Open Source Maketplace Insights
John Thompson
can shed a little light on the climate of the open source community at the
moment. I believe that part of the reason that open source based startups
are failing left and right is not an issue of marketing as it's commonly
believed but more of an issue of the underlying technology.
I know that that's a strong statement to make, but I have evidence to back
it up! At one of the major corps(5000+ employees) that I consult for, we
wanted to integrate Linux into our server pool. The allure of not having
to pay any restrictive licensing fees was too great to ignore. I
reccomended the installation of several boxes running the new 2.4.9
kernel, and my hopes were high that it would perform up to snuff with the
Windows 2k boxes which were(and still are!) doing an AMAZING job at their
respective tasks of serving HTTP requests, DNS, and fileserving.
I consider myself to be very technically inclined having programmed in VB
for the last 8 years doing kernel level programming. I don't believe in C
programming because contrary to popular belief, VB can go just as low
level as C and the newest VB compiler generates code that's every bit as
fast. I took it upon myself to configure the system from scratch and even
used an optimised version of gcc 3.1 to increase the execution speed of
the binaries. I integrated the 3 machines I had configured into the server
pool, and I'd have to say the results were less than impressive... We all
know that linux isn't even close to being ready for the desktop, but I had
heard that it was supposed to perform decently as a "server" based
operating system. The 3 machines all went into swap immediately, and it
was obvious that they weren't going to be able to handle the load in this
"enterprise" environment. After running for less than 24 hours, 2 of them
had experienced kernel panics caused by Bind and Apache crashing! Granted,
Apache is a volunteer based project written by weekend hackers in their
spare time while Microsft's IIS has an actual professional full fledged
development team devoted to it. Not to mention the fact that the Linux
kernel itself lacks any support for any type of journaled filesystem,
memory protection, SMP support, etc, but I thought that since Linux is
based on such "old" technology that it would run with some level of
stability. After several days of this type of behaviour, we decided to
reinstall windows 2k on the boxes to make sure it wasn't a hardware
problem that was causing things to go wrong. The machines instantly shaped
up and were seamlessly reintegrated into the server pool with just one
Win2K machine doing more work than all 3 of the Linux boxes
Needless to say, I won't be reccomending Linux/FSF to anymore of my
clients. I'm dissappointed that they won't be able to leverege the free
cost of Linux to their advantage, but in this case I suppose the old adage
stands true that, "you get what you pay for." I would have also liked to
have access to the source code of the applications that we're running on
our mission critical systems; however, from the looks of it, the Microsoft
"shared source" program seems to offer all of the same freedoms as the
GPL.
As things stand now, I can understand using Linux in academia to compile
simple "Hello World" style programs and learn C programming, but I'm
afraid that for anything more than a hobby OS, Windows 98/NT/2K are your
only choices.
Thank you,
John Thompson
E J
Well here is my professional experience with Linux.
I don't think you can shoe horn Windows into embedded systems. Windows CE
does not like to play with other standard protocols very well and you can't
strip out the parts of Windows CE that you don't need.
There are cross compiler for many chips. So far, I have not seen Windows CE
for my embedded Power PC.
It was amazing seeing these tiny linux living on compact flashes on a single
computer board mounting network hard disk. Only the basic linux commands
work but they do work.
You are going to see more embedded system coming out based on Linux.
Check out Tivo. It is based on Linux.
I have a linksys router. It has linux built in.
Hampa Hug
[snip]
> I consider myself to be very technically inclined having programmed in VB
> for the last 8 years doing kernel level programming. I don't believe in C
> programming because contrary to popular belief, VB can go just as low
> level as C and the newest VB compiler generates code that's every bit as
> fast. I took it upon myself to configure the system from scratch and even
> used an optimised version of gcc 3.1 to increase the execution speed of
> the binaries.
gcc 3.1 is scheduled for April 15 2002. The current version is 3.0.1. It
hasn't been out long and I would recommend against using it in production
systems until the dust of a new major version has cleared.
I won't argue about VB's suitability for system programming. However, not
having at least some level of experience with C (and the programming
environment used under Linux) is not an optimal precondition for doing
system administration in a production environment of the size that you
describe.
> I integrated the 3 machines I had configured into the server
> pool, and I'd have to say the results were less than impressive... We all
> know that linux isn't even close to being ready for the desktop, but I had
> heard that it was supposed to perform decently as a "server" based
> operating system. The 3 machines all went into swap immediately, and it
> was obvious that they weren't going to be able to handle the load in this
> "enterprise" environment. After running for less than 24 hours, 2 of them
> had experienced kernel panics caused by Bind and Apache crashing! Granted,
Bind caused a kernel panic? That's hard to believe. Every single kernel panic
I've seen in four years working with Linux was caused by a) faulty hardware
or b) bad configuration.
> Apache is a volunteer based project written by weekend hackers in their
> spare time while Microsft's IIS has an actual professional full fledged
> development team devoted to it. Not to mention the fact that the Linux
> kernel itself lacks any support for any type of journaled filesystem,
> memory protection, SMP support, etc, but I thought that since Linux is
You've got your facts wrong. Linux supports all of the above.
> based on such "old" technology that it would run with some level of
> stability. After several days of this type of behaviour, we decided to
If it's stability you wanted then I don't understand why you went for
the latest software. Perhaps kernel 2.2.x along with gcc 2.95.x would
have been more suitable?
> reinstall windows 2k on the boxes to make sure it wasn't a hardware
> problem that was causing things to go wrong. The machines instantly shaped
> up and were seamlessly reintegrated into the server pool with just one
> Win2K machine doing more work than all 3 of the Linux boxes
Sounds like you seriously misconfigured those Linux boxes to me.
> Needless to say, I won't be reccomending Linux/FSF to anymore of my
> clients. I'm dissappointed that they won't be able to leverege the free
> cost of Linux to their advantage, but in this case I suppose the old adage
> stands true that, "you get what you pay for." I would have also liked to
> have access to the source code of the applications that we're running on
> our mission critical systems; however, from the looks of it, the Microsoft
> "shared source" program seems to offer all of the same freedoms as the
> GPL.
If all you want is passively looking at your system's source code that may
even be true. But then again, what's the good?
> As things stand now, I can understand using Linux in academia to compile
> simple "Hello World" style programs and learn C programming, but I'm
> afraid that for anything more than a hobby OS, Windows 98/NT/2K are your
> only choices.
W98? ROTFL.
> Thank you,
> John Thompson
perfectly welcome,
Hampa Hug
Vilmos Soti
> I integrated the 3 machines I had configured into the server
> pool, and I'd have to say the results were less than impressive... We all
> know that linux isn't even close to being ready for the desktop, but I had
I wouldn't say that Linux isn't *CLOSE* to being ready for the
desktop. Not there, in my opinion, but also not far.
> Apache is a volunteer based project written by weekend hackers in their
> spare time while Microsft's IIS has an actual professional full fledged
Apache is not written by weekend hackers. First, they have been in the
business for quite a few years. They were/are system admins and started
to write a server for their needs. They knew what they needed and sat
down to write it.
> development team devoted to it. Not to mention the fact that the Linux
> kernel itself lacks any support for any type of journaled filesystem,
> memory protection, SMP support, etc, but I thought that since Linux is
This is total bullshit. The Linus line of kernels have ReiserFS which is
a journalling fs. Also, SGI's xfs, IBM's jfs, and ext3 are also there.
It is just a question if time when they become part of the official
kernel.
As of lack of memory protection. With this single statement you
announced to us that you have no idea what you are talking about.
SMP support. Linux had SMP support starting somewhere in the 1.3.x kernels
in the mid '90-s. Definitely before NT4 came out. It was not very good,
but it was already there.
Vilmos
Thaddeus L Olczyk
wrote:
>I work as a consultant for several fortune 500 companies, and I think I
>can shed a little light on the climate of the open source community at the
>moment.
Yes. You have revealed a lot.
So let us begin.
>I consider myself to be very technically inclined having programmed in VB
>for the last 8 years doing kernel level programming. I don't believe in C
>programming because contrary to popular belief, VB can go just as low
>level as C and the newest VB compiler generates code that's every bit as
>fast.
In my experience ( 10 years on windows, 15 years before that of all
sorts of nonproffesional programming--meaning I was a student and
student labor is often free ) 98% of all VB programmers are morons.
I've even had technical leads at interviews say things like, "We can't
afford many C/C++/Delphi programmers. We get a few and the rest
of the work, the easy stuff, is done by VB programmers." Of the other
2% ( who can be very competent technically ) they all make sure to
include other programming languages in their credentials (
C/C++/Delphi/Pascal/LISP/Forth/FORTRAN/MODULA-2/Eiffel/Smalltalk ...)
to make sure that people know that they are part of that 2%.
It is dubious at best that you have ever done kernel level programming
( in fact I doubt that you are really clear as to what kernel level
programming is ), so please clarify for us. Name companies and kernel
level projects for us.
BTW the skills you list ( VB programmer ) do not coincide with the
skills of the job that you were doing ( system/network administrator
).
> I took it upon myself to configure the system from scratch and even
>used an optimised version of gcc 3.1 to increase the execution speed of
>the binaries.
Bzzzt.
There is no gcc 3.1.
So whatever version of gcc you used, how did you build it?
>I
>reccomended the installation of several boxes running the new 2.4.9
>kernel,
So how did you compile it?
( Frankly I doubt you did. If you did then you will be able to tell
me what MTRR is. One thing that it is, is something that you will
learn about when configuring the kernel. )
BTW you don't mention why you chose to compile a 2.4.9 kernel
instead of the stock precompiled 2.4.3/2.4.4 kernel that comes with
new distros, and would probably suffice?
Your one of those VB morons who thinks you know everything about
computing when you dont even know 2% of windows. You brain is fried
to the point that you equate newest with best.
However I believe this part to be an out right lie. There is virtually
no way that you compiled the 2.4.9.
( BTW if you actually did compile the 2.4.9 kernel, which compiler--
including version did you use ?)
> Not to mention the fact that the Linux
>kernel itself lacks any support for any type of journaled filesystem,
>memory protection, SMP support, etc, but I thought that since Linux is
>based on such "old" technology that it would run with some level of
>stability.
Uhm. Yes it does have SMP support. If you had built the kernel as you
claim, you would know this since SMP support is one of the first few
options you have to choose when configuring the kernel for compile.
As for journaling filesystems. There are three presently available:
ext3, reiserfs ( both open source specific ) and XFS. JFS is in beta.
XFS ( SGI ) and JFS ( IBM ) are both ports of filesystems that
have been used in large commercial systems far longer than any
Wincrap journaled filesystem has even existed.
As for memory protection. The linux kernel started out as a project
to learn about protection/memory management. 10 years ago. It has
kept up.
So what else don't you know about linux kernels. Most of it I guess.
>We all
>know that linux isn't even close to being ready for the desktop, but I had
>heard that it was supposed to perform decently as a "server" based
>operating system.
You know that but then your a VB moron so you really know nothing.
>The 3 machines all went into swap immediately, and it
>was obvious that they weren't going to be able to handle the load in this
>"enterprise" environment.
Hint put in a stick of memory, any size. Computers need memory to run.
This is about the only thing to cause a linux machine to swap on boot.
That and doing a real shitass job of compiling the kernel.
> After running for less than 24 hours, 2 of them
>had experienced kernel panics caused by Bind and Apache crashing!
Applications crashing don't cause kernel panics.
The only reason that anything such as this could ever happen is if
some did an extremely shitass job of compiling the kernel. Like a
person with the IQ of a kindergarden child. (Oops. Sorry. I apologise
to all children. )
>Granted,
>Apache is a volunteer based project written by weekend hackers in their
>spare time while Microsft's IIS has an actual professional full fledged
>development team devoted to it.
You mean like the person I met at Lucent. He was famous throughout
the company for writing a very large and critical part of 4ESS and
5ESS. When I met him, he was working one day a out of every two weeks
and spent the rest of his time learning new technologies so that he
could get a job come layoffs. Not his fault, that was what they gave
him to do and he was considered quite productive compared to people
who worked every day. I doubt it is very different at Microshit.
As for Microshits vaunted "professional full fledged development
team". They can't even write a decent framework, instead saddling
us with MFC. This full fledge development team gave us Win 1.0-3.0,
Win 95, Win95b2,Win 98,WinSE, WinME. Funny the longer they
worked the worse they got. They gave us NT1.0-NT3.0. They gave us
CE1.0-2.0, PenWindows, and whatever happened to Cairo. Everything
they have delivered over the last ten years has been way late and way
overbudget ( except for XP which they had to get out the door to avoid
the DoJ). They had to start labeling their operating systems by years
to force the "professional development staff" to release their OSs in
some resemblence of time.
As for Apache, it's being run on more web sites than IIS, is more
robust and has far fewer security holes. It has run on more servers
and handled heavier loads than IIS. Of course the problem was with
Apache and not some VB moron who was configuring the system.
>
>Needless to say, I won't be reccomending Linux/FSF to anymore of my
>clients.
Sad to say that Fortune 500 companies are hiring VB morons to
do their system analysis/system administration for them.
>
>As things stand now, I can understand using Linux in academia to compile
>simple "Hello World" style programs and learn C programming, but I'm
>afraid that for anything more than a hobby OS, Windows 98/NT/2K are your
>only choices.
Acadameia like JPL perhaps, who trust their multi-billion dollar space
systems to it.
>I believe that part of the reason that open source based startups
>are failing left and right is not an issue of marketing as it's commonly
>believed but more of an issue of the underlying technology.
>
What you believe is irrelevant, you're a VB moron.
The real reason that startups are failing is because they are being
run by these people with dreams of making it rich who only know a bit
of the technology. The best analogy is Apple, created by two guys
a techie (Woz ) and a business-techie (Jobs). As long as the techie
was in charge of the technical aspects (Apple II) the company did
great. When the business-techie (Lisa,Mac) took over the company went
into the toilet.
The one company I worked for the guy knew all the technologies
by name. But when asked what was the one thing that the application
had to do to make sure that it the company would survive, he could
never state it. Instead he would start to, then he would say and I
want this technology in there. Early on we could have had three or
four use-cases to implement, we could have implemented them and
he would have a demo which he could have used to raise more money.
Instead the developers just pissed around creating pieces that never
fit together.
On the radio this week I heard a guy talk about his startup. One of
the key pieces depended on technology that a company a friend
works at is developing. The friend says the technology won't be ready
for prime-time for at least ten years. Yet this guy acted as if it
worked.
That is why a lot of startups are failing.
I have been using computers for over 25 years. Most of the systems
that I used I have had to configure some important piece if not most
of it. Configuring Linux ( which includes things like rebuilding the
kernel ) is one of the easiest system I've ever had to work at.
Sure it's easy to configure Windows using dialog-boxes 98% of the
time. But when a company needs that other 2% they will discover
how badly they wasted their money paying for the cheap "dialog-box
kiddie"/VB moron instead of the expensive knowledgable system
administrator/system analyst. Hey you get what yopu pay for, and
in case you don't get it, all software is cheap compared to people.
Peter T. Breuer
> John Thompson wrote:
> [snip]
>> I consider myself to be very technically inclined having programmed in VB
>> for the last 8 years doing kernel level programming. I don't believe in C
It's clear that this person has a view of himself that is relative to
his working environment. No coder would do what he did. As professor of
computer science and telecommunications engineering, and an expert in
hardware specification and software engineering in particular, I can
tell you that I've met people with that kind of inflated view of
themselves before, and yes, to their eternal puzzlement, they are
always unable to do the simplest things (or the most difficult)!
>> programming because contrary to popular belief, VB can go just as low
>> level as C and the newest VB compiler generates code that's every bit as
This would meake many people laugh. VB is not a suitable programming
language for the areas in which C is used (kernels ..) - no more than
ava is. Mind you, compiled java can go very well in applications.
>> fast. I took it upon myself to configure the system from scratch and even
>> used an optimised version of gcc 3.1 to increase the execution speed of
This is silly and pointless - you want the OLDEST version of gcc, not
the newest. The standard is gcc 2.95 (or 2.91). I use gcc 2.8.1 and
2.7.3 . And you should know that optimization in general does you no good
on a 686 platform! Compiling for 386 and removing nops and doing loop
unrolling would probably be best. If you don't know why, 20 demerits to
your self-esteem.
> gcc 3.1 is scheduled for April 15 2002. The current version is 3.0.1. It
> hasn't been out long and I would recommend against using it in production
> systems until the dust of a new major version has cleared.
I would't dare to use 3.0 yet.
> I won't argue about VB's suitability for system programming. However, not
It's unsuitable for o/s programming, as we know. How many articles in
systems journals claim that the author used VB to do o/s programming?
None. For systems tasks, shell scripts, perl, VB are all much of a
muchness. I prefer shell.
> having at least some level of experience with C (and the programming
> environment used under Linux) is not an optimal precondition for doing
> system administration in a production environment of the size that you
> describe.
Well, of course not.
>> I integrated the 3 machines I had configured into the server
>> pool, and I'd have to say the results were less than impressive... We all
>> know that linux isn't even close to being ready for the desktop, but I had
>> heard that it was supposed to perform decently as a "server" based
>> operating system. The 3 machines all went into swap immediately, and it
You used a 2.4.9 kernel. It's known that its VM is broken and will do
that under several coditions. Go to 2.4.4 if you want sane VM
behaviour, or try alan coxes kernels. Why are you using 2.4 in a
production env? It's safe enough for someone who knows what they are
doing, but you don't. In your place I'd back down to 2.2 in a hurry, or
turn off swap.
>> was obvious that they weren't going to be able to handle the load in this
>> "enterprise" environment. After running for less than 24 hours, 2 of them
>> had experienced kernel panics caused by Bind and Apache crashing! Granted,
Impossible. They can't be the cause, thy must be the effect. If you
make such a basic mistake about cause and effect, your qualifications
must be questioned. Something is wrong! Bad memory, perhaps? Overheating?
All very common.
> Bind caused a kernel panic? That's hard to believe. Every single kernel panic
> I've seen in four years working with Linux was caused by a) faulty hardware
> or b) bad configuration.
>> Apache is a volunteer based project written by weekend hackers in their
:-). Some kind of joke? Apache is written by all the sysadmins in the
world, working together. It's vastly better authored than MSs amateur
offering, as well as running on every system out there. Are you a FUD
monger?
>> spare time while Microsft's IIS has an actual professional full fledged
>> development team devoted to it. Not to mention the fact that the Linux
>> kernel itself lacks any support for any type of journaled filesystem,
>> memory protection, SMP support, etc, but I thought that since Linux is
> You've got your facts wrong. Linux supports all of the above.
Certainly!
>> based on such "old" technology that it would run with some level of
>> stability. After several days of this type of behaviour, we decided to
> If it's stability you wanted then I don't understand why you went for
> the latest software. Perhaps kernel 2.2.x along with gcc 2.95.x would
> have been more suitable?
>> reinstall windows 2k on the boxes to make sure it wasn't a hardware
>> problem that was causing things to go wrong. The machines instantly shaped
>> up and were seamlessly reintegrated into the server pool with just one
>> Win2K machine doing more work than all 3 of the Linux boxes
> Sounds like you seriously misconfigured those Linux boxes to me.
And to me.
>> Needless to say, I won't be reccomending Linux/FSF to anymore of my
Anyone who produces such a silly result merely disqualifies themselves
from serious consideration. Do you think the rest of the world is
hallucinating? Well, yes - most humans do have a tendency to think so!
Specially ones not used to computer science, I have observed. Most
people initially blame the compiler, not themsleves.
>> clients. I'm dissappointed that they won't be able to leverege the free
>> cost of Linux to their advantage, but in this case I suppose the old adage
>> stands true that, "you get what you pay for." I would have also liked to
>> have access to the source code of the applications that we're running on
>> our mission critical systems; however, from the looks of it, the Microsoft
>> "shared source" program seems to offer all of the same freedoms as the
>> GPL.
> If all you want is passively looking at your system's source code that may
> even be true. But then again, what's the good?
>> As things stand now, I can understand using Linux in academia to compile
>> simple "Hello World" style programs and learn C programming, but I'm
>> afraid that for anything more than a hobby OS, Windows 98/NT/2K are your
>> only choices.
> W98? ROTFL.
This is simply one of those paid FUD mongers. Why would he bother to
post such nonsense otherwise! One has to question why these people go
around on the linux newsgroups doing this. Does he think that my
several hundred linux systems handling http and email and a dozen
services each are will 'o the wisps? If so, then I've been seeing
things for the last eight years.
Peter
Eric Y. Chang
...
: pool, and I'd have to say the results were less than impressive... We all
: know that linux isn't even close to being ready for the desktop, but I had
: heard that it was supposed to perform decently as a "server" based
: operating system. The 3 machines all went into swap immediately, and it
: was obvious that they weren't going to be able to handle the load in this
: "enterprise" environment. After running for less than 24 hours, 2 of them
: had experienced kernel panics caused by Bind and Apache crashing! Granted,
Oddly enough, most of my experiences with Linux has been for desktop use.
When trying to use it as a server, we had the same experiences: Linux
crashed hourly, while Win2K was rock solid. The odd thing was that only
the RH 7.1 boxes failed. The kernels were still running, but the stack
was dead, and there was no way to telnet in. Since these machines did not
have consoles, they had to be power cycled.
A quick search of google revealed some sort of kernel bug, but all the
posts were of the "me too" type, so we did not know how to fix it.
Some clever system administrator figured out that the problem was
caused by a flaky Win2K DNS server causing all the Linux NFS mounts
to go down. When the Win2K box was rebooted, everything was OK.
Another odd thing was that it only affected the RH 7.1 installs.
About a week or two later, the Linux servers started dying like
flies. This time, we rebooted the Win2K server, AND installed a
caching DNS server on Linux. No problems to date :-) :-) :-)!!!!!!
So the upshot of all this is that you are probably a bit right.
This whole affair occurred on a live installation and it was
scary to see it go down right before our eyes. Even though it
was Win2K that was ultimately responsible, lack of knowledge
about Linux was the real cause of the disaster. In other words,
Linux was not a good solution due to the relative inexperience
and unawareness of the system administrators involved. Just like
you. That doesn't mean that it is useless. It works for so many
installations that something must be OK, but I would not stake
my reputation on it, at least for use as a server, if there is
not an experienced system administrator available. I would,
however, trust it fully on the desktop.
...
: As things stand now, I can understand using Linux in academia to compile
: simple "Hello World" style programs and learn C programming, but I'm
^^^^^^^^^^^
Many of the really sophisticated stochastic simulation programs, finite
element modeling programs, and irreversible thermodynamics computation
programs fall in this category since they don't use etched look buttons
and cursors that cast a shadow. Yes, they are often written by people
who are in academia and are learing C programming :-), but you tend to
trivialize them. These programs are not trivial, and your attempts to
devalue them is distressing.