Linux-Advocacy Digest #910, Volume #25 Sun, 2 Apr 00 09:13:03 EDT
Contents:
Re: Linux bugs!!! (Jim Dabell)
Re: Linux vs Windows development man-hours? ("Robert Moir")
Re: Linux vs Windows development man-hours? ("Robert Moir")
Re: Need help on compiling Linux stats (Simon Brooke)
Re: Linux vs Windows development man-hours? (Robert Morelli)
Re: Windows 2000 has "issues" (Joseph)
Re: Windows 2000: nothing worse ("Erik Funkenbusch")
----------------------------------------------------------------------------
From: Jim Dabell <[EMAIL PROTECTED]>
Subject: Re: Linux bugs!!!
Date: Sun, 02 Apr 2000 13:15:50 +0100
Reply-To: [EMAIL PROTECTED]
Mig Mig wrote:
>
> Jim Dabell wrote:
> > Mig Mig wrote:
> > >
> > > Cihl wrote:
> > > > This is a follow-up of the thread "Nice link" which was started by me :-)
> > > > I wish to propose the following challenge for everyone:
> > > >
> > > > Find a way, ANY way of crashing the entire Linux operating system while
> > > > operating as a non-root user! Please post anything you can find in this
> > > > thread. I'm sure many people will be interested!
> > >
> > > Thats rather easy by reading the kernel mailing list.. i remneber one crash
> > > caused by a user writting direct to video-memory
> >
> > And exactly how could a normal user do that? You need special
> > permissions to write directly to video ram.
>
> Try search the kernel archives.. im sure Alan Cox discussed the problem in
> around may-june 1999
Sorry, can't find it. Still, if it was discussed that long ago, it's
bound to have been fixed by now. Oh hold on, are you talking about
sending bad data to /dev/3Dfx? That's an artifact of how the Voodoo 1
chip and Glide work, AFAIK, and can't be worked around without
significant performance loss.
In any case:
a) It's an external driver
b) root can restrict it to "trusted" people
c) there's no way of exploiting it unless you can execute code on the
computer.
Let's compare that to MS:
a) The aux/con drivers are built into all DOS derivatives, with no way
of disabling them.
b) Anybody on the machine can access them
c) You can exploit them with a simple web page or email
Genuine question: have MS fixed the bug yet?
Jim
------------------------------
From: "Robert Moir" <[EMAIL PROTECTED]>
Subject: Re: Linux vs Windows development man-hours?
Date: Sun, 2 Apr 2000 12:09:40 +0100
"Robert Morelli" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
[...]
> It's only been in the past couple of years that Microsoft has tried to
establish
> pure research labs. For them, it's an afterthought and a publicity
stunt, like
> the philanthropy that also only began recently. Companies like IBM and
ATT have
> always had a strong focus on fundamental research. This difference is
very
> significant in terms of production. IBM and ATT were both once
monopolistic like
> Microsoft is now. But at least IBM has contributed countless innovations
in everything
> from CPU design to database technology. Microsoft is a monopolist, but
it hasn't
> redeemed itself with innovations. Whatever people they may have hired for
a showcase
> research lab in the last couple of years aren't a factor in the
development of Windows.
Few years? I remember working on projects back in 1995 that used the
*results* of work done by MSR. So even if you assume they bent over and
pulled good hard code out of their arses at 5 mins notice they've been going
5 years at least, and I happen to know there was a few years work done
before I was invited to the party. So how long is "a few years" to you -
MSR's main lab was founded in 1991.
------------------------------
From: "Robert Moir" <[EMAIL PROTECTED]>
Subject: Re: Linux vs Windows development man-hours?
Date: Sun, 2 Apr 2000 12:12:38 +0100
One problem. Choice, the thing that Linux supports.
I don't WANT to use Linux. No matter how free it is.
"Charlie Ebert" <"Charlie Ebert" [EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
I'm going to delete a bunch of conversation here and just get to
the point.
Microsoft is a fairly large company with around 800 actual people
involved in software development of their products.
Of which, half of them are working on other products such as
Microsoft Office, flight simulators, so on and so forth.
That leaves about 450 software programmer/analysts with around
50 more people working as full time system analysts engaged on
the 2000 project. Check the Microsoft Web site for details please.
Don't get into some idiotic contest with me please.
The average development body at Microsoft committes him/herself
to a 55 hour week.
Linux on the other hand is managed at several levels.
There's the Kernel level managed by Linus, whereby you have
about 200 people involved, helping him maintain the kernal.
Then you have the different distributions such as Suse with 350
people, RedHat with 400 people, Debian with 300 people, Turbo linux
with 300 people, Mandrake with 100 people, so on and so forth.
Then let's view the drivers situation. The manufacturers write
hardware drivers for Windows operating systems and then they
submit them to the test via Microsofts independent contracted testing
agency.
Linux either has a programmer volunteer to write the driver himself or
the hardware manufacturer does it himself for linux.
So Linux is once again clearly ahead in manhours!
Linux started almost 10 years after Microsoft did. They are 10
years behind. But that's not completely true as most of the
Systems design was already layed out for them. All they had
to do was copy the model for their own.
So in code maturity, Linux is just about 2 years behind Microsoft.
Let's say that, if it's even 2 years.
Now there are other people working on Linux code out there submitting
patches to these distributions I've talked about earlier. There's
nobody from the Microsoft camp doing that.
There are 25 or more ENGLISH distribution organizations for Linux per
the WWW.LINUX.ORG site. There are MORE which are non-english.
Let us say there are around 50 of them total in the world.
Each organization has an average of 200 people working on their
code. And for every organization, 100 more people from outside
submit bug patches.
The linux organizations vary from full time jobs at companies,
to volunteer efforts such as with Debian. I will average them
all out to a 40 hour work week which I see as fair.
So 50 distributions times 200 people times 40 hours per week =
400,000 man hours per week there.
And 100 people times 50 distributions time 2 hours per week =
10,000 of free wheeling bug patching help.
Linus's group not even being included here folks.....
200 programmers time 30 hours a week average times 1 kernel =
6,000 hours a week there as a good guess. See Linus's group is
NOT a corporation. This isn't their job...
The total man hours spent on Linux is an estimated 425,000 a week.
Microsoft has 500 employees times 55 hours per week =
27,500 man hours per week.
Notice how I avoided the time spent writing drivers for specific
hardware from this equation. See, Microsoft spends very little
time doing this. They mainly translate other peoples work.
While the Linux crew, if not handed code, will have to write it.
So we have Microsoft at 27,500 man hours per week.
Then we have Linux at 425,000 per week, EASY!
That means Microsoft Corporation is roughly 7% the size of the
Linux community and the Linux community is at most 2 years behind
them in technology.
If I were Bill Gates, I would be making plans for getting out
of the way of Linux. I would be making post Microsoft OS plans
for the company.
Microsoft has an absolute ZERO chance for survival as an operating
systems company 10 years from now.
The writing is on the wall. $300 + bucks for Windows 2000
Workstation, and even at retail prices $65 for a good distribution of
Linux. And if you know how to FTP install you can avoid paying
the $65 and get it for free.
10 years from now, we will not be talking about Microsoft corporation.
10 years from now, if Microsoft is still a corporation it will be an
applications vendor at best.
The BIG guns of the future are clearly Linux, FreeBSD and the BSD
clan, and the GNU's HERD.
I don't think Solaris nor BEOS, nor Microsoft will be here 10 years
from now. They won't have a place in that market.
I think the real questions to be answered will be what is the
future of FreeBSD and BSDI....?
What power can the GNU HERD posses over the linux kernel which
would make us jump over there.....?
Forget about Microsoft. They have only 1-2 more OS releases
to go before it's boot hill for the company.
And I know Bill Gates knows this. This is WHY there is only
one OS offered at Microsoft now, Windows 2000! No more 95, 98 then NT
nonsense. The days of thowing resources to the wind are
over. Everyman on Windows 2000! It's a fight to the death
for Microsoft Corporation.
Isn't it amazing how one guy, Linus Torvalds conquered the world with
his Kernel.
Bill Gates wanted to accomplish that.
With a Free operating system, what's the purpose in any commerical
venture development such as Windows 2000??? How about even Solaris???
Suppose they maintain 10,000 customers each, can they
break even as they have to compete with the monster that is LINUX?
Commerical OS situations will be about as welcome as used toilet
paper in 10 years.
And if you don't believe that, tell me how I'm wrong here!
Show me what Windows 2000, 98, Solaris, or whatever has on
my Mandrake 7.0 (Air) Distribution.
Show me how anyone of them beats Suse 6.4!!!!
You can't. They don't...
Between man-hours, price, SUPPORT, Microsoft stands no chance
of survival in the years to come.
Linux has better Support. I get my answers in litterally minutes!
I get patches the next day if not that afternoon!
Bill Gates should start thinking along these lines of becoming
an applications vendor for Linux before it's too late.
That's where the money will start rolling in in just 2 years time.
Charlie
------------------------------
Crossposted-To: alt.os.linux,uk.comp.os.linux
Subject: Re: Need help on compiling Linux stats
From: Simon Brooke <[EMAIL PROTECTED]>
Date: Sun, 02 Apr 2000 12:34:02 GMT
"Tom Steinberg" <[EMAIL PROTECTED]> writes:
> I have spent the last couple of days reading endless Linux articles looking
> for any and every bit of statistical data I can find concerning Linux. I'm
> not getting very far very fast, so I've returned to the wonderful and
> generous people on usenet ( butter butter ) to ask for suggestions and info.
> Absolutely anything numerical is helpful, but I am especially looking for:
>
> 1) Info/guesses about the Linux and opensource user base. Growth patterns
> over time and projections are especially needed.
This really is extemely difficult and I doubt that anyone has any
useful figures.
The following resources are all of doubtful utility:
<URL:http://counter.li.org/>
- not widely used, so of little utility
<URL:http://www.netcraft.co.uk/survey/>
- monthly survay of 13 million Web servers; does not counf
Linux as such but counts Apache; an unquantifiable but
probably large proportion of Apache sites run Linux.
Statmarket <http://www.statmarket.com/> will sell you information,
but I doubt whether their information is better than anyone
else's. They offer sample data at
<URL:http://statmarket.com/html/sample_operating_system.html>
which gives their estimate of the percentage of machines running Linux
as 0.22%. Seeing their data appears to be based on analysis of visits
to websites by browsers, and Linux machines are at present more likely
to be servers than clients, this is unlikely to be a realistic estimate.
> 2) Value of markets, quantified cost benefits, market impact analyses of
> opensource.
The LinuxCare IPO prospectus is probably as good a source as you will
find for this sort of data, as the largest single component of the
financial value of the open source market is likely to be in support.
> 3) Stock valuations of opensource ( esp. Linux ) firms.
> 4) Revenue streams associated with opensource.
Few open source companies have much trading history. One of those few
was Cygnus, but they've been taken over by RedHat. RedHat and Caldera
are both publicly quoted on NASDAQ, as are LinuxCare and one or two
others, so their market valuation is available, but all are is in my
opinion presently inflated above their real worth.
It's very difficult to derive true values for any of these figures,
for two reasons:
1: The numbers don't exist in the first place. There is no effective
central bureau for collating Linux stats. The number of machines on
which Linux has ever been installed is likely to be significantly
greater than the number of Linux distributions which have ever been
sold, because, apart from people who have used magazine cover disks or
direct downloads from the Internet, it's probable that many boxed
copies of distributions have been loaded onto multiple machines. At
the same time, Linux isn't always a love-at-first-sight experience for
newbies, so the number of machines which normally run Linux is likely
to be significantly smaller than the number which have had Linux
installed.
2: There are many people trying to cash in on the hype. This year,
particularly, dozens of commercial entities who wouldn't know what a
cron job was if it fell on them out of a tree are trying to sell
Linux, or services related to Linux. There's a frenzy of hype
happening at present, much of it very uninformed. I was particularly
interested in Netcraft's recent dogfood analysis, which I will quote:
'On a related note, it's notable that the Linux evangelism
bandwagon is bringing in people more prepared to envangelise the
operating system than run it themselves. At the last Linux Expo in
London, more of the exhibitor companies were running Microsoft-IIS
than Linux on their own sites. Recently queried sites on our
server query form include; linux.ora.com which runs Solaris, while
linuxbeacon.com is an early adopter of Windows 2000 along with the
shortlived www.linuxanswers.co.uk and the parody domain names
www.slashdot.org.uk and www.freshmeat.org. However, contrary to
what one might expect, www.linuxsucks.org runs Linux.'
I have to say I'm looking forward to your paper.
--
[EMAIL PROTECTED] (Simon Brooke) http://www.jasmine.org.uk/~simon/
'there are no solutions, only precipitates'
------------------------------
Date: Sun, 02 Apr 2000 06:06:09 -0400
From: Robert Morelli <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: Re: Linux vs Windows development man-hours?
Charlie Ebert wrote:
>
> I'm going to delete a bunch of conversation here and just get to
> the point.
>
> Microsoft is a fairly large company with around 800 actual people
> involved in software development of their products.
>
> Of which, half of them are working on other products such as
> Microsoft Office, flight simulators, so on and so forth.
>
> That leaves about 450 software programmer/analysts with around
> 50 more people working as full time system analysts engaged on
> the 2000 project. Check the Microsoft Web site for details please.
> Don't get into some idiotic contest with me please.
As I'm sure you've gathered from my post, I'm no apologist for Microsoft.
But I just can't believe that the numbers you're claiming are realistic.
According to their web site, Microsoft has about 35,000 employees of whom
about 14,000 are in R&D. Where are you getting 800? Now, 14,000 is not
realistic either.
You mention flight simulators. It's actually well known that Microsoft
contracts out Flight Simulator, so no Microsoft programmers are involved with
that.
> The average development body at Microsoft committes him/herself
> to a 55 hour week.
>
> Linux on the other hand is managed at several levels.
> There's the Kernel level managed by Linus, whereby you have
> about 200 people involved, helping him maintain the kernal.
>
> Then you have the different distributions such as Suse with 350
> people, RedHat with 400 people, Debian with 300 people, Turbo linux
> with 300 people, Mandrake with 100 people, so on and so forth.
When you say RedHat has 400, is that number comparable to Microsoft's 35,000
total, their 14,000 in R&D, or their putative 800 programmers?
> Then let's view the drivers situation. The manufacturers write
> hardware drivers for Windows operating systems and then they
> submit them to the test via Microsofts independent contracted testing
> agency.
>
> Linux either has a programmer volunteer to write the driver himself or
> the hardware manufacturer does it himself for linux.
>
> So Linux is once again clearly ahead in manhours!
I don't follow your argument here. Let me point out that I don't care whether
work is done by a Microsoft employee or a green Martian. It's man-hours that
go into Windows versus man-hours that go into Linux that I care about.
The only way Linux could be winning on the driver front is if there are more
supported devices under Linux, or if the drivers have more features. I
suppose this is conceivable if we're only talking Windows 2000, but I don't
think this is an area where either system is going to pull ahead.
> Linux started almost 10 years after Microsoft did. They are 10
> years behind. But that's not completely true as most of the
> Systems design was already layed out for them. All they had
> to do was copy the model for their own.
>
> So in code maturity, Linux is just about 2 years behind Microsoft.
> Let's say that, if it's even 2 years.
I'm not so sure. Microsoft had worked with IBM on OS/2 until their break.
The head of the NT team had previously developed VMS. Also, for several
years, the word was that Microsoft intended to "redo UNIX right," meaning
that they intended to use the best features of UNIX and improve on them.
Microsoft had access to at least as much of a foundation from previous
systems as Linux did. I've worked near university CS departments for many
years, and I can tell you that for the reasons I've stated, NT was
actually expected to be a serious advance in OS design. In the years
between about 1993 and 1996 that expectation disappeared as it became clear
that there was nothing radically new in NT and that its initial releases
were hardly even usable in the contexts where UNIX was serving. In fact,
it was about then that people stopped expecting it to be an advance and
started fearing that it would displace UNIX for economic reasons without
attaining its quality. In any case, the only argument I can put forward
in favor of Linux in this context is that Microsoft actually introduced
needless limitations by departing from what had been done before. As one
extremely smart computer scientist I knew (who had once expected great
things of NT) put it: "Microsoft has botched everything they've ever done.
NT was no exception." That was in 1997.
Another issue which you don't mention is that Linux seems to be accelerating
in terms of support. I assume you consider Linux began in 1990. NT began at the
same time, plus or minus a few years. But when Linux began it was one person,
while the initial NT team had, I think, about 300 people. (That latter number
probably includes people who draw icons and such.) The initial Linux team was
1 person. Each year that passes, the Linux community seems to grow significantly.
I'm sure the NT team has also grown some. Microsoft also has huge financial
resources to enlarge the team up to a point. Ultimately, they wouldn't be able
to compete at all if Linux were to harness the full potential of the open source
concept. I really don't have any idea whether that potential will be realized
though.
> Now there are other people working on Linux code out there submitting
> patches to these distributions I've talked about earlier. There's
> nobody from the Microsoft camp doing that.
>
> There are 25 or more ENGLISH distribution organizations for Linux per
> the WWW.LINUX.ORG site. There are MORE which are non-english.
> Let us say there are around 50 of them total in the world.
>
> Each organization has an average of 200 people working on their
> code. And for every organization, 100 more people from outside
> submit bug patches.
An important question here is whether this work is contributing to the Linux
base, or just the specific issues related to the distribution. If the latter
is the case, it's redundant work.
> The linux organizations vary from full time jobs at companies,
> to volunteer efforts such as with Debian. I will average them
> all out to a 40 hour work week which I see as fair.
>
> So 50 distributions times 200 people times 40 hours per week =
> 400,000 man hours per week there.
> And 100 people times 50 distributions time 2 hours per week =
> 10,000 of free wheeling bug patching help.
>
> Linus's group not even being included here folks.....
> 200 programmers time 30 hours a week average times 1 kernel =
> 6,000 hours a week there as a good guess. See Linus's group is
> NOT a corporation. This isn't their job...
>
> The total man hours spent on Linux is an estimated 425,000 a week.
>
> Microsoft has 500 employees times 55 hours per week =
> 27,500 man hours per week.
>
> Notice how I avoided the time spent writing drivers for specific
> hardware from this equation. See, Microsoft spends very little
> time doing this. They mainly translate other peoples work.
> While the Linux crew, if not handed code, will have to write it.
>
> So we have Microsoft at 27,500 man hours per week.
>
> Then we have Linux at 425,000 per week, EASY!
>
> That means Microsoft Corporation is roughly 7% the size of the
> Linux community and the Linux community is at most 2 years behind
> them in technology.
>
> If I were Bill Gates, I would be making plans for getting out
> of the way of Linux. I would be making post Microsoft OS plans
> for the company.
>
> Microsoft has an absolute ZERO chance for survival as an operating
> systems company 10 years from now.
>
> The writing is on the wall. $300 + bucks for Windows 2000
> Workstation, and even at retail prices $65 for a good distribution of
> Linux. And if you know how to FTP install you can avoid paying
> the $65 and get it for free.
>
> 10 years from now, we will not be talking about Microsoft corporation.
> 10 years from now, if Microsoft is still a corporation it will be an
> applications vendor at best.
>
> The BIG guns of the future are clearly Linux, FreeBSD and the BSD
> clan, and the GNU's HERD.
>
> I don't think Solaris nor BEOS, nor Microsoft will be here 10 years
> from now. They won't have a place in that market.
>
> I think the real questions to be answered will be what is the
> future of FreeBSD and BSDI....?
>
> What power can the GNU HERD posses over the linux kernel which
> would make us jump over there.....?
>
> Forget about Microsoft. They have only 1-2 more OS releases
> to go before it's boot hill for the company.
Someone at IBM estimated maybe 2 years, maybe 5, in a recent interview. But I
don't know if this was based on numbers, or just a hunch.
> And I know Bill Gates knows this. This is WHY there is only
> one OS offered at Microsoft now, Windows 2000! No more 95, 98 then NT
> nonsense. The days of thowing resources to the wind are
> over. Everyman on Windows 2000! It's a fight to the death
> for Microsoft Corporation.
>
> Isn't it amazing how one guy, Linus Torvalds conquered the world with
> his Kernel.
>
> Bill Gates wanted to accomplish that.
>
> With a Free operating system, what's the purpose in any commerical
> venture development such as Windows 2000??? How about even Solaris???
> Suppose they maintain 10,000 customers each, can they
> break even as they have to compete with the monster that is LINUX?
>
> Commerical OS situations will be about as welcome as used toilet
> paper in 10 years.
>
> And if you don't believe that, tell me how I'm wrong here!
> Show me what Windows 2000, 98, Solaris, or whatever has on
> my Mandrake 7.0 (Air) Distribution.
>
> Show me how anyone of them beats Suse 6.4!!!!
>
> You can't. They don't...
>
> Between man-hours, price, SUPPORT, Microsoft stands no chance
> of survival in the years to come.
>
> Linux has better Support. I get my answers in litterally minutes!
> I get patches the next day if not that afternoon!
>
> Bill Gates should start thinking along these lines of becoming
> an applications vendor for Linux before it's too late.
> That's where the money will start rolling in in just 2 years time.
>
> Charlie
------------------------------
Date: Sun, 02 Apr 2000 05:03:16 -0400
From: Joseph <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy,comp.os.os2.advocacy
Subject: Re: Windows 2000 has "issues"
"Austy Garhi (n. d'e-pl.)" wrote:
>
> [EMAIL PROTECTED] wrote:
>
> > Can we say "shrinkwarp?"
>
> HUH??
MS has long been criticized/characterized by corporations as behaving
like a shrinkwrap software software company. They don't understand
mission criticial markets.
------------------------------
From: "Erik Funkenbusch" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Windows 2000: nothing worse
Date: Sun, 2 Apr 2000 03:20:23 -0500
Perry Pip <[EMAIL PROTECTED]> wrote in message
news:vRBF4.1252$[EMAIL PROTECTED]...
> "Erik Funkenbusch" <[EMAIL PROTECTED]> wrote in message
> news:rmZE4.1243$[EMAIL PROTECTED]...
>
>
> > Read Inside Windows NT. MIPS used to advertise that they were the ones
> that
> > NT was developed on. But since you're incapable of doing any research
> > yourself, here:
> >
> > http://www.microsoft.com/WINDOWS2000/news/fromms/kanoarchitect.asp
> >
> > "We tested ourselves by not doing the x86 version first. We did the RISC
> > (Reduced Instruction Set Computing) stuff first. "
>
> Fair enough. However this has little pertinence to the fact that the
current
> version of NT, i.e. Win2K, is not supported on any architecture other than
> IA-32. Has it ever ocurred to you that portability has been lost due to
new
> features added.
Considering that there was an Alpha version of Win2k up until RC2 when
Compaq decided to abandon NT Alpha support, no. It never occured to me.
Additionally, Microsoft claims that the 64 bit version of NT will ship when
Merced.. er I mean Itanium ships this fall.
> > > >NT4 runs on PPC,
> > >
> > > Proof please. Provide URL's to resellers for NT4 on PPC.
> >
> > My NT 4.0 CD has 4 directories. I386, PPC, MIPS, and Alpha, with
> > executables for each.
>
> So you would recommend to your client or boss or whoever you work for to
run
> NT on PPC or MIPS?? If not, it is rather misleading as an adovacte to say
> "runs on". The reality is their are no service packs, security updates,
etc.
> for these architectures. No resellers are selling PPC or MIPS systems with
> NT installed. It is more accurate to say "used to run on".
Considering that IBM no longer makes machines capable of running NT on PPC,
no. I think the same is true of MIPS machines.
> > > >Alpha
> > >
> > > Only as 32 bit. Running NT on Alpha is as effective as running
> > > Win311/Dos622 on a Pentium III.
> >
> > And your point is? How does it's speed relate to the fact that it
exists?
>
> My point is if you are running a 32 bit OS on a 64 bit chip you're not
even
> close to using the full capability of the chip.
What does that have to do with the discussion of whether the OS is portable
or not?
------------------------------
** 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
******************************
