Linux-Advocacy Digest #712, Volume #27 Sun, 16 Jul 00 04:13:03 EDT
Contents:
Re: Linsux as a desktop platform (T. Max Devlin)
Re: Linsux as a desktop platform (T. Max Devlin)
Re: Linsux as a desktop platform (T. Max Devlin)
Re: Would a M$ Voluntary Split Save It?
Re: Linsux as a desktop platform (ZnU)
Re: one step forward, two steps back.. (Jim Broughton)
Re: Linsux as a desktop platform (T. Max Devlin)
Re: Linsux as a desktop platform (T. Max Devlin)
Re: Linsux as a desktop platform ("Christopher Smith")
----------------------------------------------------------------------------
From: T. Max Devlin <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.mac.advocacy,comp.os.ms-windows.advocacy,comp.unix.advocacy
Subject: Re: Linsux as a desktop platform
Date: Sun, 16 Jul 2000 02:47:01 -0400
Reply-To: [EMAIL PROTECTED]
Said Gary Hallock in comp.os.linux.advocacy;
>"T. Max Devlin" wrote:
>
>> Said The Ghost In The Machine in comp.os.linux.advocacy;
>> [...]
>> >However, at least the kernel has the option of forcibly moving
>> >things around; on a pure CMT system, this isn't really possible.
>>
>> So would you say that a suggestion might be warranted that it might be
>> feasible and beneficial to consider making it unnecessary, as well as
>> impossible, instead of simply taking the opposite approach of making it
>> possible?
>>
>
>Gibberish. Look, if you are trying to suggest that a CMT system could be
>designed to allow forcibly moving thongs around, then no that is not
>possible. Why, you ask? Because it would require preempting processes
>which, by definition, is not CMT.
No, I am trying to suggest that a CMT system which does not need to
"forcibly move things around" might be possible. A CMT with a more
comprehensive mechanism for allowing non-active processes to effectively
pre-empt the active process without requiring an external scheduler.
Perhaps token passing in place of polling might be a suitable
comparison, or at least analogy. Do you see what I mean?
--
T. Max Devlin
Manager of Research & Educational Services
Managed Services
ELTRAX Technology Services Group
[EMAIL PROTECTED]
-[Opinions expressed are my own; everyone else, including
my employer, has to pay for them, subject to
applicable licensing agreement]-
====== Posted via Newsfeeds.Com, Uncensored Usenet News ======
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
======= Over 80,000 Newsgroups = 16 Different Servers! ======
------------------------------
From: T. Max Devlin <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.mac.advocacy,comp.os.ms-windows.advocacy,comp.unix.advocacy
Subject: Re: Linsux as a desktop platform
Date: Sun, 16 Jul 2000 02:57:20 -0400
Reply-To: [EMAIL PROTECTED]
Said void in comp.os.linux.advocacy;
>On Sat, 15 Jul 2000 22:11:39 -0400, T. Max Devlin <[EMAIL PROTECTED]> wrote:
>>Said The Ghost In The Machine in comp.os.linux.advocacy;
>> [...]
>>>However, at least the kernel has the option of forcibly moving
>>>things around; on a pure CMT system, this isn't really possible.
>>
>>So would you say that a suggestion might be warranted that it might be
>>feasible and beneficial to consider making it unnecessary, as well as
>>impossible, instead of simply taking the opposite approach of making it
>>possible?
>
>No. You're continuing to ignore the principle that system-wide
>resources should be managed by the operating system. If you think you
>can gainsay this principle, you'd better tell me exactly why, with
>real-world examples and no handwaving whatsoever.
Shared media Ethernet LANs are not "run" by the hub, which is a passive
repeater; the shared resource is not managed by a central mechanism, but
by all the components autonomously following an explicit set of rules.
Token Ring LANs provide many of the benefits of a polling system for
sharing, but use a daisy-chain token passing method which seems strongly
related to me to the CMT/PMT issue in principle, if not application.
Yes, I think we can gainsay this principle. Obviously, these are
examples from my own background in networking, but the principle of the
value of autonomous cooperation as superior to centralized control seems
well substantiated by many developments in technology over the past
thirty years.
In fact, I'd go so far as to say there is no such principle as
"system-wide resources should be managed by the operating system".
Merely (I know you'll hate me, but) an assumption that this is always
the best, and possibly the only, possible approach. I would even
consider it the default, based on the way humans tend to go about a
problem. I think it *may* be possible that "system-wide resources
should be managed by cooperative operation of autonomous scheduling
processes" might be every bit as robust an approach, properly
implemented, though potentially more susceptible to improper
implementation.
--
T. Max Devlin
Manager of Research & Educational Services
Managed Services
ELTRAX Technology Services Group
[EMAIL PROTECTED]
-[Opinions expressed are my own; everyone else, including
my employer, has to pay for them, subject to
applicable licensing agreement]-
====== Posted via Newsfeeds.Com, Uncensored Usenet News ======
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
======= Over 80,000 Newsgroups = 16 Different Servers! ======
------------------------------
From: T. Max Devlin <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.mac.advocacy,comp.os.ms-windows.advocacy,comp.unix.advocacy
Subject: Re: Linsux as a desktop platform
Date: Sun, 16 Jul 2000 02:59:52 -0400
Reply-To: [EMAIL PROTECTED]
Said Colin R. Day in comp.os.linux.advocacy;
>"T. Max Devlin" wrote:
>>
>> But what is "rare" at 60 seconds a minute might turn out to be extremely
>> frequent at millions of cycles a second, don't you think? CMT doesn't
>> all by itself offer anything near what you know the technical
>> requirement are, and I recognize that. But CMT never purported to. It
>> is CMT plus the implementation of scheduling, *cooperatively*, the apps,
>> which is of value, and can indeed, as proven in the marketplace, provide
>> effective solutions which end users can benefit from, even if engineers
>> want to insist they're "stupid".
>>
>
>And how did the market do this? Did people choose Mac because it
>had CMT as opposed PMT? Did people choose applications on the
>basis of the apps' CMT performance, and were they even aware that
>that was an issue?
No, and that's literally why my question came up. It didn't seem to
make near as much difference as would be indicated by the resistance to
CMT as a valid approach. That's the point that confuses me.
--
T. Max Devlin
Manager of Research & Educational Services
Managed Services
ELTRAX Technology Services Group
[EMAIL PROTECTED]
-[Opinions expressed are my own; everyone else, including
my employer, has to pay for them, subject to
applicable licensing agreement]-
====== Posted via Newsfeeds.Com, Uncensored Usenet News ======
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
======= Over 80,000 Newsgroups = 16 Different Servers! ======
------------------------------
From: <[EMAIL PROTECTED]>
Date: Sun, 16 Jul 2000 06:00:51 GMT
Subject: Re: Would a M$ Voluntary Split Save It?
Crossposted-To:
comp.os.os2.advocacy,comp.os.ms-windows.nt.advocacy,comp.sys.mac.advocacy
>>>>>>>>>>>>>>>>>> Original Message <<<<<<<<<<<<<<<<<<
On 7/15/00, 11:38:26 PM, "JS/PL" <[EMAIL PROTECTED]> wrote=20
regarding Re: Would a M$ Voluntary Split Save It?:
> > It is mysterious to me because whether or not bundling IE with Windo=
ws
> > to kill Netscape by trying to "cut off their air supply" is not a
> > question of designing an OS, but of building and selling a product. =
=20
It
> > is, indeed, incredibly lousy OS design, IMHO, but that is, indeed,
> > beside the point. Microsoft is free to make stupid design decisions=
.
> > They are not free to force them on the market through monopolization=
=20
or
> > tying.
> Yes- they are free to tie whatever they want with their software=20
(*including
> a ham sandwich), and it is their right to have their product=20
distributed in
> a un altered state. Every software maker has the right to insist that =
2nd
> party distributors not re-write the software before distribution.
> It is ludicrous to assert that OEM's should be allowed to delete MSIE =
before
> distribution or remove the Microsoft Windows start up splash screen as=
=20
some
> wanted to do a few years back.
We know a ham sandwich isn't windows so there is a technical=20
distinction you totally avoid - which is why you're wrong. IE and=20
Windows are two distinct products. OEMs have every right to not be=20
forced to include MSIE if they choose to sell the monopoly product,=20
windows. That's the law. Mandating they include MSIE with Windows is=20
illegal. Bolting MSIE to Windows (or bolting a ham sandwich) doesn't=20
make MSIE or the ham sandwich part of the OS. Maybe now you understand=20
the significance of the ham sandwich argument.=20
OEMs also have a right to sell PCs that do not boot up into windows=20
directly with out being punished by MS monopoly power. MS has no=20
right to control the OEM's product which is what they sell to we the=20
consumer.=20
Why make such a MS centric argument about rights and not defend a=20
OEM's right to design and brand their own hardware? This is so phony=20
an argument to pretend there is one company with rights.
> Netscape didn't have any supposed "air supply" cut off by Microsoft
> including a browser with the OS. Netscape was included on nearly every=
=20
OEM
> master restore disk anyway, and they had deals with many major=20
software
> makers to be included on many install disks. And Netscape was always=20
freely
> downloadable! It was basically free even before they officially gave=20
it
> away. Free in the sense that they was an unspecified "trial period"=20
with
> absolutely no nag screen or reminder- ever.
You're facts are wrong. I'm not sure what the point is of trying to=20
deny the obvious. Netscape's browser did generate revenue and=20
companies did pay for Navigator and so did governments lest they be=20
sued for stealing software. Netscape was ruined - bought by AOL they=20
were so damaged by MS. In a court room MS was found guilty of=20
limiting Netscape by bolting MSIE into windows and ordered to be=20
broken up - Why argue others have to accept as fact some extreme,=20
non-tractable story? If the facts don't work for you you still cannot=20
change them.
------------------------------
From: ZnU <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.mac.advocacy,comp.os.ms-windows.advocacy,comp.unix.advocacy
Subject: Re: Linsux as a desktop platform
Date: Sun, 16 Jul 2000 07:00:51 GMT
In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED]
wrote:
> Said Gary Hallock in comp.os.linux.advocacy;
> >"T. Max Devlin" wrote:
> >
> >> Said The Ghost In The Machine in comp.os.linux.advocacy;
> >> [...]
> >> >However, at least the kernel has the option of forcibly moving
> >> >things around; on a pure CMT system, this isn't really possible.
> >>
> >> So would you say that a suggestion might be warranted that it might be
> >> feasible and beneficial to consider making it unnecessary, as well as
> >> impossible, instead of simply taking the opposite approach of making
> >> it
> >> possible?
> >>
> >
> >Gibberish. Look, if you are trying to suggest that a CMT system could
> >be
> >designed to allow forcibly moving thongs around, then no that is not
> >possible. Why, you ask? Because it would require preempting processes
> >which, by definition, is not CMT.
>
> No, I am trying to suggest that a CMT system which does not need to
> "forcibly move things around" might be possible. A CMT with a more
> comprehensive mechanism for allowing non-active processes to effectively
> pre-empt the active process without requiring an external scheduler.
CMT has no mechanism allowing an inactive process to preempt the active
process.
> Perhaps token passing in place of polling might be a suitable
> comparison, or at least analogy. Do you see what I mean?
An processes can't pass a token if it doesn't have the CPU. If it has
the CPU, it doesn't need to pass a token.
--
The number of UNIX installations has grown to 10, with more expected.
-- The Unix Programmer's Manual, 2nd Edition, June 1972
ZnU <[EMAIL PROTECTED]> | <http://znu.dhs.org>
------------------------------
From: Jim Broughton <[EMAIL PROTECTED]>
Subject: Re: one step forward, two steps back..
Date: Sun, 16 Jul 2000 03:05:10 -0400
Reply-To: [EMAIL PROTECTED]
MH wrote:
>
> Just checking in. (= Haven't been here in a LONG time. I used Linux for a
> couple of years back in the RH 4.2 to 5.1 days. Ditched it because I just
> didn't have the time, was learning windows development, going to school, and
> the 'linux allure' just didn't grab & wipe off on me as easily as it does
> some. Not to mention the email threats from COLA I got one day. Nice bunch.
> (-:
>
> So, I decided what the heck, I've got a little time this summer - I'll try
> it again.
> Quick update from the front:
>
> The distros are doing exactly as I knew they would three years ago. Trying
> to emulate windows installation programs. It isn't working. They're better
> in some regards, worse in others. I've installed W2K, and on this PC, winME.
> The installation was flawless.
>
> RH still insists that my external modem is missing on each boot, (when it's
> there), sometimes the sound works, sometimes it doesn't. Half of what I
> installed is buried somewhere - not on the menus. The default installs I
> think are a good idea. Trouble is, some of them leave you hanging with a
> useless setup or bomb out trying to deliver the latter, or won't let you
> setup things they didn't install very easily.
> Having said that, however, it has improved for a novice.
>
> The window managers are trying to emulate windows. It isn't working.
> Neither Gnome or KDE comes close. I can see the point of trying, but if
> you're going to do it, do it right or don't do it at all. It's not right.
> The menu systems are a complete mess. Why does gnome have to automagically
> plaster that useless bar across the bottom by default no matter what WM you
> use? I know these things can be configured by hand in the config files, ..
> but I thought we were doing GUIs here, remember? Drag and drop support?
> Nope. Half baked at best. We're doing GUIs here, remember? We're better than
> windows, remember? Not at gui's your not. Not even in the same ballpark in
> the same league, in the same decade.
>
> Memory useage. When I ran Linux last, it was RH 5.1 on a P100 with 64MBs of
> ram. This box would NEVER swap. I was running Afterstep for the WM. Even
> that terrible excuse for a browser NN, when running with three or four other
> apps...no swap. none.. nada. I loved it. Now? HA. KDE, NN only, on a PPRO
> 200 with 96MBs of ram- I'm generating 50+ MB swap files. Not only that, the
> system does not seem much faster than the P100. The MS bloat syndrome has
> come home to roost. As I have said before, Linux was agile, and stable
> because it was lean and well tuned. The apps (small & lean) most people ran
> were tried and true. Now the $$ has taken over and wants the desktop.
> They're trying to emulate windows. It isn't going to work. You run big GUI
> based apps on top of big GUI'd window managers and you have created the same
> problems you have in windows. Only worse because windows has had years
> 'tuning' this slop to the point that it's getting practically stable
> now. --well, in windows terms any way!
>
> If it wasn't for bedroom hackers wanting their PC's desktop to look
> different because they have the newest theme of the month, with skins
> flapping off the walls I don't think many home users would be coming to
> linux at all. The distro's are just turning out bloated slop that doesn't
> have HALF the features of windows. Thank god for building your own Linux
> setup. Otherwise, I'd ditch it for good. This 'takeover the desktop' thing
> is just as I predicted. A mess. The iceWM is all I need with as many
> terminals as my screen will hold, running gcc or simple editors, writing
> scripts so on and so on. Internet? Big apps? Forget it. NN is just plain
> sub standard compared to IE. Have you ever really used IE? It's another
> world. Talk about features and ease of use. Star office? I've seen it. I
> wouldn't install that bug fest if Larry blew my dog. You think office is
> bad? Whew. Star office is chilling.
> Desktop domination?-- it's a LONG way off, if at all. And I think that's a
> good thing for Linux. And computer users in general.
Oh gawd another wintroll.
JIM
--
If Sense were common everyone would have it!
------------------------------
From: T. Max Devlin <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.mac.advocacy,comp.os.ms-windows.advocacy,comp.unix.advocacy
Subject: Re: Linsux as a desktop platform
Date: Sun, 16 Jul 2000 03:11:47 -0400
Reply-To: [EMAIL PROTECTED]
Said void in comp.os.linux.advocacy;
>On Sat, 15 Jul 2000 22:47:06 -0400, T. Max Devlin <[EMAIL PROTECTED]> wrote:
>>
>>You are correct. I cannot give examples, only reasons, to back up my
>>statements.
>
>But everybody knows that you can derive all sorts of garbage with
>ironclad logic if you start with false assumptions.
Yes. So why not simply point out the false assumptions if you can
identify them, rather than ridicule what you see as garbage? I know, I
know, you think you did. But I don't call "everybody with any knowledge
of the subjects knows you're wrong" as pointing out a false assumption.
>>If they are easily refuted with logic, then I capitulate,
>>with my apologies. But I won't take anyone's word for whether what they
>>present contradicts my position;
>
>That's your failure. You've picked wrong postulates, but you think
>you're hot shit because your logic is valid. This is the worst sort of
>sophistry.
I do not engage in sophistry. Merely free inquiry, with an expectation
of civility.
>>More specifically, Ethernet interconnects transceivers of indeterminate
>>and unknown number (limited by physical resources, essentially). Rather
>>than have them all plug in to a switch which then interconnects the
>>transceivers, a shared media (bus or hub) is used, which acts passively
>>and does not "control" the channel.
>
>Wrong. Everyone who can afford to uses switches nowadays. It's a bogus
>analogy anyway.
Thanks for the clear and concise appraisal of my reasoning.
Everybody can afford to use switches these days, yes. And most all of
their links are at <.1% utilization almost all the time, except for the
few that are at 100% utilization all the time that act as bottlenecks.
The wide implementation of Ethernet switches was often an inappropriate
response to bad application design. I say inappropriate merely because
it did not solve the problems, not through any technical bias. The rest
are a result of the fact that everyone who can afford LANs can usually
afford switches, thanks to the cheap ASIC technology which makes those
switches possible.
--
T. Max Devlin
Manager of Research & Educational Services
Managed Services
ELTRAX Technology Services Group
[EMAIL PROTECTED]
-[Opinions expressed are my own; everyone else, including
my employer, has to pay for them, subject to
applicable licensing agreement]-
====== Posted via Newsfeeds.Com, Uncensored Usenet News ======
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
======= Over 80,000 Newsgroups = 16 Different Servers! ======
------------------------------
From: T. Max Devlin <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.mac.advocacy,comp.os.ms-windows.advocacy,comp.unix.advocacy
Subject: Re: Linsux as a desktop platform
Date: Sun, 16 Jul 2000 03:13:38 -0400
Reply-To: [EMAIL PROTECTED]
Said [EMAIL PROTECTED] in comp.os.linux.advocacy;
[...]
>>Thanks for pointing it out; I am aware of how difficult I've been. This
>>post was quite embarrassing in retrospect. But these are advocacy, not
>>technical, groups.
>
>I beg your pardon? You think you can blindly advocate things without
>providing technical reasons here?
No, nor did I say that.
>Oh...wait - you've done exactly that. Never mind.
Nor have I done that.
>>What changed from 1988 to now is the
>>commercialization of microcomputer technology, where whether something
>>is popular is potentially due to massive cluelessness as the possibility
>>that it is technically valid.
>
>In other words, your theory is that products come to market due to
>product popularity rather than because it's what you would have
>chosen. [...]
[...These statements are not worth commenting upon. Try to get that
hair out of your ass...]
--
T. Max Devlin
Manager of Research & Educational Services
Managed Services
ELTRAX Technology Services Group
[EMAIL PROTECTED]
-[Opinions expressed are my own; everyone else, including
my employer, has to pay for them, subject to
applicable licensing agreement]-
====== Posted via Newsfeeds.Com, Uncensored Usenet News ======
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
======= Over 80,000 Newsgroups = 16 Different Servers! ======
------------------------------
From: "Christopher Smith" <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.mac.advocacy,comp.os.ms-windows.advocacy,comp.unix.advocacy
Subject: Re: Linsux as a desktop platform
Date: Sun, 16 Jul 2000 17:26:39 +1000
"T. Max Devlin" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Said Colin R. Day in comp.os.linux.advocacy;
> [...]
> >Sorry, I meant clients as customers, not machines. A customer pays
> >you to run a complex simulation. The simulation does not require your
> >attention after you start it, but it does require the lion's share of the
> >CPU. If the customer pays enough, I will put up unresponsive apps
> >to run the simulation.
>
> This seems a direct argument for Shortest Job First, in which some
> explicit notion of which thing is expected to finish first is given
> priority. The problem, of course, is determining the notion, and
> deriving it.
Which is, naturally, next to impossible without knowing a great deal about
whatever it is you're running. Vanilla SJF is a poor algorithm to use on
any system where jobs may potentially take a very long time [1][2].
Hence multiple queues would probably solve this problem best, with the
process in a low priority queue in the former case and a high priority queue
in the latter case.
> >> One of the coolest utilities I've seen for
> >> handling Internet stuff is a download app called GoZilla. This handy
> >> little database inserts itself in the way whenever I click a link in a
> >> browser that would start a download. It keeps track of my downloads
and
> >> makes managing them in all sorts of ways much easier. But that's not
> >> why I got it. The thing I love is that each download window for each
> >> separate file has a little slider in it, which enables me to
> >> individually and directly control (as well as set defaults and policies
> >> with download categories) how much of my available throughput that one
> >> particular download is allowed to take.
> >
> >But how does this relate to CMT? This seems to be network bound
> >rather than CPU bound. Can one do this in CMT? One can do this
> >with CPU time in Linux, at least if one is root (shouldn't be a problem
> >on a personal desktop machine), although I don't know if there is a
> >nice GUI front end.
>
> One can do anything in CMT one can do in PMT, if one designs a suitable
> CMT system.
In the context of this discussion, no, you can't. The reason is because
ultimately CMT is at the mercy of the applications.
There are a few simple and specific situations where a CMT scheduler would
be a better choice, but these are irrelevant to the types of computers we're
talking about.
> I do indeed like the potential for elegance and flexibility
> in leaving CPU use to a daisy-chain system of autonomous processes using
> an explicit and minimal set of rule mechanisms, rather than an external
> scheduler. I understand the engineering difficulties involved,
> generally, I think, but that isn't related to GoZilla. The manipulation
> of the scheduler, though, is indeed the most direct mechanism for
> implementing this kind of functionality in the current PMT system. It
> merely requires integration into the UI to provide more efficient and
> adaptive configuration by the operator. Because only the operator knows
> if they have a client waiting; the scheduler is as ignorant as the app
> would be in a CMT system.
All PMT OSes have methods for manipulating priorities. Almost all have
included programs for manipulating priorities while a process os running.
Of those, the vast majority have GUI tools.
As usual, your hypothetical problem was determined and fixed a very long
time ago.
You also ignore the fact it is a very rare case for the vast, vast majority
of people that a priority needs to be manipulated. Most people who need to
be manipulating priorities, know what they are and how to do it.
> >That's why there is nice/renice.
>
> Yes, but the nice value, it turns out, doesn't have anywhere near the
> impact as the other components of the algorithmic scheduling. I'm not
> sure really how much, though, as I'm not a CS graduate, and I really
> suck at math. Here's the piece where it comes up, I think:
>
> p_usrpri = PUSER + (p_cpu/4) + 2.p_nice
>
> And the description says that nice "cancels out some of the effects of
> high CPU loading". I'm not really sure how much impact that has on the
> over-all efficiency of algorithmic scheduling.
It won't affect the *efficiency* at all. You'll still maximise the usage of
system resources, assuming[1] you're running enough stuff to actually do
that.
A plethora of, say, xterms running at the highest priority that are all
waiting for user input will still result in a mostly idle CPU and a highly
responsive system.
It might affect the responsiveness of the system, if you bump something that
actually uses a lot of cycles, but if you're messing with priorities that's
a risk you've implicitly decided to take.
> Then again, I guess I'm still not sure how much impact algorithmic
> scheduling has overall on the handling of bottlenecks, or where those
> bottlenecks are.
It reduces their impact on the efficient use of system resources.
However, if a process is waiting for some slow device like a printer, it
will still have to wait for it, so if that's the only process you're using
then you *will* notice the delay. Nothing can change that fact, although
PMT will lessen it from the OSes point of view.
The fundamental different between PMT and CMT is that while that process is
waiting for the printer (or some other device), on a PMT system the rest of
the processes on the system can potentially do useful work (assuming they
have some to do).
> So I'm obviously floundering at this point.
You've been floundering since your fist post. Even now, after many subtle
and not-so-subtle suggestions to actually go and learn even the *basics* of
the topic under discussion, then sit down and think about them for a while,
you obviously haven't.
The intricacies of CPU scheduling is not something you can even learn about,
let alone understand, in "20 minutes".
> But I'm
> told that some of these assumptions may not have been questioned in more
> than twenty years, and I think some rather broad changes in use of
> computing resources have occurred since then.
Not really. Everything's gotten faster and bigger, but there haven't been
any fundamental changes (at least not in mainstream computing) in the way
computer resources are used - and faster and bigger is what makes the few
advantages CMT has over PMT irrelevant. That's why PMT is still around
after 30-some years.
[1] On that note, I remember hearing a story in one of my OS lectures about
an old, old computer (talking early 70s here IIRC) at a University in the
States using a similar system. Apparently when it was being shut down after
many years of faithful service, they discovered a job that had been
submitted *years* beforehand and never been started, due to its predicted
length and the SJF algorithm.
[2] "Very long time" being a relative measure. A system which runs nothing
but days-long jobs obviously has a different concept of "very long time" to
a system being used interactively, like a desktop.
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.advocacy) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
