> Nope, he meant to say it is made in king county, perhaps !!!
Oh, shit, I forgot that we are dealing with a future emperor here…
Anton Bassov
Peter,
The Windows IO model does NOT derive from microkernel architecture. It
derives from VMS (released 1978), which derives from RSX-11M (released
1973… after a whopping 18 month development effort!). The IO model (with
IRPs being forwarded from device to device, and with DPCs being queued as an
alternative to running at DIRQL) in these operating systems is VERY similar
to that of Windows NT. I can state this as a fact, having written kernel
mode code for all of them.
Excellent. Now we can look forward to extending this ‘discussion’ to the
relative merits of RT-11 Fork Routines, why VMS was (insert favorite
adjective), TOPS-20 vs. TOPS-10, whether or not the 9-bit byte is better
than the 8-bit byte, if the Meta-Key is the great-grandmother of the
Windows-Key, yadda, yadda, yadda.
Personally, I miss my 11/73 and MicroVAXII. But I am not going to miss this
thread…
-dave
I realize that you are a superior architect to Cutler, and thus I’m sure… somewhere somehow… you have a track record or credentials or SOMEthing to back that up. Other than hot air, that is.
Well, despite what you label as the very obvious " cumbersome and inefficient" nature of Cutler’s I/O Subsystem designs, those operating systems have managed to limp along,and become 2 or 3 of the most successful in history and – without question – supporting the broadest range of devices ever supported.
Amazing, given such a bad design.
I am arguing that decomposition of functionality into individual functional layers with well-defined boundaries is a long-standing engineering principal, and while it is a principal that is USED in microkernel-based operating systems, it is not a principal RESTRICTED to microkernel-based operating systems.
TOPS-20 vs. TOPS-10
[/quote]
TOPS-10! DEFINITELY. And especially running on a DecSystem10 using Emitter Coupled Logic. Whoooo the power supply on the sucker!
I have an 11/24 in my basement. It runs RSX-11M on dual RL02s. I haven’t fired it up in over 20 years. I’ll sell it to anybody who’d like to buy it, for… oh… only HALF of what it cost in 1983.
I never liked the MicroVAX series. Too darn slow.
Peter
OSR
I was not clear enough …
Redmond,WA. is in king county of that state …
It’s not an attack to anyone by any means …
-pro
> Nope, he meant to say it is made in king county, perhaps !!!
Oh, shit, I forgot that we are dealing with a future emperor here…
Anton Bassov
NTDEV is sponsored by OSR
For our schedule of WDF, WDM, debugging and other seminars visit:
http://www.osr.com/seminarsTo unsubscribe, visit the List Server section of OSR Online at
http://www.osronline.com/page.cfm?name=ListServer
I have an 11/24 in my basement. It runs RSX-11M on dual RL02s.
Oh fancy you! How ‘new-school’. I would have pegged you for a 11/40 with
RK05s and the vector graphics unit. (I wrote my first attempt at
understanding Foley & van Dam SGP in Pascal to render the game of ‘Life’ on
one of those. Ah the wonder of it all.)
I never liked the MicroVAX series. Too darn slow.
Dare we say the ‘Yugo’ of the VAX family 
Ok, I have had my fun. It’s Friday. I’ll let you and Anton finish the
cage match while this east-coaster heads for a Guinness. I can’t stand to
watch anymore. Make sure that one of you mops up and shuts out the lights.
Cheers,
-dave
That’s the best post in this thread… my own included, of course.
Peter
OSR
Peter,
I realize that you are a superior architect to Cutler,
Well, judging from your numerous posts on the subject, Cutler seems to be just your idol…
I just wonder what you are going to do if, at some point, Cutler changes his opinion on this or that topic.
For example, he may decide that that fork -exec sequence that is used by all UNIX-like systems is, actually, better than spawning; or, probably, he may abandon his idea of uniform IO model and decide that different types of devices should have different interfaces that are appropriate for a given device class; or, probably, he may decide that virtual file system is a wonderful idea; or, to frighten you even more, he may look at Linux kernel sources some day and get impressed with some of its parts (for example, he may fall in love with a radix-tree-based mechanism that , given a PFN, allows you to quickly locate everyone who happens to map a given page into its address space, i.e. something that Windows kernel is unable to do).
What are you going to do then???
I am arguing that decomposition of functionality into individual functional layers with
well-defined boundaries is a long-standing engineering principal,
Actually, I don’t really know how the above statement is related to our discussion of Windows IO model.
If you don’t mind, could you please explain to us why, in your opinion, the above principle would be violated if, instead of sending IRPs to one another, drivers just registered their callbacks with the appropriate subsystem the way it is done under Linux, as well as under some Windows subsystems (NDIS is the very first example that comes to my mind)…
Anton Bassov
Anton,
I’m not the kind of person to suggest anyone anything here, but sorry …
This is not nice by any means OKAY???
Everyone playing around kernel mode day-in-and-day-out are giving you the
options ( try it out then come back ), show them what you got … Isn’t it
enough of a respect they are giving you???
If you have alternative means, use it. I don’t see any point in telling
Doran with a butt-load of suggestion and arguing with him. Show him your
work, you will be an architect in MS… If you really really hate the guts
of MS ( for one reason or other), then things ends here… Hang on in
Linux kernel threads … why fight here???
Really this is not good…
-pro
Peter,
> I realize that you are a superior architect to Cutler,
Well, judging from your numerous posts on the subject, Cutler seems to be
just your idol…I just wonder what you are going to do if, at some point, Cutler changes
his opinion on this or that topic.For example, he may decide that that fork -exec sequence that is used by
all UNIX-like systems is, actually, better than spawning; or, probably,
he may abandon his idea of uniform IO model and decide that different
types of devices should have different interfaces that are appropriate for
a given device class; or, probably, he may decide that virtual file system
is a wonderful idea; or, to frighten you even more, he may look at Linux
kernel sources some day and get impressed with some of its parts (for
example, he may fall in love with a radix-tree-based mechanism that ,
given a PFN, allows you to quickly locate everyone who happens to map a
given page into its address space, i.e. something that Windows kernel is
unable to do).What are you going to do then???
> I am arguing that decomposition of functionality into individual
> functional layers with
> well-defined boundaries is a long-standing engineering principal,Actually, I don’t really know how the above statement is related to our
discussion of Windows IO model.
If you don’t mind, could you please explain to us why, in your opinion,
the above principle would be violated if, instead of sending IRPs to one
another, drivers just registered their callbacks with the appropriate
subsystem the way it is done under Linux, as well as under some Windows
subsystems (NDIS is the very first example that comes to my mind)…Anton Bassov
NTDEV is sponsored by OSR
For our schedule of WDF, WDM, debugging and other seminars visit:
http://www.osr.com/seminarsTo unsubscribe, visit the List Server section of OSR Online at
http://www.osronline.com/page.cfm?name=ListServer
Can we tone back the personal attacks please? You seem to really have it in for Doron and Peter recently, for whatever reason.
Facts and civil discussion to try and get your opinion across are one thing, but dropping to that level is really quite another, and not very appropriate for this forum, I think.
- S
-----Original Message-----
From: xxxxx@lists.osr.com [mailto:xxxxx@lists.osr.com] On Behalf Of xxxxx@hotmail.com
Sent: Friday, September 05, 2008 7:37 PM
To: Windows System Software Devs Interest List
Subject: RE:[ntdev] Microkernel discussion continued…
Peter,
I realize that you are a superior architect to Cutler,
Well, judging from your numerous posts on the subject, Cutler seems to be just your idol…
I just wonder what you are going to do if, at some point, Cutler changes his opinion on this or that topic.
For example, he may decide that that fork -exec sequence that is used by all UNIX-like systems is, actually, better than spawning; or, probably, he may abandon his idea of uniform IO model and decide that different types of devices should have different interfaces that are appropriate for a given device class; or, probably, he may decide that virtual file system is a wonderful idea; or, to frighten you even more, he may look at Linux kernel sources some day and get impressed with some of its parts (for example, he may fall in love with a radix-tree-based mechanism that , given a PFN, allows you to quickly locate everyone who happens to map a given page into its address space, i.e. something that Windows kernel is unable to do).
What are you going to do then???
I am arguing that decomposition of functionality into individual functional layers with
well-defined boundaries is a long-standing engineering principal,
Actually, I don’t really know how the above statement is related to our discussion of Windows IO model.
If you don’t mind, could you please explain to us why, in your opinion, the above principle would be violated if, instead of sending IRPs to one another, drivers just registered their callbacks with the appropriate subsystem the way it is done under Linux, as well as under some Windows subsystems (NDIS is the very first example that comes to my mind)…
Anton Bassov
NTDEV is sponsored by OSR
For our schedule of WDF, WDM, debugging and other seminars visit:
http://www.osr.com/seminars
To unsubscribe, visit the List Server section of OSR Online at http://www.osronline.com/page.cfm?name=ListServer
Prokash,
I don’t see any point in telling Doran with a butt-load of suggestion and arguing with him.
Sorry, but he asked about it himself - he asked me how, in my opinion, could they achieve their goal
of allowing stacks with multiple third-party drivers that have to interact with one another had not they used their current IO model that is based upon IRPs. The subtle message of his post was “this would be impossible” Therefore, I just showed him the very first possibility that got into my head…
why fight here???
Well, apparently this is a bit selfish approach, but to be honest, the only reason why I do it is because, in technical terms, I learn an AWFUL LOT of general OS-design-related stuff (i.e. something I am really interested in) this way…
Anton Bassov
Ken,
Can we tone back the personal attacks please? You seem to really have it in
for Doron and Peter recently, for whatever reason.
Please note that it has nothing to do personally either with Peter or with Doron - what I DO attack is the very attitude " it was done by Cutler (Torvalds, Ritchie, etc) himself, and, hence, is something absolute that cannot be questioned, let alone argued about". This is nothing more than just a logical fallacy known as “argumentum ad verecundiam” (appeal to authority), and somehow it happened that both Peter and Doron used “ad verecundiam” arguments in their recent couple of posts…
Anton Bassov
Actually, it seemed to me like Doron was spending his spare time responding to confrontational mailinglist postings with his views on the advantages and disadvantages of how and why the I/O manager structures driver calls.
I suppose that’s just my interpretation though
- S
-----Original Message-----
From: xxxxx@lists.osr.com [mailto:xxxxx@lists.osr.com] On Behalf Of xxxxx@hotmail.com
Sent: Friday, September 05, 2008 9:37 PM
To: Windows System Software Devs Interest List
Subject: RE:[ntdev] Microkernel discussion continued…
Ken,
Can we tone back the personal attacks please? You seem to really have it in
for Doron and Peter recently, for whatever reason.
Please note that it has nothing to do personally either with Peter or with Doron - what I DO attack is the very attitude " it was done by Cutler (Torvalds, Ritchie, etc) himself, and, hence, is something absolute that cannot be questioned, let alone argued about". This is nothing more than just a logical fallacy known as “argumentum ad verecundiam” (appeal to authority), and somehow it happened that both Peter and Doron used “ad verecundiam” arguments in their recent couple of posts…
Anton Bassov
NTDEV is sponsored by OSR
For our schedule of WDF, WDM, debugging and other seminars visit:
http://www.osr.com/seminars
To unsubscribe, visit the List Server section of OSR Online at http://www.osronline.com/page.cfm?name=ListServer
Well then try your best to get on to pure kernel development in any form
( Windows, Linux, MAC, symbian, QNX …). Get the ball rolling!
Becoming violent does not help! Otherwise, you would be wrapped up
insulting people and hurting yourself. And this applies to anyone, not
just you. Meanwhile we all try to learn from each other!!!
-pro
xxxxx@hotmail.com wrote:
Prokash,
> I don’t see any point in telling Doran with a butt-load of suggestion and arguing with him.
>Sorry, but he asked about it himself - he asked me how, in my opinion, could they achieve their goal
of allowing stacks with multiple third-party drivers that have to interact with one another had not they used their current IO model that is based upon IRPs. The subtle message of his post was “this would be impossible” Therefore, I just showed him the very first possibility that got into my head…> why fight here???
>Well, apparently this is a bit selfish approach, but to be honest, the only reason why I do it is because, in technical terms, I learn an AWFUL LOT of general OS-design-related stuff (i.e. something I am really interested in) this way…
Anton Bassov
NTDEV is sponsored by OSR
For our schedule of WDF, WDM, debugging and other seminars visit:
http://www.osr.com/seminarsTo unsubscribe, visit the List Server section of OSR Online at http://www.osronline.com/page.cfm?name=ListServer
That is pretty much it. It was not to discredit, disparage, or dismiss the linux or unix driver model. While the irp model does have its warts, it does have its advantages as well, but anton made a blanket statement that irps suck and blow chunks and I was trying to point out some of the advantages that irps have (and have today, not in some theoretical “if I were to do it” land). It all depends on who you want to target and what your goals of the design are. Uber simplicity is great, but extensibility is also a consideration to be made, esp if you want an ecosystem built around your software (free or not).
So, anton, you don’t need windows source to prove that you have the chops to make this all work. You have the linux kernel and all the drivers in it to modify at your heart’s content. That is an awesome playground to party in. why not create a callback based, layered driver model that does sync and async io in linux to prove your point. Hell, the end result easily be that the current linux model is just fine and all that extensibility jazz is not worth it.
Having owned up 3 different stacks in windows, I have created mixed hybrid models which have both irps and callback, with some interfaces sync and async. I had ideal “if I were to do it” designs that met with reality and practicality and were altered accordingly. There is nothing like putting your money where your mouth is and just doing it. So, again, I invite you to go do it (and don’t put up fake barriers that blame me, msft or anyone else for blocking you) and tell us how you did. I am truly interested in the outcome, I am sure you will have a lot to tell us.
Have a good weekend.
d
-----Original Message-----
From: xxxxx@lists.osr.com [mailto:xxxxx@lists.osr.com] On Behalf Of Skywing
Sent: Friday, September 05, 2008 6:56 PM
To: Windows System Software Devs Interest List
Subject: RE: RE:[ntdev] Microkernel discussion continued…
Actually, it seemed to me like Doron was spending his spare time responding to confrontational mailinglist postings with his views on the advantages and disadvantages of how and why the I/O manager structures driver calls.
I suppose that’s just my interpretation though
- S
-----Original Message-----
From: xxxxx@lists.osr.com [mailto:xxxxx@lists.osr.com] On Behalf Of xxxxx@hotmail.com
Sent: Friday, September 05, 2008 9:37 PM
To: Windows System Software Devs Interest List
Subject: RE:[ntdev] Microkernel discussion continued…
Ken,
Can we tone back the personal attacks please? You seem to really have it in
for Doron and Peter recently, for whatever reason.
Please note that it has nothing to do personally either with Peter or with Doron - what I DO attack is the very attitude " it was done by Cutler (Torvalds, Ritchie, etc) himself, and, hence, is something absolute that cannot be questioned, let alone argued about". This is nothing more than just a logical fallacy known as “argumentum ad verecundiam” (appeal to authority), and somehow it happened that both Peter and Doron used “ad verecundiam” arguments in their recent couple of posts…
Anton Bassov
NTDEV is sponsored by OSR
For our schedule of WDF, WDM, debugging and other seminars visit:
http://www.osr.com/seminars
To unsubscribe, visit the List Server section of OSR Online at http://www.osronline.com/page.cfm?name=ListServer
NTDEV is sponsored by OSR
For our schedule of WDF, WDM, debugging and other seminars visit:
http://www.osr.com/seminars
To unsubscribe, visit the List Server section of OSR Online at http://www.osronline.com/page.cfm?name=ListServer
I find it quite ironic that you are complaining about “argumentum ad
verecundiam” when it was you yourself that stated:
I guess that before he starts arguing “why Linus sucks” he has to write
his own OS and to give its source code to the world for free.
So, you are saying that you can’t appeal to authority for matters of
argument in fact you just can’t argue unless you are one of the handful of
first hand authorities. Oh unless your name is Anton I guess?
Bill M.
wrote in message news:xxxxx@ntdev…
> Ken,
>
>> Can we tone back the personal attacks please? You seem to really have it
>> in
>> for Doron and Peter recently, for whatever reason.
>
> Please note that it has nothing to do personally either with Peter or with
> Doron - what I DO attack is the very attitude " it was done by Cutler
> (Torvalds, Ritchie, etc) himself, and, hence, is something absolute that
> cannot be questioned, let alone argued about". This is nothing more than
> just a logical fallacy known as “argumentum ad verecundiam” (appeal to
> authority), and somehow it happened that both Peter and Doron used “ad
> verecundiam” arguments in their recent couple of posts…
>
>
> Anton Bassov
>
> I find it quite ironic that you are complaining about “argumentum ad verecundiam”
when it was you yourself that stated: " I guess that before he starts arguing “why Linus sucks”
he has to write his own OS and to give its source code to the world for free".
This is not an “argumentum ad verecundiam” . Why??? Simply because the statement “Linus sucks” is about Linus Torvalds himself, rather than about any of his technical ideas. It would be the one if I used this argument while defending his model or his implementation. However, I was defending Linus Torvalds himself. Similarly, Peter’s statement would not be “argumentum ad verecundiam” at all if, instead of criticizing the shortcoming of Windows IO model, I was criticizing Dave Cutler in person. However, I was not doing it, was I …
Anton Bassov
Doron,
So, anton, you don’t need windows source to prove that you have the chops to make this all work.
You have the linux kernel and all the drivers in it to modify at your heart’s content. That is an
awesome playground to party in. why not create a callback based, layered driver model that does
sync and async io in linux to prove your point.
…
So, again, I invite you to go do it (and don’t put up fake barriers that blame me, msft or
anyone else for blocking you) and tell us how you did. I am truly interested in the outcome,
Well, indeed, it sounds pretty exciting…
At the same time, it seems like a bit too much work for one person, don’t you think? Therefore, if you are “truly interested in the outcome”, I invite you to participate in it. Let’s split out efforts the following way - I will write a driver model the way I see it, and you will write demo stack(s) and app(s) to actually take advantage on it. This model will peacefully coexist with the existing one - all one needs to do in order to try something like that under Linux is to register a special filesystem with VFS. At this point, the top of our stack will become accessible to the standard read/write/ioctls/etc so that apps will be able to deal with our devices, and what we do inside of it is our own business, so that we can easily put an entire layered driver model there. For the sake of the experiment I think it makes sense to abstract ourselves from dealing with the actual hardware (at least at the original stage) - instead, let’s create an imaginary “Toaster” subsystem that our stacks will register with at the lower edge, i.e. do it the way WDK sample does.
Are you ready to try it??? It looks like a great fun…
The “only” problem that I can immediately foresee here is that if we discover that it works just great, you will say “Well, this is just a virtual system that does not deal with the actual hardware - even if it works here it does not necessarily mean it will work in the real life”…
Anton Bassov
You known, I find little fun or insight to be gained by debating technical issues with you Anton. Sometimes I *think* we might have a good discussion, but your discussion degenerates to mere argument. This is because you seem to want to ARGUE and find a way to WIN, not DISCUSS what we’re talking about clearly and technically and come to a resolution – even if that resolution is “we agree to disagree.”
The above quote is a good example. This issue under discussion isn’t whether i venerate Cutler, or even Cutler’s intellectual prowess. If you GO ON to read what I wrote, RIGHT AFTER what you quoted…
Now… instead of resorting to lame ad hominem attacks, you might try addressing the issue: How IS it that several operating systems that have I/O subsystems that are “cumbersome and inefficient” have managed to become so obviously successful? DO you answer this? Ahhhh… no. You tell me that I worship Cutler and “what would I do if he changes his mind”?
As I’ve said before, every time an issue is raised that you can’t talk yourself out of, you don’t concede the point and move on, you slither the argument in a different direction. As a result, we never come to conclusion on these threads, and I – for one – never learn much of anything.
Another excellent example of this tactic is… in the face of my refutation of your statement that the Windows I/O model was essentially a microkernel based design, you attempted to twist the argument by replying:
The discussion wasn’t about violating the principal of functional decomposition into layers. The argument was about your fallacious and baseless asssertion that the Windows I/O subsystem was a microkernel design, implemented in a non-microkernel operating system, and thus needlessly inefficient.
When I informed you that your assertion was indeed incorrect – by citing the heritage of the windows I/O subsystem design – and tried to pin you down on that point, you attempt to switch the topic.
I can see you fancy yourself an OS architect. You WANT to be an OS architect. But you’ve got no track record – at least none that you’ve told us about – that demonstrates that you are a competent OS architect. And you INSIST on making assertions that are plainly wrong such as your assertion that the Windows NT I/O subsystem was designed for a microkernel, when I DEFINITIVELY demonstrated to you that this is not the case.
Despite all that, you’re entitled to your opinion, even if it’s essentially all conjecture and hot air. But given the fact that you don’t want to discuss the issues civily and in a reasonably disciplined manner, I’m out of this discussion.
This might be educational for YOU, but it isn’t for ME. And while I don’t mind mentoring you a bit, I won’t be attacked in the process.
Peter
OSR
Anton,
The problem with that idea is you have already dictated a lot of the
model. Having been on the architecture team of more than one OS (though not
successful ones), there is a heck of a lot baggage you are bringing to a
discussion with this. For instance:
Peripheral Model
You don’t mention this, but from my guess you are looking at the
standard PC model. That is one approach, but should it be designed for
something like Infiniband, or smart I/O processors. For instance I worked
for a mini-computer firm that created smart peripherals, and we choose to
make the disk and network models reflect the hardware interfaces of these
peripherals even though it slowed down some of the older devices.
Smart I/O
You are assuming a POSIX I/O model and ignoring all the others. This
is a common problem since both UNIX and Windows use similar I/O models, but
it is far from the only one. For instance, the same mini-computer company
had peripherals for serial ports that we downloaded state machines to. The
idea was a to allow editing to a large degree never to leave the perpheral.
I have not tracked things as well as I should but throuugh out the 80’s
there were still a fair amount of research on I/O models that moved some
smarts out of the application.
Fault Tolerance
There are a lot of things in the high availability / fault tolerant
space that impact I/O design. Yes people do graft them on in regular OS’es
with things lime MPIO but is this a goal for your model. Every body bitches
about crashes so perhaps it should be.
Synchronization model
You should know the OS’es synchronization model before you design an
I/O model. While things like semaphores, events, spin-locks are common they
are not the only models of synchronization.
I did not want to spend a lot of time on this or it could have gone on for
pages, but your idea of designing a driver model withough considering the
rest of the OS, shows a lack of experience in the problems of OS design.
–
Don Burn (MVP, Windows DDK)
Windows 2k/XP/2k3 Filesystem and Driver Consulting
Website: http://www.windrvr.com
Blog: http://msmvps.com/blogs/WinDrvr
Remove StopSpam to reply
wrote in message news:xxxxx@ntdev…
At the same time, it seems like a bit too much work for one person, don’t
you think? Therefore, if you are “truly interested in the outcome”, I
invite you to participate in it. Let’s split out efforts the following way -
I will write a driver model the way I see it, and you will write demo
stack(s) and app(s) to actually take advantage on it. This model will
peacefully coexist with the existing one - all one needs to do in order to
try something like that under Linux is to register a special filesystem with
VFS. At this point, the top of our stack will become accessible to the
standard read/write/ioctls/etc so that apps will be able to deal with our
devices, and what we do inside of it is our own business, so that we can
easily put an entire layered driver model there. For the sake of the
experiment I think it makes sense to abstract ourselves from dealing with
the actual hardware (at least at the original stage) - instead, let’s create
an imaginary “Toaster” subsystem that our stacks will register with at the
lower edge, i.e. do it the way WDK sample does.
Peter,
Sometimes I *think* we might have a good discussion, but your discussion degenerates
to mere argument.
Are you sure this is my fault??? Look at the statements that you made on this thread:
[begin quote 1]
You are, once again, grossly mis-informed and talking about something about which you apparently know little.
[end quote 1]
[begin quote 2]
I realize that you are a superior architect to Cutler, and thus I’m sure… somewhere somehow…
you have a track record or credentials or SOMEthing to back that up. Other than hot air, that is.
[end quote 2]
In your latest post you keep on playing the same lines:
[begin quote 3]
I can see you fancy yourself an OS architect. You WANT to be an OS architect.
[end quote 3]
[begin quote 4]
you’re entitled to your opinion, even if it’s essentially all conjecture and hot air.
[end quote 4]
As you can see it yourself, you are not arguing against the* things* that I say - instead, somehow it happens that you always speak about me IN PERSON. To be honest, judging from the above statements, one would hardly suggest that you are interested in a good discussion - instead, one would make a conclusion that you are much more interested in an “ugly bar-room brawl” that you are trying to provoke with the above statements all the time, rather than in a properly argumented technical discussion…
Now… instead of resorting to lame ad hominem attacks,
Please look again at the quotations that I have provided, so that you will realize how ridiculous the above complaint about “ad hominem” arguments is…
How IS it that several operating systems that have I/O subsystems that are
“cumbersome and inefficient” have managed to become so obviously successful?
DO you answer this?
Well, actually, I know only of 2 of them. Somehow it happened that the first one, despite some references to it as to “the only true OS ever designed”, never reached the level of success that is enjoyed by UNIX-like systems. The second one reached popularity only ON THE DESKTOP market and only because of the dominant market position of a company that was selling the OS in question. The trend seems to be changing now, and the only thing the company in question is able to do is to speak about “more than 60 patents” that Linux supposedly violates ( while failing to name any of them, of course) - that’s instead of improving their system, in technical terms…
So my question is - where is the tremendous success of Cutler’s designs that you are speaking about, compared to all other designs in existence ??? Sorry, but I cannot see any in so far…
But given the fact that you don’t want to discuss the issues civily and in a reasonably
disciplined manner, I’m out of this discussion.
Again, look at the quotes that I have provided above, and you will realize how ridiculous the above statement is…
This might be educational for YOU, but it isn’t for ME.
Well, judging from the quotes that I have provided above, you don’t really want it to be educational for yourself, do you. Sorry, but this is not my fault - I am trying my best to remain on the technical side
and to avoid personal references as much as I can…
And while I don’t mind mentoring you a bit, I won’t be attacked in the process.
Look - I am not attacking *YOU*, right? What I did attack is that the *argument* that you used on this particular_ occasion_, because this argument is nothing more than a logical fallacy…
Anton Bassov