Slashdot Mirror
CNN on Microsoft and Linux
noise writes "Article that details Microsoft's continuing anti-Linux campaign up through Ed Muth's comments last week. There are some nice OS and web server usage statistics, as well as some information on the difficulties that MS will likely have with the 64-bit version of NT. "
While this article is pretty good, I have issues about a few points. First, Beowulf clusters make heavy use of message passing libraries such as MPI and PVM for clustering (these are not fault-tolerant or failover by nature). These are also available on Windows NT. Second, for those who are parallel programmers, you know that ray tracers are by nature embarrassingly parallel. Algorithms like that adapt well to any network and are scalable by nature. The IBM benchmark basically compares execution times of a DEC Alpha vs that of an x86. Those of you who program Alphas know that extra effort needs to be used when writing code for the Alpha so that you get good performance. Seeing the IBM cluster running an LU decomposition or a CG solver would be more interesting. Also, seeing code that was coded with lots of care for the Alpha against similar code for the x86 would be interesting. Third, those of you who have dealt with performance measuring know that 100Mbit Ethernet and 10Mbit Ethernet have almost the same latency. The main advantage is an increase in bandwidth. For parallel computing many Linux clusters use Myrinet. Similarly, Windows NT clusters use Giganet and Myrinet. A similar cluster to this IBM cluster with either of those two networks is far more scalable on communication intensive applications (embarrassingly parallel applications won't notice that much difference as there is little communication). There are a variety of other issues that I have with the article. I am not advocating either OS, just pointing out some points that came to my mind while reading this article. Many of the other points are valid and overall, the article is pretty good. I cringe many times when I hear people mentioning Beowulf, especially in reference to the IBM 'scalability' benchmark recently (pure marketting hype). Just for reference, the Cray T3E using MPICH (commonly used on Linux) gets ~6.5 (yep, six point five) microseconds latency and ~321 megabytes/sec bandwidth (that's right, three hundred twenty one megabytes per second). Compare this to the typical TCP/IP MPICH on Linux of ~180 microseconds of latency and ~10 megabytes/sec bandwidth and then think what a communication intensive application would do.... Now, compare this with carefully written code on an Alpha-450 (the kind in the Cray-T3E 450 - should get around 400 MFLOPS easily) to what you normally get on a PII-Xeon (less than 100 MFLOPS). I must admit, though, bad code on an Alpha-450 would be around 60 MFLOPS.
Wake up people. Mr Petreley is hardly "unbiased". He's been pounding away at Microsoft for years, as editor of "Linuxworld", and before that the ill-fated "Network Computer World". I like to see people pounding away at Microsoft, but that's where he's coming from.
Furthermore this story is a bit off-base, as his stories often are. Is Linux considered "splintered" because it comes in 64-bit flavours? Of course not, so why should Windows be? And surely the ability of some future Win64 to emulate 32 bit code should not be counted as detrimental.
Also, counting the numbers of machines running various Web servers and operating systems and attached to the Internet can't be extrapolated to, say, what IT people like to use generally. So let's not get carried away.
Anyway, what I'd really like to know is how many big NT sites use it because they receive "incentives" from Microsoft in cash and kind. At least, four years ago this was happening at some companies I was close to.
roc+sd@cs.cmu.edu
"Ease of use" depends on who you ask. On the desktop, Microsoft likely still has the upper hand--but KDE and GNOME mean that said upper hand is slipping. Things look even bleaker for Microsoft if uptime is a part of ease of use. No system is easy to use when crashed or wedged.
Much of the article was looking at NT the server, though. When running a server (as opposed to running a client that happens to talk to the server), the base assumption is that the machine is run by a professional computer operator, such as a sysadmin. From the sysadmin's perspective, a Unix or Linux server tends to be _much_ friendlier than an NT server. Most of this comes from the fact that the GUI is not wed to the backend as it is on Windows.
For one thing, the server is friendly enough to talk to you at _your_ console, not _its_ console. If I'm running a dozen NT boxes and have to do something to all of them, I have to log into a dozen consoles. To do the same job on Unix, I use X to "beam" windows to my machine from all of them.
Another "ease of use" advantage in the server world is scriptability. Most anything you can do on a Unix machine, you can do by manipulating text, whether it's manipulating data streams or configuration files. Since text generation is a simple task, this means that you can build your own macros and UIs to do tasks specific to your installation. When talking to an NT application, you have to click the buttons in their preferred order.
At the low end of the scale, desktops for casual users, Microsoft holds the upper hand for ease of use. At the high end of the scale, servers, Unix beats NT. The turning point is probably close to where the power users sit, depending on one's definition of power user.
I partially agree with you, but I'd like to share an experience we had in our office converting a machine over from Win95 to Linux.
We have a group of employees that don't work in the office and just come and go at different times in the day. We had one win95 machine for about 10 of these employees to come in on break and check their email and surf the web.
With win95 the machine was always crashing, and we had many problems with win95 not being multi-user.
Eventually we moved the machine into Linux and set up each person with their own username, password and Windowmaker desktop. Yes, we didn't even bother with KDE or GNOME (well we did have GNOME installed for one of the power users, but he actually requested that we remove it and just leave Windowmaker). We setup a button for Netscape, a button for an xterm and a button for an office app we occasionally use. We'll probably add one more button to let them use WordPerfect.
The largely computer illiterate users couldn't be happier! We set them up with a Windowmaker theme of their choice and showed them how to log into the school server and check their email with pine. We haven't had a single problem with this setup. Most of them probably really don't understand that they are even using a Unix varient. They just click on one or two buttons and get the things done that they need to do, and they know they won't crash the machine or break anything accidently.
This may not be a typical office experience, but it shows how easy it was to migrate users that weren't familiar with computers at all. I suspect if they had been like many Win95 users they would have complained bitterly because the system doesn't work exactly like Windows does.
I think many people forget that Windows was hard to train people on as well. In fact, I have a hard time sitting in front of a Mac machine. It is definately not intuitive to me. It took me quite some time to figure out how to eject the floppy disk! What common sense was I supposed to use to determine that to eject the disk I drag the icon into the trash can?
Linux wasn't intuitive for me to learn, but now that I have learned it, it's no harder than anything else I've done with computers.
Microsoft definately had a clue when they were still competing with Word Perfect. It let you map the Word Perfect keybindings to the new Word ones as well as giving the user help throughout to help them migrate. It's just as easy to do the same thing in reverse, and I hope it will be done!
The more you know, the less you understand.
Remember when 32 bit processors came out the world was perfectly content to stick to 16 bit Win 3.1 for years until Microsoft "invented" the 32 bit OS. When Intel finally "invents" 64 bit chips the managers and CEO's of the world are going to be perfectly happy running a 32 bit OS for many years to come indeed They already are if you count Sparc and Alpha.
Still, I found the article to have a lot of anti-MS FUD in it, which was disappointing. Linux doesn't need to spread FUD to "win." MicroSoft is something that should be ignored, not beaten up upon.
For example, for the author to claim he is quoting "one of Linux's most ardent supports", when he is quoting a MS executive's trial testimony is bogus. MicroSoft doesn't support Linux, although they do want to say that someone could be considered a competitor to them. Yes, MicroSoft is being bogus also, but what MicroSoft does is irrelevant and can't excuse what Linux supporters do.
As for Linux not having a "long term road map", that is true. Linus recently said that he wasn't sure what was going to go into v2.3, and what will actually end up in it is anyones guess. I don't think this is a problem, but to turn around and effectively say "Yeah, but MicroSoft doesn't have one either!!11!!!" is lame.
As far as the "cost per transation" goes, MicroSoft is right that the price of Linux is meaningless. Anyone who has done real large scale roll-outs knows that the vast majority of the costs involves the labor and training expenses. While you can argue fairly strongly that Linux can have a much smaller labor expense to create a server/system, the training is not as easy to dismiss. Linux has real weaknesses, as well as real strengths. The price, when it comes down to it, is not that big of an issue.
The area of 64-bit support is also something that Linux may be "better than WinNT", but that isn't saying much. There is a lot of Linux software that doesn't run correctly on Alphas/Sparcs. There isn't even very many distributions that can claim to support these 64-bit machines.
While it is possible for MicroSoft to mess up their 64-bit support, I find it somewhat unlikely. The problems with going from 16-bit segments to a flat 32-bit model are much different than going from a flat 32-bit model to a flat 64-bit model.
All in all, I found reading this article to be much like a one-night-fling. Yeah, it was a lot of fun, but I kind feel "dirty" afterwords. There was just too much anti-MicroSoft stuff, and not enough "this is why you should use Linux". What little of the later there was, it was mostly done by tearing down MS.
SPF support for most open source mail servers can be found at libspf2.
It would seem that any large body with a lot at stake (let's use a politician and his/her government as an example) will gamble by saying things that aren't necessarily true (or NOT saying things that are true), in the hopes that the heat will be turned off and things will go back to normal. Some examples from recent history include President Clinton's escapade with Lewinsky, and Premier Clark's [government's] numerous scandals (the ferries, the casino, etc). Until they know there's no turning back, they'll swear that black is white.
Such is the case with Microsoft. While they believed they were still on pretty solid ground they could quietly overlook any shortcomings of their own OS, and form arguments (that may or may not be convincing) to see them coming out on top. They could deny that Linux was a threat because, as far as they were concerned, there was no evidence to support that idea.
A breaking point came where they could no longer deny what was really going on, and unless they were to acknowledge what everybody could see was happening (That Linux does "challenge the industry leader"), they'd look rather silly.
Of course, your guess as to what comes of this is as good as mine. Many Politicians who fib manage to stay (or get back) in office; maybe Microsoft will as well.
But, my vote has been with Linux since kernel 1.2.x, and i'm not changing anytime soon.
"Give me liberty, or give me death, Zogwarg Queen!" - Spiff.
Okay, it looks clear that as long as Linux development continues as it is, and Microsoft doesn't find some miraculous way to dig themselves out from years of bad architecture decisions, Linux will win (has won?) the server war.
Now, however, it's time to win the desktop. Between KDE, GNOME, and WINE development, I personally see Linux at taking a serious chunk of Windows desktop market share in two years, and maybe breaking the 50% mark in three years. What you do folks think, reasonable?
Really good stuff if you have the time to read it
support gun control: take guns from cops
Actually, if you go here:
http://www.sunbelt-software.com/0399_ 2000.htm
You can see that when they noticed that the statistics were getting skewed towards Linux, they stopped using those statistics. Only the first 2000 responses were used. These responses were made before word of the survey got out to the Linux crowd.
There's no ballot stuffing here.
I think this is the first SOLID piece of evidence I've seen showing how much trouble M$ is in in the OS market. There's no M$ bashing. Just concrete facts, as far as I can see. The author deserves applause for writing such a brilliant piece. I would think, more so now than ever, the folks in Redmond are chasing their tails more and more, not knowing how to stem the tide of Linux. The simple fact is that Linux is going to roll right over them. The momentum is too strong to fight. Cheers to all you folks developing Linux apps!!
Granted. Personally, I've been using Linux for a little over 6 months now and I would never go back. (Well, actually I do...but only when I need to edit complex Office 97 documents)
I agree with you that Linux is a ways away from being what I would put on my mother's computer either, but consider it's roots.
As we all know, Linux is a variant of UNIX, which was originally written by developers for other developers. Computers weren't household appliances. They didn't need to be "idiot friendly", they needed to be "professional friendly". And, due to hardware constraints, it had to be a lean and mean.
Windows on the other hand, traces it's lineage back to the inception of the PC, and QDOS. PCs, in contrast to workstations, had to be more "idiot friendly" than UNIX, because they were not targeted exclusively at the professional.
Now, if we bring this forward to the inception of NT, we find that Microsoft has seen the writting on the wall when it comes to the "PC" market. They can see that they have saturated the home market. In order to continue making the unbelievable amounts of revenue they have been sustaining, they need to A) Force the people who have bought their product in the past to upgrade to a new version, and/or B) Move into the server market, where the big bucks are. Microsoft decided to do both. (Hence NT, and eventually, the 9x line)
Meanwhile, the commercial UNIXes were happy as clams in a pond. They WERE the server market. There were many variants available, and the competition kept things fresh, but there was little reason to make UNIX "idiot friendly". NT (and more importantly, NT's marketing) caught them off guard. All of a sudden, NT was on it's road to becoming the big game in town, and the various UNIX vendors began loosing that ever so important market share.
Now, we bring this forward to present day. Many companies who standardized on NT have knowingly, or unknowingly been forced to bring back UNIX. Linux is rising in popularity and, for the first time, there IS a reason to make it "idiot friendly", if only to end Microsoft's stranglehold on the PC market.
The bad news is, Microsoft has been making "user-friendly" OSes complete with user candy for quite some time now. Arguably, they're pretty good at it.
The flip side of the coin is that UNIX/Linux has been doing networking for quite some time now. And I think we'll all agree, it's pretty damn good at it.
So what we have is two seperate systems, built for completely disparate environmnents encroaching on each other's living space. (Imagine a whale being forced to walk on land and an elephant being forced to swim.) And at this point, they either evolve, or you end up with a lot of beached whales and drowned elephants.
It comes down to a question of how well each can adapt to their new environment. Can NT become more reliable, scalable, and robust? Can UNIX become more user friendly, intuitive, and simple?
Linux, being open source has a distinct advantage. It can (and has) evolve very quickly. Look at where Linux was in terms of use-ability a year ago, compared to now. Then, look at NT in terms of stability, and performance a year ago, compared to now.
So, while Linux still has a long way to go, I'd much rather wait for it to become more user-friendly than I would for NT to become more scalable.
These are interesting times. Enjoy them.
"What do you mean, invalid parameters? 9000Gigs of RAM and it can't answer a simple question!" -- Earthworm Jim
Several people have pointed out that this survey could have been slashdotted.
Well, it wasn't. Here is the intro to the survey results from Sunbelt:
Sunbelt March99 Survey Results
over 1999 NT users
And now, the answers to the survey questions! Before anything else,
thanks for your many thousands of survey answers. This is a hot
topic for sure!!
First though, some background data. This survey was meant to
get an idea how things are looked at BY the NT community, FOR
the NT community. Well, that was somewhat naive I have to admit
. The questions were created while looking at discussions
between NT system administrators and were definitely written
from that viewpoint. Some claim that they were biased. There may
be some truth in that if you look at it from the Linux POV.
Sunbelt does not claim any kind of scientific validity. This
survey is a snapshot and not a random sample at all. It is not
objective but that was not the idea in the first place.
BUT, I guess if you step into a war you should expect some bullets
flying around, and I did get some flak from people. Most of these
people are both running NT and Linux and know them well. Personally
I know NT but have not much experience with Linux. We have a few
'closet' Linux users in Sunbelt though, both in sales and Tech .
Anyway, what happened is that we sent the invitation to do the
survey, and very quickly got thousands of responses back. We
followed in real-time what the results were, every 500 responses
or so. Extremely surprising numbers to start with. But even more
strange after the first few hours. Numbers started suddenly to
change and become slanted toward more Linux than before. Very odd
from a statistical perspective I remember from my stats course
in college.
Now, it so happens that the software we use to do the survey
queries the browser that was used to fill out the survey. So
we know what O/S it is running on, and the IP address it comes
from. The Linux user community had gotten word about the survey
and was getting into gear to show that Linux had support.
So anyway, we took the first 2000 survey results and did our
analysis on those. These were from predominantly NT users with
some people answering from a Linux machine. That will give some
idea about the current state of Linux use and the results are
revealing.
/snip/
And here is my own favorite set of statistics, about the supposed
advantages/disadvantages of Linux:
How important are these perceived Linux Benefits?
Scale= Very Important / Important / Of Less Importance
Linux is free
775.00 / 523.00 / 542.00
Stability
1584.00 / 195.00 / 58.00
Not Microsoft
340.00 / 290.00 / 1178.00
Better performance
1356.00 / 381.00 / 89.00
Open Software
1148.00 / 459.00 / 225.00
Other
619.00 / 294.00 / 294.00
How important are these perceived Linux drawbacks?
(Same scale)
Lack of Applications
463.00 / 449.00 / 882.00
Not Mission Critical yet
314.00 / 333.00 / 1099.00
Lack of Third Party utilities
324.00 / 352.00 / 1084.00
Different distributions / GUI's
261.00 / 299.00 / 1197.00
Lack of standardization
386.00 / 315.00 / 1060.00
Open Software
257.00 / 281.00 / 1214.00
Lack of Tech Support
324.00 / 312.00 / 1119.00
Too 'Bleeding Edge'
129.00 / 256.00 / 1347.00
Steep Learning Curve
238.00 / 401.00 / 1116.00
Hardware incompatibilities
371.00 / 404.00 / 985.00
User unfriendly
270.00 / 385.00 / 1101.00
It's clear he has a good handle on the strengths of Linux in comparison to today's WinNT, as well as the NT yet to be developed.
As usual, it's noted that so much FUD from the Micros~1 people downplay the importance and significance that Linux has in current Internet applications as well as home-users' desktops. MS seems to have just discovered that there indeed are GUIs for Linux, and there are applications comparable to most anything that MS has already produced. And Linux almost always does it better.
To quote:
And the reasons stating the superiority of Linux go on and on. Really, if you look at those numbers and have used Linux, the numbers shouldn't be a surprise to you. Can you really beat an OS that's not only free, but has thousands of programmers continually fixing, updating, and supporting it?
viva Open Source! viva Linux!
I only post comments when someone on the internet is wrong.
Be honest now, if Hotmail used Linux instead of FreeBSD for their web servers, do you think Solaris would have been mentioned at all?
No, it's just that the significance of Microsoft using Solaris of all things is much greater than that it's a unix. It's Sun's unix, and Microsoft hates Sun.
That they're using Solaris is especially telling to the audience of this piece. The use of FreeBSD may be of interest to us, but mentioning it to the PHB's just dilutes the impact.
-- Alastair
Linux still has a long way to go before it will beat Microsoft's ease of use. I have been using Linux for about a year now, and I still find myself spending a lot of time trying to figure out the operating system.
Furthermore, it's still quite a ways from being something that I would want to put on my mom's computer.
A lot has happened in the last year, but I would say it'll be 2 or 3 more years before it will make significant gains on "user's" desktops. Perhaps Corel's distribution will be more "idiot friendly." If so, that will be a reason for many techies to turn against the Corel distribution, and shoot themselves in the head.
This confirms what I've been telling friends for a while now: Linux will beat NT handily in the server market, and this will be obvious to everyone very, i.e. in 1999 or 2000 at the latest. Microsoft has been hemoraging credibility right and left over the past year, and their one-year-and-growing tardiness in delivering NT5.0, er Win2000 is going to hurt them. Unfortunately, M$ will continue to dominate the desktop client market for quite some time, due to momentum and the difficulty and expense of retraining users to use a new client platform. (No flames please, this is just my opinion.) I suspect that this market will decrease in significance as networks of "information appliances" take on more and more of the functionality of the traditional standalone desktop computer. (The PC will survive in the form of a gateway between your home network and the internet, but currently Linux is a much better gateway platform than is Windows.) Microsoft's biggest mistake was to try to position "Windows" as the solution to all possible computing needs. It is clearly architecturally and structurally deficient to scale from palmtops to supercomputers. The customers would have been better served if M$ had concectrated on client side software (the only area where a GUI is an advantage) instead of trying to be all things to all people, and failing to completely implement any of it's promises in the process.
"Freedom means freedom for everybody" -- Dick Cheney
That's a bunch of bs! KDE is adaquate for mom to use whenever she visits. All I had to tell her was her username, password, how to start ppp (kppp), and log off. I only had to show once
and she was on her own. She even waited for "System halted" to show up before turning off the
computer.
The thing I find most difficult for windows users (moving to linux) to understand is the multiuser concept. When I tell mom and dad to log off and let someone else use it, they ask me "why?" This is something new to them.