IBM Obligation to Disclose the Instruction Set
Ze'ev Atlas
Aug 4, 2001
Hi everybody
I've just read an article that implies that IBM obligation to disclose its
hardware instruction set is to be expired soon (if it has not yet expired.)
The article foresaw the demise of Amdahl and Hitachi mainframe clones as a
result of that.
1. Is it true that IBM might not be obligated to disclose the full
instruction set?
2. If so, while I do not worry about Amdahl and Hitachi that might or might
not have cross licensing agreements with IBM, I do worry about the evolution
of Hercules-390 into the zSeries era. Is there any possible remedy. In
essence, should we worry?
3. Is it at all possible to make a pact with IBM that will clarify the
relations between them and our community?
Thanks
Ze'ev Atlas
8:31 pm
RE: IBM Obligation to Disclose the Instruction Set
Fish
Aug 4, 2001
> -----Original Message-----
> From: Ze'ev Atlas [mailto:ZAtlas@...]
> Sent: Saturday, August 04, 2001 8:32 PM
> To: 'hercules-390@yahoogroups.com '
> Subject: [hercules-390] IBM Obligation to Disclose the Instruction Set
>
>
> Hi everybody
> I've just read an article that implies that IBM obligation to disclose its
<snip>
Do you have a URL?
--
"Fish" (David B. Trout)
fish@...
8:58 pm
RE: IBM Obligation to Disclose the Instruction Set
Ze'ev Atlas
Aug 4, 2001
Hi
There is no URL yet, maybe next month. A printed copy is available to Naspa
memebers. It is Mr. Steve Thompson's article: "What Does the z/Architecture
Mean to Mainframe Users?", published in Naspa's Technical Support, June
2001. I assume you are familiar with that publication, Sam Golob is a
regular there. The implication appears in the last paragraph, subtitled
Summary.
If he is correct, we'll depend on the good will of IBM even more than now!
Ze'ev Atlas
-----Original Message-----
From: Fish
To: hercules-390@yahoogroups.com
Sent: 8/4/01 11:58 PM
Subject: RE: [hercules-390] IBM Obligation to Disclose the Instruction Set
> -----Original Message-----
> From: Ze'ev Atlas [mailto:ZAtlas@...]
> Sent: Saturday, August 04, 2001 8:32 PM
> To: 'hercules-390@yahoogroups.com '
> Subject: [hercules-390] IBM Obligation to Disclose the Instruction Set
>
>
> Hi everybody
> I've just read an article that implies that IBM obligation to disclose
its
<snip>
Do you have a URL?
--
"Fish" (David B. Trout)
fish@...
Community email addresses:
Post message: hercules-390@yahoogroups.com
Subscribe: hercules-390-subscribe@yahoogroups.com
Unsubscribe: hercules-390-unsubscribe@yahoogroups.com
List owner: hercules-390-owner@yahoogroups.com
Files and archives at:
http://groups.yahoo.com/group/hercules-390
<http://groups.yahoo.com/group/hercules-390>
Get the latest version of Hercules from:
http://www.conmicro.cx/hercules <http://www.conmicro.cx/hercules>
Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service
<http://docs.yahoo.com/info/terms/> .
10:06 pm
RE: IBM Obligation to Disclose the Instruction Set
jalvo@...
Aug 5, 2001
(Puzzled... Amdahl (Fujitsu) and Hitachi have withdrawn from the
mainframe clone market during the last several years. They support
existing customers but nothing new. So what does the article mean by
predicting an event which has already happened?
john
On Sat, 4 Aug 2001 23:31:33 -0400 , "Ze'ev Atlas" <ZAtlas@...>
wrote:
> Hi everybody
>I've just read an article that implies that IBM obligation to disclose its
>hardware instruction set is to be expired soon (if it has not yet expired.)
>The article foresaw the demise of Amdahl and Hitachi mainframe clones as a
>result of that.
>1. Is it true that IBM might not be obligated to disclose the full
>instruction set?
>2. If so, while I do not worry about Amdahl and Hitachi that might or might
>not have cross licensing agreements with IBM, I do worry about the evolution
>of Hercules-390 into the zSeries era. Is there any possible remedy. In
>essence, should we worry?
>3. Is it at all possible to make a pact with IBM that will clarify the
>relations between them and our community?
>Thanks
>Ze'ev Atlas
>
>
>Community email addresses:
> Post message: hercules-390@yahoogroups.com
> Subscribe: hercules-390-subscribe@yahoogroups.com
> Unsubscribe: hercules-390-unsubscribe@yahoogroups.com
> List owner: hercules-390-owner@yahoogroups.com
>
>Files and archives at:
> http://groups.yahoo.com/group/hercules-390
>
>Get the latest version of Hercules from:
> http://www.conmicro.cx/hercules
>
>Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
6:07 am
RE: IBM Obligation to Disclose the Instruction Set
Ze'ev Atlas
Aug 5, 2001
Hi everybody
John is correct but this was not my main point. I cannot care less about
both Amdahl and Hitachi. My main point was the availability or disclosure
of the full information about the IBM hardware and hardware/firmware
instruction sets as is related to Hercules-390. However, I've just read one
more article (I am catching up with reading this summer) from Mr. Mark
Lilliycorp In Naspa's Technical Support, February 2001. In his column
Industry Insight the article "The Best and Worst of Times?" Mr. Lillycorp
indicates that the aforementioned corporations did indeed leave the
mainframe market and even began to fight against it. He also lamented (my
interpretation) the general fact that this market is not considered dynamic
and not atracting new blood. As a result, IBM is now without competitors in
this market.
This, as we know, may be convinient for the shortsighted monopolist, but it
is a death spell to the market as a whole. IBM had been in this situation
in the past and may or may not be shortsighted. It is possible to point
that to IBM and actually have them in the long run, endorse Hercules-390 as
the "official" competitor. One that they have to provide with full
disclosure in order to prevent being a monopoly, but they should and would
use their marketing muscles to compete against. Except when they license it
for their own educational and demo needs.
It sounds like an unachivable goal, but we should somehow try to convince
IBM at least to negotiate some degree of official recognition.
Ze'ev Atlas
-----Original Message-----
From: jalvo@...
To: hercules-390@yahoogroups.com
Sent: 05/08/01 02:07
Subject: Re: [hercules-390] IBM Obligation to Disclose the Instruction Set
(Puzzled... Amdahl (Fujitsu) and Hitachi have withdrawn from the
mainframe clone market during the last several years. They support
existing customers but nothing new. So what does the article mean by
predicting an event which has already happened?
john
On Sat, 4 Aug 2001 23:31:33 -0400 , "Ze'ev Atlas" <ZAtlas@...>
wrote:
> Hi everybody
>I've just read an article that implies that IBM obligation to disclose
its
>hardware instruction set is to be expired soon (if it has not yet
expired.)
>The article foresaw the demise of Amdahl and Hitachi mainframe clones
as a
>result of that.
>1. Is it true that IBM might not be obligated to disclose the full
>instruction set?
>2. If so, while I do not worry about Amdahl and Hitachi that might or
might
>not have cross licensing agreements with IBM, I do worry about the
evolution
>of Hercules-390 into the zSeries era. Is there any possible remedy.
In
>essence, should we worry?
>3. Is it at all possible to make a pact with IBM that will clarify the
>relations between them and our community?
>Thanks
>Ze'ev Atlas
>
>
>Community email addresses:
> Post message: hercules-390@yahoogroups.com
> Subscribe: hercules-390-subscribe@yahoogroups.com
> Unsubscribe: hercules-390-unsubscribe@yahoogroups.com
> List owner: hercules-390-owner@yahoogroups.com
>
>Files and archives at:
> http://groups.yahoo.com/group/hercules-390
<http://groups.yahoo.com/group/hercules-390>
>
>Get the latest version of Hercules from:
> http://www.conmicro.cx/hercules <http://www.conmicro.cx/hercules>
>
>Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/ <http://docs.yahoo.com/info/terms/>
>
Community email addresses:
Post message: hercules-390@yahoogroups.com
Subscribe: hercules-390-subscribe@yahoogroups.com
Unsubscribe: hercules-390-unsubscribe@yahoogroups.com
List owner: hercules-390-owner@yahoogroups.com
Files and archives at:
http://groups.yahoo.com/group/hercules-390
<http://groups.yahoo.com/group/hercules-390>
Get the latest version of Hercules from:
http://www.conmicro.cx/hercules <http://www.conmicro.cx/hercules>
Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service
<http://docs.yahoo.com/info/terms/> .
7:59 pm
RE: IBM Obligation to Disclose the Instruction Set
stuart@...
Aug 6, 2001
Mr. Atlas:
Let me warn you, I beleive your ignorance of the law in the US
concerning anti-competitive activity, anti-trust, the Sherman Act,
the Lanham Act, the congressional record regarding the legislative
intent of the congressman encacting either of the above entitled
legislative acts, and monopolistic behavior as viewed through the
lense of the Supreme Court of the United States of America, is only
eclipsed by your shear arogance in attempting to proport you are
capable of demonstrating even the most limited knowledge base in the
subject area you raise.
Let us continue disecting your idea from that purview, shall we?
Let us start with the basics; monopolies are NOT ILLEGAL IN THE US!
However, anti-competitive behavior yeilding a monopoly is. What does
this mean?
It means that if you produce a product so well, that no one can
create a competitive product (either due to their lack of engineering
abilities or production abilities) and go out of business, you have
legally done nothing wrong. There is no law in this country against
having a product with no competition in such a scenario.
However, market supressive and anti-competitive behavior can be quite
illegal in certain instances. In order for the Department of Justice
to view IBM in violation of anti-trust legislation, they must be
inhibiting the market in some demonstrable manner.
Once evidence showing the effects of such anti-competitive behavior
is demonstrated to a court of competant jurisdiction, then there can
be an arguement made for temporary restraining orders, permanent
injunctive relief, civil penatilites, charges of corporate criminal
liability, or judicial action to punatively impact the offending
party.
Such evidence must show in real damages (lost revenues, i.e. how you
didnt make money, or were forced to loose it). Now, how you are going
to argue that a product which is free and non-commercial in its
nature (inclusive of the documented goals of it's main developers and
copyright holders being non-commercial), I cant even begin to fathom.
However, even if you were to envision such a viable argument, you
would need the Department of Justice to buy into it, and be willing
to litigate a case on behalf of your damages, as I am sure you know
(being the legal expert you seem to be) that enforcement of the
Sherman Act can only be done by the US Government, and not a private
party.
Absent the above needed constraints you would see no relief until
such time as a court of competent jurisdiction ruled in your favor
with respect to the infringing activities, and this does not take
into consideration the lengthy appeals process.
If all of the above came true (it never would, but let us pretend),
then I doubt IBM would be willing to resist any behavior until
compelled to do so by a Federal Court. Somehow, I dont think you are
appointed to the Federal Bench and in control of adjudicating a case
against IBM, hence, they wont give a rats rectum about what you have
to say.
Hercules is not a viable competitor, you are asking IBM to conspire
with you in creating a "bogus competitor", such that IBM will be able
to escape some imagined wrong (of which there is none) that you think
they are perpatrating. Are you nuts? Do you really think IBM will
just "miss" the gapping whole in any legal theory (and using those
terms "legal theory" with regard to your idea is quite a stretch) you
might be interested in pointing out to them "as a favour"?
On top of all that, after successfully beating their competitors'
divisions into insolvency (in the areas discussed regarding IBM S/390
emulation [in harware]), you wish them now to give away their
patents, copyrights, and trade secrets? Surely you are smart enough
to realize that such patents, copyrights, and trade secrets are worth
more now than they were before.
By the way Mr. Atlas, in case you were wondering, "Mr. Grant" is whom
is buried in Grant's Tomb.
Stuart
Beverly Hills, CA, USA
--- In hercules-390@y..., Ze'ev Atlas <ZAtlas@t...> wrote:
> Hi everybody
> John is correct but this was not my main point. I cannot care less
about
> both Amdahl and Hitachi. My main point was the availability or
disclosure
> of the full information about the IBM hardware and hardware/firmware
> instruction sets as is related to Hercules-390. However, I've just
read one
> more article (I am catching up with reading this summer) from Mr.
Mark
> Lilliycorp In Naspa's Technical Support, February 2001. In his
column
> Industry Insight the article "The Best and Worst of Times?" Mr.
Lillycorp
> indicates that the aforementioned corporations did indeed leave the
> mainframe market and even began to fight against it. He also
lamented (my
> interpretation) the general fact that this market is not considered
dynamic
> and not atracting new blood. As a result, IBM is now without
competitors in
> this market.
> This, as we know, may be convinient for the shortsighted
monopolist, but it
> is a death spell to the market as a whole. IBM had been in this
situation
> in the past and may or may not be shortsighted. It is possible to
point
> that to IBM and actually have them in the long run, endorse
Hercules-390 as
> the "official" competitor. One that they have to provide with full
> disclosure in order to prevent being a monopoly, but they should
and would
> use their marketing muscles to compete against. Except when they
license it
> for their own educational and demo needs.
> It sounds like an unachivable goal, but we should somehow try to
convince
> IBM at least to negotiate some degree of official recognition.
> Ze'ev Atlas
>
> -----Original Message-----
> From: jalvo@m...
> To: hercules-390@y...
> Sent: 05/08/01 02:07
> Subject: Re: [hercules-390] IBM Obligation to Disclose the
Instruction Set
>
> (Puzzled... Amdahl (Fujitsu) and Hitachi have withdrawn from the
> mainframe clone market during the last several years. They support
> existing customers but nothing new. So what does the article mean by
> predicting an event which has already happened?
>
> john
>
>
> On Sat, 4 Aug 2001 23:31:33 -0400 , "Ze'ev Atlas" <ZAtlas@t...>
> wrote:
>
> > Hi everybody
> >I've just read an article that implies that IBM obligation to
disclose
> its
> >hardware instruction set is to be expired soon (if it has not yet
> expired.)
> >The article foresaw the demise of Amdahl and Hitachi mainframe
clones
> as a
> >result of that.
> >1. Is it true that IBM might not be obligated to disclose the full
> >instruction set?
> >2. If so, while I do not worry about Amdahl and Hitachi that
might or
> might
> >not have cross licensing agreements with IBM, I do worry about the
> evolution
> >of Hercules-390 into the zSeries era. Is there any possible
remedy.
> In
> >essence, should we worry?
> >3. Is it at all possible to make a pact with IBM that will
clarify the
> >relations between them and our community?
> >Thanks
> >Ze'ev Atlas
> >
> >
> >Community email addresses:
> > Post message: hercules-390@y...
> > Subscribe: hercules-390-subscribe@y...
> > Unsubscribe: hercules-390-unsubscribe@y...
> > List owner: hercules-390-owner@y...
> >
> >Files and archives at:
> > http://groups.yahoo.com/group/hercules-390
> <http://groups.yahoo.com/group/hercules-390>
> >
> >Get the latest version of Hercules from:
> > http://www.conmicro.cx/hercules
<http://www.conmicro.cx/hercules>
> >
> >Your use of Yahoo! Groups is subject to
> http://docs.yahoo.com/info/terms/
<http://docs.yahoo.com/info/terms/>
> >
>
>
> Community email addresses:
> Post message: hercules-390@y...
> Subscribe: hercules-390-subscribe@y...
> Unsubscribe: hercules-390-unsubscribe@y...
> List owner: hercules-390-owner@y...
>
> Files and archives at:
> http://groups.yahoo.com/group/hercules-390
> <http://groups.yahoo.com/group/hercules-390>
>
> Get the latest version of Hercules from:
> http://www.conmicro.cx/hercules
<http://www.conmicro.cx/hercules>
>
> Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service
> <http://docs.yahoo.com/info/terms/> .
1:56 am
RE: IBM Obligation to Disclose the Instruction Set
peter_flass@...
Aug 6, 2001
And lawyers wonder why the public views them as being
somewhere between used-car salesmen and serial rapists on the
totem pole.
Surely there was a better way to say this...
--- In hercules-390@y..., stuart@b... wrote:
> Mr. Atlas:
>
> Let me warn you, I beleive your ignorance of the law in the
US
> concerning anti-competitive activity, anti-trust, the
Sherman Act,
> the Lanham Act, the congressional record regarding the
legislative
> intent of the congressman encacting either of the above
entitled
> legislative acts, and monopolistic behavior as viewed
through the
> lense of the Supreme Court of the United States of America,
is only
> eclipsed by your shear arogance in attempting to proport
you are
> capable of demonstrating even the most limited knowledge
base in the
> subject area you raise.
>
> Let us continue disecting your idea from that purview,
shall we?
>
5:46 am
RE: IBM Obligation to Disclose the Instruction Set
Ze'ev Atlas
Aug 6, 2001
Hi everybody
From an innocent idea of "there is a potential problem to the Hercules
implementation, so therefore let's try to convivce IBM the it is a good idea
to keep us running", the discussion went to blaming me that my ignorance of
the law in the US... , is only eclipsed by my shear arogance. I am not a
lawyer (and after reading this rant I must say thanks God!) neither am I
arogant. I did not say and did not even mean that we could enforce
non-monopolistic behavior. I of course did not mean that IBM should create
bogus competitor (and after reading Stuart's letter, I am sorry I've even
said something that may be even remotely understood that way) I just said
that we should pleed with IBM that it is of their best interest to keep us
running by providing the subject matter information!
Thanks Peter for answering Stuart.
In any case, nobody yet have discussed the main problem, my opinion, that we
may be closed out of the necessary information to keep Hercules on the right
track.
Ze'ev Atlas
-----Original Message-----
From: peter_flass@... [mailto:peter_flass@...]
Sent: Mon, August 06, 2001 8:46 AM
To: hercules-390@yahoogroups.com
Subject: [hercules-390] Re: IBM Obligation to Disclose the Instruction Set
And lawyers wonder why the public views them as being
somewhere between used-car salesmen and serial rapists on the
totem pole.
Surely there was a better way to say this...
--- In hercules-390@y..., stuart@b... wrote:
> Mr. Atlas:
>
> Let me warn you, I beleive your ignorance of the law in the
US
> concerning anti-competitive activity, anti-trust, the
Sherman Act,
> the Lanham Act, the congressional record regarding the
legislative
> intent of the congressman encacting either of the above
entitled
> legislative acts, and monopolistic behavior as viewed
through the
> lense of the Supreme Court of the United States of America,
is only
> eclipsed by your shear arogance in attempting to proport
you are
> capable of demonstrating even the most limited knowledge
base in the
> subject area you raise.
>
> Let us continue disecting your idea from that purview,
shall we?
>
Community email addresses:
Post message: hercules-390@yahoogroups.com
Subscribe: hercules-390-subscribe@yahoogroups.com
Unsubscribe: hercules-390-unsubscribe@yahoogroups.com
List owner: hercules-390-owner@yahoogroups.com
Files and archives at:
http://groups.yahoo.com/group/hercules-390
<http://groups.yahoo.com/group/hercules-390>
Get the latest version of Hercules from:
http://www.conmicro.cx/hercules <http://www.conmicro.cx/hercules>
Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service
<http://docs.yahoo.com/info/terms/> .
[Non-text portions of this message have been removed]
6:22 am
RE: IBM Obligation to Disclose the Instruction Set
abaddon@...
Aug 6, 2001
--- In hercules-390@y..., "Fish" <fish@i...> wrote:
> > -----Original Message-----
> > From: Ze'ev Atlas [mailto:ZAtlas@t...]
> > Sent: Monday, August 06, 2001 6:22 AM
> > To: 'hercules-390@y...'
> > Subject: RE: [hercules-390] Re: IBM Obligation to Disclose the
> > Instruction Set
> >
> >
> > Hi everybody
> > From an innocent idea of "there is a potential problem to the
Hercules
> > implementation, so therefore let's try to convivce IBM the it is a
good idea
> > to keep us running", the discussion went to blaming me that my
ignorance of
> > the law in the US... , is only eclipsed by my shear arogance. I
am not a
<snip>
Stuart is an asshole IMO. He's one of only three people on this list
that I've
had to killfile.
I'm saying nothing... Qui tacet, consentit.
On topic, I would point out that IBM (presumably) does (and presumably
will continue to) share information with at least ONE other outfit -
Flex. If they ever stopped disclosing the instruction set generally,
they would (presumably) have to continue to disclose to Flex. (a lot
of presumtion here... it's that kind of topic)
Only Flex are not exactly a competitor, as they have a deal with IBM
and serve a market segment which IBM has decided not to serve with
hardware. All of the above is subject to change... but we may find
ourselves in the position some day where we are an open source
competitor/alternative to Flex, as well as IBM... wonder if that would
affect the position re. disclosure?
Mike
11:02 am
RE: IBM Obligation to Disclose the Instruction Set
Graeme Gibson
Aug 6, 2001
>snips<
> On topic, I would point out that IBM (presumably) does (and presumably
> will continue to) share information with at least ONE other outfit -
> Flex. If they ever stopped disclosing the instruction set generally,
> they would (presumably) have to continue to disclose to Flex. (a lot
> of presumtion here... it's that kind of topic)
Hmm. Ok, step back and let's have a look at the forest. IMO it would
be completely laughable to buy a computer without knowing the
instruction set that enables the purchaser to use it. We tend to obsess
a bit about the O/S (I suspect that Herculeans see the OS as *being* the
application) but the real reason that organisations buy these machines
is to run the enormous quantity of in-house written software on them.
Software that cannot be written and maintained without full knowledge of
the instruction set. Think Intel instead of IBM and you'll see how
untenable the idea of withholding the instruction set really is, what a
bad business decision it would be, how much uncertainty it would
create. No hardware vendor wants uncertainty about their product's
accessability, usability, viability.
Furthermore, as a vendor of operating systems for their hardware, IBM
should not be seen to be in the position of using "inside" or secret
information to enable them to compete unfairly with third parties who
might choose to create competitive operating systems for that same IBM
hardware. I think there have been court cases re hardware-to-hardware
interfaces (s/360 channels? tape controllers?) where IBM had to disclose
interface details to third parties. The "instruction set" is simply
another interface where full disclosure is a necessary requirement to
avoid the general class of criticism recently levelled at M/$oft over
*their* developers having access to W*ndows API details that were not
freely available in the same timeframes to third parties who might
choose to build competitive products.
OPEN is the word. It reassures customers about the future, helps them
make plans, encourages them to spend more money on hardware, software,
consultants, programmers.
Graeme.
Disclosure: I work for an ISV whose future depends upon a viable IBM
big iron world -and- knowledge of the instruction sets for a variety of
big and little machines, amongst other things.
5:22 pm
RE: IBM Obligation to Disclose the Instruction Set
jalvo@...
Aug 7, 2001
In the past IBM introduced non-disclosed instructions, real ciscy
ones. One was tree sort supposed to speed up its program product sort.
It wasn't very effective. I knew someone who worked at SyncSort... the
reverse engineered it and used it. whenever it provided a performance
boost [on smaller machines... on larger ones it slowed things down.]
There were MVS assists for early MVS (SP1?) releases.
There were several layers of VM assists, some more microcode-y then
others.
The scheme that runs LPARS is much like microcode... it is a whole
little operating system sort of like a hard-ware assisted VM.
So there is a long history of IBM moving things into the hardware
realm. Hasn't hurt customers, who mostly worry about DB2 and CICS and
MQSeries throughput these days, but it does put pressure on vendors.
What I find most interesting is that Hercules essentially provides a
cheap MIP ( 2*1GZ linux machine, .5 GIG Ram, lots of storage, a Nic or
two costs maybe $4000 for good quality.) That gives maybe 8 390 Mips,
so we are at $500/MIP straight purchase. Does anyone have an
comparison price point? And cost of Intel computing is falling 30+% a
year. Six year from now, $60/MIP. You can't go too far out because
there are some physical barriers looming.
The balance point is that all the interesting software is much more
recent than 25 years ago... so from a legal point of view they are
safe. But it looks to me like a terrible danger is looming... a big
performance gap that will be more and more attractive to customers all
the time. Add to that the fact that most of the 390 programmers will
be retiring (or worse) in the next 10-15 years and I think it is going
to demand a transition to some new world.
john alvord
On Tue, 07 Aug 2001 11:22:23 +1100, Graeme Gibson <graeme@...>
wrote:
>>snips<
>> On topic, I would point out that IBM (presumably) does (and presumably
>> will continue to) share information with at least ONE other outfit -
>> Flex. If they ever stopped disclosing the instruction set generally,
>> they would (presumably) have to continue to disclose to Flex. (a lot
>> of presumtion here... it's that kind of topic)
>
>Hmm. Ok, step back and let's have a look at the forest. IMO it would
>be completely laughable to buy a computer without knowing the
>instruction set that enables the purchaser to use it. We tend to obsess
>a bit about the O/S (I suspect that Herculeans see the OS as *being* the
>application) but the real reason that organisations buy these machines
>is to run the enormous quantity of in-house written software on them.
>Software that cannot be written and maintained without full knowledge of
>the instruction set. Think Intel instead of IBM and you'll see how
>untenable the idea of withholding the instruction set really is, what a
>bad business decision it would be, how much uncertainty it would
>create. No hardware vendor wants uncertainty about their product's
>accessability, usability, viability.
>
>Furthermore, as a vendor of operating systems for their hardware, IBM
>should not be seen to be in the position of using "inside" or secret
>information to enable them to compete unfairly with third parties who
>might choose to create competitive operating systems for that same IBM
>hardware. I think there have been court cases re hardware-to-hardware
>interfaces (s/360 channels? tape controllers?) where IBM had to disclose
>interface details to third parties. The "instruction set" is simply
>another interface where full disclosure is a necessary requirement to
>avoid the general class of criticism recently levelled at M/$oft over
>*their* developers having access to W*ndows API details that were not
>freely available in the same timeframes to third parties who might
>choose to build competitive products.
>
>OPEN is the word. It reassures customers about the future, helps them
>make plans, encourages them to spend more money on hardware, software,
>consultants, programmers.
>
>Graeme.
>
>Disclosure: I work for an ISV whose future depends upon a viable IBM
>big iron world -and- knowledge of the instruction sets for a variety of
>big and little machines, amongst other things.
>
>
>Community email addresses:
> Post message: hercules-390@yahoogroups.com
> Subscribe: hercules-390-subscribe@yahoogroups.com
> Unsubscribe: hercules-390-unsubscribe@yahoogroups.com
> List owner: hercules-390-owner@yahoogroups.com
>
>Files and archives at:
> http://groups.yahoo.com/group/hercules-390
>
>Get the latest version of Hercules from:
> http://www.conmicro.cx/hercules
>
>Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
4:14 am
RE: IBM Obligation to Disclose the Instruction Set
John Summerfield
Aug 6, 2001
>
> There were MVS assists for early MVS (SP1?) releases.
I know there was for DFSORT; it was a bit of a concern to Amdahl
customers but OTOH a faster sort isn't everything.
>
> There were several layers of VM assists, some more microcode-y then
> others.
And for VS1.
>
> The scheme that runs LPARS is much like microcode... it is a whole
> little operating system sort of like a hard-ware assisted VM.
I suspect it's more than a little operating system. When I worked for
Amdahl (middle 80s) I was told the service processors ran Unix. Cut
down doubtless.
>
> So there is a long history of IBM moving things into the hardware
> realm. Hasn't hurt customers, who mostly worry about DB2 and CICS and
> MQSeries throughput these days, but it does put pressure on vendors.
If the hardware does it faster, it's good for the client. If
competitors could duplicate it (and I believe they could) that's good
for the client too.
>
> What I find most interesting is that Hercules essentially provides a
> cheap MIP ( 2*1GZ linux machine, .5 GIG Ram, lots of storage, a Nic or
> two costs maybe $4000 for good quality.) That gives maybe 8 390 Mips,
> so we are at $500/MIP straight purchase. Does anyone have an
> comparison price point? And cost of Intel computing is falling 30+% a
> year. Six year from now, $60/MIP. You can't go too far out because
> there are some physical barriers looming.
>
> The balance point is that all the interesting software is much more
> recent than 25 years ago... so from a legal point of view they are
> safe. But it looks to me like a terrible danger is looming... a big
> performance gap that will be more and more attractive to customers all
> the time. Add to that the fact that most of the 390 programmers will
> be retiring (or worse) in the next 10-15 years and I think it is going
> to demand a transition to some new world.
I worked at a bank a while go. I imagine MVS on every programmers' desktop could
be very attractive.
--
Cheers
John Summerfield
Microsoft's most solid OS: http://www.geocities.com/rcwoolley/
Note: mail delivered to me is deemed to be intended for me, for my disposition.
9:39 pm
RE: IBM Obligation to Disclose the Instruction Set
phil@...
Aug 7, 2001
> > There were several layers of VM assists, some more microcode-y
then
> > others.
>
> And for VS1.
MVS was always IBM's strategic operating system for the big accounts,
were the money was. If you want to research the legal issues back
then, concentrate on the hardware functions that MVS _REQUIRED_ to
run. That was were the fun was - things like Page Fault Assist were
bonuses. But you NEEDED Cross Memory Services, and that's protected
by a HARDWARE patent.
> >
> > The scheme that runs LPARS is much like microcode... it is a whole
> > little operating system sort of like a hard-ware assisted VM.
>
> I suspect it's more than a little operating system
Look at its message prefixes.
2:07 am
RE: IBM Obligation to Disclose the Instruction Set
Jay Maynard
Aug 7, 2001
On Tue, Aug 07, 2001 at 09:07:54AM -0000, phil@...
wrote:
> MVS was always IBM's strategic operating system for the big accounts,
> were the money was. If you want to research the legal issues back
> then, concentrate on the hardware functions that MVS _REQUIRED_ to
> run. That was were the fun was - things like Page Fault Assist were
> bonuses. But you NEEDED Cross Memory Services, and that's protected
> by a HARDWARE patent.
By now, I suspect that patent has expired; Cross Memory Services was
introduced before 1984, and at that point, patent duration was 17 years from
issuance.
Whether Hercules is infringing IBM patents or not is, under US patent law,
not a question a layman is competent to answer. (Strange but true...but
people have been held liable for intentional infringement for looking at
patents that were ruled to have been infringed, even though they claimed
that they believed they were not infringing those patents! The court held
that only a patent attorney is competent to have an opinion on whether or
not a patent is likely to be infringed by a particular device or program.) I
have not looked at any IBM patents, nor will I, and unless and until IBM
notifies me (or any other Hercules developer) that it believes that Hercules
is infringing and demands that we cease, I refuse to worry about it.
This discussion has gone on long enough, I believe, and it's clear by now
that nobody's going to convince anyone else of their position. I believe
that further discussion here is pointless unless someone has something truly
new to add.
5:10 am
Copyright 2001