steven R. Wheeler <swhe...@xpert.net> wrote in message
: 3847CC...@xpert.net...
: > Paul Schlyter wrote:
: > > Yup, there are lots of funny legacies in the PC hardware. Do you for
: > > instance know why the PC timer tick ticks 18.2 times per second? That's
: > > a pretty awkward number really, so why did they choose that number?
: > > Yes, there's a reason, and if you want to know, I can tell you.
: >
: >
: > It's because at that rate, you get 65536 interrupts in 24 hours for your
: That actually gives 65536 interrupts in 1 hour, and an about 9.6 seconds
: discrepancy on a 24-hour basis. Anyway, while they were at it, why did
: they use a 1.19318 MHz clock to get the 18.2 Hz rate, and not a 1.193046
: MHz clock, which would yield a few hundreths of second error a day? Both
: frequencies seem rather odd to my, admittedly scarce, hardware design
: knowledge...
One has the advantage of being a simple divide-down of a
14.3181MHz crystal that is used in Televison sets, and is thus one
of the very cheapest one can buy. I read somewhere that the nasty
edge-triggered interrupts on the PC were the result of another
"design decision" that saved something like 75 cents. The phrase
"Penny wise and Pound Foolish" springs to mind.
Mike
| alb...@agames.com, speaking only for myself
Holger Petersen
>The "call 5" convention is is still supported in NT4, I just checked.
^^^^^^^
>On the 80286, the addresses did NOT wrap at 1M, so they had to put
>hardware into the motherboard to disable address line A20 to support
But there was no version of Windows-NT on any 286´er out there?!
>the "call 5" convention.
The reasoon for the A20-Gate hase to be found elsewhere... Some rumor:
- dBase
- some BIOS-routine (I doubt, could have been debugged...)
- Microsoft (Disk-) BASIC
just guessing, Holger
bill_h
IBM PC Tech Reference Manual (First Edition, Aug 1981) on eBay.
Actually, I've been going through the 'pile' and found a few
others including:
http://cgi.ebay.com/aw-cgi/eBayISAPI.dll?ViewItem&item=212983522
''Fumbling The Future'', the inside story about the invention
of (arguably) the FIRST personal computer, the Xerox Alto. And
the total failure of the company to capitalize on the creative
work done at PARC (Palo Alto Research Center).
How about this for a hoot: The MIT ''A FORTRAN Coloring Book'',
Professor Kaufman's attempt to make Fortran programming FUN?
>From MIT Press, 1978 (mine's 6th printing, 1981). While I'm
at it I think I'll dig out my PDP-11 Fortran book, too.
Okay, 'nother topic - MITS. I've found TWO books that go into
the early days in Albuquerque, first David Bunnell's ''Personal
Computing'', 1978; and second, Forrest Mims' ''Siliconnections''
(also about 1978 or '79). Does anybody know any other EARLY
sources (books) with info about the place or the people? Of
course, there were several magazine articles, too.
If those guys will be in Albuquerque in Feb or March, (25th
Anniversary of the Jan 1975 ''announcement'' that sent Gates
out there) I'm hoping to get mine autographed - hey, I gotta
collect more than just old hardwares and softwares!
Will there be an area set aside for swapping S-100 boards and stuff?
Anybody in Albuquerque monitoring this 'group know any plans yet?
Bill
Tucson
Steven R. Wheeler
> >> Yup, there are lots of funny legacies in the PC hardware. Do you for
> >> instance know why the PC timer tick ticks 18.2 times per second? That's
> >> a pretty awkward number really, so why did they choose that number?
> >> Yes, there's a reason, and if you want to know, I can tell you.
> >
>
>
> >discrepancy in the original PC BIOS, and if it detected a "midnight
> >crossing" when you made a BIOS call to read the time, it would adjust
> >the time by that amount.
> >
> "65536 interrupts per hour" ;-)
Mea culpa. Mea maxima culpa. In extenuation, I wrote that BIOS over a
decade ago,
and was writing late at night (at least, I was tired enough that it
seemed late!).
- wheels
Charles Richmond
>
> [snip...] [snip...] [snip...]
>
> Thus it appears that the frequency of the PC's timer tick was
> selected in order to use the same quartz crystal to also generate an
> NTSC-525 colour subcarrier frequency by simply feeding the timer chip
> frequency into a x3 frequency multiplier. The good ol' CGA
> video/graphics cards are, as you may remember, able to output not
> just RGB video signals but also an NTSC-525 colour composite video
> signal, which can be fed into the video input of any NTSC-525 colour
> TV received.
>
> So there we are -- the PC's somewhat odd timer tick frequency of
> about 18.2 ticks/s is due to NTSC's choice of colour subcarrier
> frequency for 525-lines, plus IBM's desire to keep down the number of
> quartz crystals on the original IBM PC.
>
IBM wanted to minimize the number of crystals on the board. But
also this same crystal was used in many, many color TV sets. Thus
it was plentiful (i.e. common), readily available, and probably cheaper.
(I have heard of this crystal refered to as the "color burst crystal"
in relation to TV, but I *never* knew why.)
--
+-------------------------------------------------------------+
| Charles and Francis Richmond <rich...@plano.net> |
+-------------------------------------------------------------+
Paul Schlyter
Peter Seed <pet...@astondes.com> wrote:
>
>Steven R. Wheeler wrote in message <3847CC...@xpert.net>...
>>Paul Schlyter wrote:
>>> Yup, there are lots of funny legacies in the PC hardware. Do you for
>>> instance know why the PC timer tick ticks 18.2 times per second? That's
>>> a pretty awkward number really, so why did they choose that number?
>>> Yes, there's a reason, and if you want to know, I can tell you.
>>
>>Having written a BIOS, I know this one, teacher!
>
>Wow, what BIOS is that, so I don't have to use it!
>
>>It's because at that rate, you get 65536 interrupts in 24 hours for your
>>real-time clock. Actually, IIRC, there was something like a 10-second
>>discrepancy in the original PC BIOS, and if it detected a "midnight
>>crossing" when you made a BIOS call to read the time, it would adjust
>>the time by that amount.
>>
>Let's see, 65536 interrupts in 86400 seconds is one every 1.32 seconds. Try
>"65536 interrupts per hour" ;-)
Ooops -- didn't see that, I read it as if he'd written "65536 per hour".
Anyway, even 65536/hour isn't quite right -- it's 65543/hour.
Mark Zenier
Paul Schlyter <pau...@merope.saaf.se> wrote:
>
>PC's and NTSC-525 (NTSC-M) or "Why does the PC tick 18.2 times/second?"
>=======================================================================
>which is exactly the same as the NTSC-M (525 lines) colour subcarrier
>frequency of 3579545 Hz.
>
>selected in order to use the same quartz crystal to also generate an
>NTSC-525 colour subcarrier frequency by simply feeding the timer chip
>frequency into a x3 frequency multiplier.
No need for a multiplier. The 14.31818 MHz clock, +OSC, is pin B30
on the I/O Channel (ie. the ISA bus) connector.
Mark Zenier mze...@eskimo.com mze...@netcom.com Washington State resident
Mario Klebsch
>the A20 feature is helpful for the CALL 5 too, but it has mainly
>other reasons. It allows the use of the 65kB memory above the 1 MB
>boundary.
Using a siutable processor mode would have enabled the use of the
entire 16MByte address space on the 286. But sadly, the 286 version of
MS/DOS was never done.
73, Mario
--
Mario Klebsch ma...@klebsch.de
Barry Watzman
MEMORY is defined as you suggested, but this is NOT stored in, nor is it part
of, a PROGRAM. A CP/M .COM program is loaded at 100H and the program does
NOT contain what you called the PSP (which is purely an MS-DOS structure, by
the way, and not a part of CP/M). The operating system does maintain the
calls, jumps and FCB's between 0h and 100h (the first 256 bytes), but they
are maintained by the program and are not part of or loaded with a CP/M
".COM" file.
Evandro Menezes wrote:
> On Sun, 28 Nov 1999 11:53:29 -0500, Barry Watzman <Wat...@neo.rr.com>
> wrote in <38415E08...@neo.rr.com>:
>
> >Huh ???? a .COM file in CP/M HAS NO
> >DEFINED STRUCTURE, AT ALL. It is simply loaded, arbitrarily at 100H and
> >control is transferred to 100H. Period. End of structure.
>
> As a matter of fact, the first 256 bytes of a program memory space was
> a structure known as the PSP and has always been fully documented,
> with its FCB's and command line buffer, etc. It's still present in
> Windows 3.11 and in the DOS box in Windows Millennium and 2000...
John Elliott
>Using a siutable processor mode would have enabled the use of the
>entire 16MByte address space on the 286. But sadly, the 286 version of
>MS/DOS was never done.
Wasn't that because OS/2 1.x was going to take over?
Barry Watzman
a "color burst" of 8 cycles of 3.579545 MHz added to the "back porch" of the
horizontal sync pulse. This is used to sync the color subcarrier local
oscillator of the TV receiver prior to the start of each scan line. You can
see this clearly if you look at a composite color video signal with an
oscilloscope. You are correct that these crystals are extremely common and
cheap, hundreds of millions of them are made every year.
Barry Watzman
Charles Richmond wrote:
> Paul Schlyter wrote:
> >
> > [snip...] [snip...] [snip...]
> >
> > Thus it appears that the frequency of the PC's timer tick was
> > selected in order to use the same quartz crystal to also generate an
> > NTSC-525 colour subcarrier frequency by simply feeding the timer chip
Paul Schlyter
Charles Richmond <rich...@plano.net> wrote:
> Paul Schlyter wrote:
>>
>> [snip...] [snip...] [snip...]
>>
>> Thus it appears that the frequency of the PC's timer tick was
>> selected in order to use the same quartz crystal to also generate an
>> NTSC-525 colour subcarrier frequency by simply feeding the timer chip
>> frequency into a x3 frequency multiplier. The good ol' CGA
>> video/graphics cards are, as you may remember, able to output not
>> just RGB video signals but also an NTSC-525 colour composite video
>> signal, which can be fed into the video input of any NTSC-525 colour
>> TV received.
>>
>> So there we are -- the PC's somewhat odd timer tick frequency of
>> about 18.2 ticks/s is due to NTSC's choice of colour subcarrier
>> frequency for 525-lines, plus IBM's desire to keep down the number of
>> quartz crystals on the original IBM PC.
>
> There is something that is *not* begin pointed out here. Yes,
> IBM wanted to minimize the number of crystals on the board. But
> also this same crystal was used in many, many color TV sets. Thus
> it was plentiful (i.e. common), readily available, and probably cheaper.
this text, which I plan to later convert to HTML and make available
on my web site (it's about (analogue) TV systems in general, including
now obsolete systems such as the French 819 line system or the English
405 line system). If you're interested, look for it in a month or two.
And this crystal frequency wasn't only selected in PC's but in other
electronic equipment too. Consider for instance the standard ISO
7816-3 for smartcards, which specifies that the clock frequency into
the smart card should be divided by 372 to get the bps rate of the
I/O to/from the smartcard. The factor 372 is pretty awkward as a
dividing factor, so why was it selected? Because the NTSC color
subcarrier frequency, 3579545 Hz, divided by 372 becomes 9622.432...
, which is close enough to 9600 to be used as a base for the common
bps rate 9600 bps. Thus, by choosing 372 as a divisor, NTSC color
subcarrier crystals could be used as a freuency base in smart card
readers, making them cheaper to produce.
> (I have heard of this crystal refered to as the "color burst crystal"
> in relation to TV, but I *never* knew why.)
receiver synchronize their local color subcarrier oscillators.
At transmission, the color subcarrier itself is suppressed most of
the time - only the sidebands are transmitted. In addition, TWO
color signals are transmitted, phase shifted 90 degrees to one
another -- in both cases the carrier itself is suppressed, and only
the sidebands are transmitted. The net result will be a subcarrier
whose amplitude signals the color saturation and whose phase signals
the color hue. This method is known as Quadrature Modulation. If
the "color" is neutral (white, gray, or black), the color subcarrier
will vanish completely at that region in the picture.
So how does the receiver get any phase reference, so it correctly can
convert the phase of the color subcarrier into a hue? Here's where
the color burst enters the picture: Once every line, after the line
synch pulse but before any image information is transmitted, a short
"burst" of pure color subcarrier is transmitted. The phase of this
"burst" becomes a phase reference to the receiver, which must
remember the phase during the remainder of the line, until the next
line when another "color burst" is transmitted.
Richard Plinston
: ma...@klebsch.de (Mario Klebsch) wrote:
:>Using a siutable processor mode would have enabled the use of the
:>entire 16MByte address space on the 286. But sadly, the 286 version of
:>MS/DOS was never done.
: Wasn't that because OS/2 1.x was going to take over?
The main reason that there was not a 286 version of DOS, ie
an MS-DOS compatible OS that could utilise 286 memory space
was that the 286 was brain-dead and MS-DOS was brain-dead in
a different way.
The main issue is that MS-DOS and most programs do segment
arithmetic and 'normalisation'. This relies on each segment
register value being a step along actual memory.
The segment model of the 286 is completely different and
segment manipulation just doesn't work (deliberately by
design). Because there are only 8192 (or so) selectors
it is impossible to emulate a 8086 by remapping these.
Once it is accepted that it is impossible to run 8086
programs in 286 mode (except in the penalty box) it
is realised that programs need to be completely
rewritten to work in 286 mode, such as to run as
native Unix-286 or OS/2 1.x or Windows standard mode.
This is not to say that MS (and others) didn't try to
make a 286 MS-DOS. MS-DOS 4.0 (the real one based on 3.1
not the much later 4.01) was a multitasking version of
DOS that ran in 8086 mode. This was supposed to lead to
a 286 version but 4.1 was still 8086 and this was
abandoned when OS/2 development started.
(MS-DOS 4.0 and 4.1 were known as 'European DOS' and
were very limited in distribution).
--
Paul Schlyter
Mark Zenier <mze...@netcom.com> wrote:
> In article <828p23$ra7$1...@merope.saaf.se>,
> Paul Schlyter <pau...@merope.saaf.se> wrote:
>
>>PC's and NTSC-525 (NTSC-M) or "Why does the PC tick 18.2 times/second?"
>>=======================================================================
> ...
>> which is exactly the same as the NTSC-M (525 lines) colour subcarrier
>> frequency of 3579545 Hz.
>>
>> selected in order to use the same quartz crystal to also generate an
>> NTSC-525 colour subcarrier frequency by simply feeding the timer chip
>> frequency into a x3 frequency multiplier.
>
> on the I/O Channel (ie. the ISA bus) connector.
Assaf Sarfati
>
> In article <3847...@aston-server2.astondes.com>,
> Peter Seed <pet...@astondes.com> wrote:
> >
> >Steven R. Wheeler wrote in message <3847CC...@xpert.net>...
> >>Paul Schlyter wrote:
> >>> Yup, there are lots of funny legacies in the PC hardware. Do you for
> >>> instance know why the PC timer tick ticks 18.2 times per second? That's
> >>> a pretty awkward number really, so why did they choose that number?
> >>> Yes, there's a reason, and if you want to know, I can tell you.
> >>
> >>Having written a BIOS, I know this one, teacher!
> >
> >Wow, what BIOS is that, so I don't have to use it!
> >
> >>It's because at that rate, you get 65536 interrupts in 24 hours for your
> >>real-time clock. Actually, IIRC, there was something like a 10-second
> >>discrepancy in the original PC BIOS, and if it detected a "midnight
> >>crossing" when you made a BIOS call to read the time, it would adjust
> >>the time by that amount.
> >>
> >Let's see, 65536 interrupts in 86400 seconds is one every 1.32 seconds. Try
> >"65536 interrupts per hour" ;-)
>
> Ooops -- didn't see that, I read it as if he'd written "65536 per hour".
>
> Anyway, even 65536/hour isn't quite right -- it's 65543/hour.
>
> --
minimum number of components): the CPU ran at 4.77 MHz (it could run at
5 MHz) because the 808x clock-chip created a x3 clock and divided it by
3 to create a 33% duty-cycle CPU clock; this same 14-odd MHz (it's still
there on the ISA bus!) was divided by 4 and used by the CGA to create
the color-burst signal for NTSC composite video (3.5something MHz).
The CPU's 4.77 MHz clock was divided by 4 and fed to the 8254 timer
(which IIRC couldn't use a clock above 2 Mhz); this 16-bit counter
created the 1/18.2 timer interrupt when it reached Terminal Count. "So
what if it's an odd number? let the S/W guys fix it!"
Regards
Assaf Sarfati
Mario Klebsch
>In comp.os.cpm John Elliott <j...@seasip.demon.co.uk> wrote:
>: ma...@klebsch.de (Mario Klebsch) wrote:
>:>Using a siutable processor mode would have enabled the use of the
>:>entire 16MByte address space on the 286. But sadly, the 286 version of
>:>MS/DOS was never done.
>: Wasn't that because OS/2 1.x was going to take over?
>The main reason that there was not a 286 version of DOS, ie
>an MS-DOS compatible OS that could utilise 286 memory space
>was that the 286 was brain-dead and MS-DOS was brain-dead in
>a different way.
AFAIK, the main difference betwenn the 8086 and the 80286 was, that
the simple formula to calculate the physical address from
segment&offset was changed to a more sophisticated scheme. This should
only affect memory menagement, which should be part of the OS.
It would be no problem to programs, that left the segments alone and
just relied on their OS to do the memory management. Unfortunately, it
was very popular at that time, to do memory management in the
applicatiopn.
IMHO the real problem was, that the programmers did not want to live
in the segmented world, and their desperate trys to escape
segmentation cemented the 1M address space and the 640k
barrier. Typical case of shooting themseklfes into their foot (and not
recognizing it). :-(
Holger Petersen
>> No need for a multiplier. The 14.31818 MHz clock, +OSC, is pin B30
>> on the I/O Channel (ie. the ISA bus) connector.
>
>OK, then a /4 divider will be needed instead, to get 3.579545 MHz
Yes No
There is nothere any 3,5 MHz in an PC.
The one and only Xtal on the original PC-Motherboard was 14.31... MHz.
it was connected to the 8284 "Clock generator and driver for 808X".
That frequency was routed to the ISA-Bus (there it is still causing
grief for CE-compliance :-). The 8284 divided it by 3 to get 4,77 MHz
for the CPU. This frequency was further divided by 4 to get those
1.19,,, MHz input to the Timer.
Yours, Holger
PS: My (Taiwan) CGA Card has no Xtal at all; and I believe, that the
original did neither, as one was supposed to turn an variable Cap
on the motherboard (in the near of the 14,3 MHz Xtal) to adjust
the color on one愀 home TV-Set
Paul Schlyter
> pau...@merope.saaf.se (Paul Schlyter) writes:
>
>>> No need for a multiplier. The 14.31818 MHz clock, +OSC, is pin B30
>>> on the I/O Channel (ie. the ISA bus) connector.
>>
>> OK, then a /4 divider will be needed instead, to get 3.579545 MHz
> Yes No
>
> There is nothere any 3,5 MHz in an PC.
generate a color NTSC composite video signal.
> The one and only Xtal on the original PC-Motherboard was 14.31... MHz.
> it was connected to the 8284 "Clock generator and driver for 808X".
> That frequency was routed to the ISA-Bus (there it is still causing
> grief for CE-compliance :-). The 8284 divided it by 3 to get 4,77 MHz
> for the CPU. This frequency was further divided by 4 to get those
> 1.19,,, MHz input to the Timer.
>
> Yours, Holger
>
> PS: My (Taiwan) CGA Card has no Xtal at all; and I believe, that the
> original did neither, as one was supposed to turn an variable Cap
> on the motherboard (in the near of the 14,3 MHz Xtal) to adjust
As I said above, the CGA card divides the 14.31818 MHz clock by 4
to get the 3.579545 MHz it needs for generating NTSC color composite
video.
Lee Riemenschneider
=>t...@tomra.OBVIOUS.no (Tor Sjowall) writes:
=>
=>
=>>The "call 5" convention is is still supported in NT4, I just checked.
=> ^^^^^^^
=>
=>
=>>On the 80286, the addresses did NOT wrap at 1M, so they had to put
=>>hardware into the motherboard to disable address line A20 to support
=>
=>But there was no version of Windows-NT on any 286´er out there?!
=>
NT gets its roots from OS/2, which originally ran on the '286. At that time
it was a Microsoft and IBM joint venture.
--
===========================================================================
Lee W. Riemenschneider lee.w.riem...@delphiauto.com
Software Engineer OS/2 user PURDUE alumnus and die-hard FAN!
===========================================================================
Grant Edwards
>> (I have heard of this crystal refered to as the "color burst crystal"
>> in relation to TV, but I *never* knew why.)
>
>That's because this frequency is transmitted in bursts, to make the
>receiver synchronize their local color subcarrier oscillators.
[...]
I've said it before and I'll say it again. Color TV will never
work.
The whole color subcarrier thing is bad enough -- but at least
as an EE, I can see that it might have a chance of working. In
the lab. For short periods of time. If you tweak it
constantly.
The whole phosphor-dot/shadow-mask scheme completely absurd,
and doesn't stand a snowball's chance in hell of ever working
reliably.
Who needs color anyway? All the good movies are black and
white...
--
Grant Edwards grante Yow! I'm an East Side
at TYPE...
visi.com
Evandro Menezes
The 1st time I got a 286 data sheet in my hands when I noticed it
could address 16MB I expected to see the physical address calculation
from segments and offsets to be shifted 4 bits to the left (seg << 8 +
off instead of seg << 4 + off)... Wouldn't it have been interesting?
;^)