Slashdot Mirror
Microsoft Clarifies Linux Myths
mikeraz told us about Microsoft's new page on
Linux Myths. Designed to convince the seasoned business professional that Linux doesn't measure up to the hype. Some good points. Some not-so-good points. Care to comment?
Update: 10/05 06:43 by CT : Mandrake has comments on
his page.
Hmm. This page looks an awful lot like a private anti-Linux pro-Microsoft page I ran across recently. I'll see if I can find the URL and post it here.
My responses:
1) RE: Benchmarks. There are three types of lies: lies, damned lies, and benchmarks.
2) WTF is a "proof point"?
3) A journalling file system does _not_ prevent data loss! It simply makes fscks faster (a worthy goal in itself, but much short of the walking-on-water powers normally associated with journalling file systems).
4) I wouldn't brag about the uptime gaurentees provided by HP, at least- they gaurentee that uptime _only_ running a small number of "certified" apps and only one a small number of machines and configurations. And you pay for it. And they don't count scheduled downtime. And 99.9% is sucky- that's like 9 hours of (unscheduled) downtime a year. No wondering clustering is so important to NT.
5) The only TCO study I've ever seen showing that NT is cheaper than Unix (which was included in the MSDN, BTW) among several other humorous assumptions, assumed that anyone using Unix on the desktop also needed to have a Windows box on their desk as well for office apps, email (Unix doesn't have email, don't you know...) etc. In other words, TCO(NT) = TCO(Unix) + TCO(NT). Pay no attention to that man behind the curtain!
6) I'll be the first person to agree that ACLs are better than Unix's protection scheme. The only problem with NT's protection scheme is that you can't use it. Little known products like Microsoft Office need all sorts of rights, like the right to write in C:\Windows. Yeah, if you know what you're doing, you can mostly lock down an NT box (leaving only a few gaping security holes absolutely required by the applications), but most people _don't_. Yeah, Unix's protection scheme could be signifigantly improved, but at least it's used.
6) I like the weasling on the C-2 security ratings. Last I heard, NT only got C-2 by disabling the networking. And 4.0 wasn't certified for years- so long the guy who was originally suing them was threatening to take them to court. Oh, and C-2 doesn't mean that much- the only lower rating is D (no security at all, ala DOS). Some versions of Unix can even hit B levels of security (there's a version of HP-UX that does this, IIRC). The only reason Linux hasn't gotten C-2 is simply that no one has walked it through the paperwork (i.e. for the same reason it's not "Unix").
7) Um, NT doesn't support PnP either. My experience has been that Linux has support for more hardware than NT. And is gaining on 98 (especially since 98 seems to be dropping support for older hardware). And hardware support is mainly done by the hardware vendors, not by Microsoft- Linux is starting to pick up support from the hardware vendors. For instance, I'm willing to bet that Linux will be native (and 64-bit, but that's cheating) on Merced before Windows will...
8) Funny, I'm using Linux on the desktop even I type this (at work, even). For my work Windows would be _less_ effective. But I'm willing to agree that I'm a special case. "Easier to administrate" is a tricky concept- administrating one Windows box (or one Mac) is easier than administrating one Linux/Unix box, agreed. But administrating 500 Linux/Unix boxes is signifigantly easier than administrating 500 Windows boxes. Things which make life _more_ difficult in the single-box (like using text-files for configuration) suddenly change to be helps in the 500-box situation (for example, letting you write scripts to automatically reconfigure all those systems).
Once you start ignoring administration costs, I'd pit Gnome/Enlightenment for user friendliness against anything Windows has to offer. Especially considering there is no user interface rule that Microsoft themselves don't violate (for example, the DDK viewers don't even have menu bars, just tool bars).
And the question is not "how many applications are on a given OS", but "are the applications you need on the given OS?"
Reality: Linux Needs Real World Proof Points Rather than Anecdotal Stories
...
Customer Testimonials
See how these leading companies and organizations have deployed Windows NT Server 4.0:
Nasdaq
Barnes & Noble
Dell Computer Corp
The Boeing Company
First Union
Chicago Stock Exchange
How ironi.
-----
First, it's worth noting that Linux is a UNIX-like operating system. Linux fundamentally relies on 30-year-old operating system technology and architecture.
Humans fundamentally rely on billion-year-old architecture in their genes, so they should be scrapped and reinvented.
Linux was not designed from the ground-up to support symmetrical multiprocessing (SMP), graphical user interfaces (GUI), asynchronous I/O, fine-grained security model, and many other important characteristics of a modern operating system.
Linux started out sleek, lean, and unbloated, and then evolved into a viable OS. In contrast, NT started out with bloated features you don't need in a server (running a GUI 24/7), and can barely support its own weight (finding bugs in NT is like finding a blade of grass in a cubic acre of hay).
-----
Reality: Free Operating System Does Not Mean Low Total Cost of Ownership
It does if "when we speak of free software, we speak of freedom, not price". Unless the latest version of NT comes with working source code, I can't tweak the OS, or any underlying components to my liking.
-----
Myth: The Linux SWAP file is limited to 128 MB RAM
Reality: Go read the mkswap(8) man page: the OLD (v0) swap spaces could only be 128MB (on an i386). The new ones (v1) support up to 2GB. Get your facts right, Microsoft. This has been around since kernel 2.1.117.
-----
There are no commercially proven clustering technologies to provide High Availability for Linux.
Can you say beowulf?
-----
This is made complex due to the fact that there isn't a central location for security issues to be reported and fixed.
Try dselect update install on any Debian system, or the RedHat errata, etc.
------
No Linux products are listed on the U.S. Government's evaluated product list.
Because we don't care about the U.S. Gov't stupid encryption "security" policies, approval would only make Linux look more insecure.
-----
Configuring Linux security requires an administrator to be an expert in the intricacies of the operating system and how components interact.
Or an admin with a port-scanner.And you tell me an NT box is secure without an admin who knows about the system? Nice try.
-----
Reality: Linux Makes No Sense at the Desktop
We never said it would... but my sister seems to like her Windows-free desktop (and she is totally OS-indifferent).
-----
Windows NT 4.0 currently supports over 39,000 systems and devices on the Hardware Compatibility List
Assuming they were bought in the last 3 years, otherwise, it's obsolete.
-----
Linux does not support important ease-of-use technologies such as Plug and Play, USB, and Power Management
Sad, my Linux box seems not to exist.
-----
The Linux operating system is not suitable for mainstream usage by business or home users.
We never said it was good for mainstream by home use, we projected by current developmental progess that it WILL be soon.
And business? Go ask RedHat if businesses don't like their software.
--------
"I already have all the latest software."
What they're talking about is that you queue an I/O, and you get a notification when it completes. I.e, instead of calling write and expecting a result, you call something like write that sets a semaphore, and semaphore is released when the bits are actually on disk and the filesystem metadata reflects that. In classic Unix, you really can't do this - write just stuffs the bits in the buffer, and only blocks if there's no buffer space, so a return from write doesn't mean the data is on disk. You can call fsync, but then you have to wait. You can wait in a thread, but this gets ugly if you're doing a lot of transactions at once - you start filling up your process table. The same is true on read - you can't really doselect on a disk file, for example - it'll always say ready, even though it isn't, and even if you hacked select to say "yes, there's data in the readahead buffer," there's no way to ask how much, or to say "do you have the 4096 bytes starting at location 131072 yet?"
This is a really nice feature, and is something for which I believe there is a Posix spec. Unfortunately, neither Linux nor NetBSD currently implement it. Database vendors in particular consider this to be A Big Deal. The ISC DHCP server has this problem, because it can't grant an address to a DHCP client until it's sure the record of the grant is on disk, which has to be done with fsync. This is probably our biggest bottleneck. :'(
Welll Ive read all this before, hell I have seen it all before. ok lets review some details
....
1. Anecdotal stories:
Well Reliance Mutual look to be moving to Linux. And then there is the My own Company (Who promote NT) whose website runs FREEBSD, then How about the fact that Rodenstock Germany (Big Optical Company) work with Linux as a development platform. The US army think NT is poor performer. No major, and for this I mean AOL, UUNET, DEMON ISP will use NT as its backbone in service provision.
2. Linux Community:
The report uses it to sound denegrating but lkets see. For each app, utility and service there is a clear line of responsibility, opportunity and information. Heck we even know who co-ordinates on the Kernel updates, patches and we can redress the distributors and autors or modules for any extra help if we see fit.
3. Promises of SMP
Well excuse me for not running a superserver but most mid level and small businesses, lets face it they are the ones paying for MS licenses, will not need or require that level of service. Those looking for hight end server perfromance are most likley running 64Bit systems (can anyone say Alpha? Unix?)
4. Late release of FUD
most likely.
As a lecturer in Visual Basic and a seasoned tech support professional in NT and windows I will say this. Out of 20 Students on my part time course in VB over half are llooking to install Linux. Of the 14 People who consult me about installing and upgrading their machines, or getting into the "industry" all will want to install Linux by christmas.
Ummmmmm....
This is the Barn, the horse is gone! do you wish to close the doors?
And thats why Firecrackers and kittens don't mix.
There is a little something I learned about marketing in school:
-AP
I'm sure this post is preceded by the usual trolls and flamers, so let's just ignore them.
I just finished reading the article and considered writing a point-by-point analysis. As Rob pointed out there are some good points and some bad points made. It shouldn't take much time to evaluate each claim and write a reasonable response. So should we?
The fact is we are starting to see more and more of these (hey, it's almost Halloween again, isn't it? (: ). Another fact is that some of these claims clearly are FUD. "Security on linux is an all-or-nothing proposition" (paraphrased from the document). Huh? Claims like this are flatly wrong, either intentionally or accidentally (should we expect everyone to know about sudo and groups?).
I believe that what linux truly needs is some sort of a PR "department". No not a fuddish (cool word) marketing department or bloodthirsty advocates for a particular distribution. Just a comprehensive document developed by a variety of people that addresses some of these common concerns and myths. Does such a document exist? Where?
-Scott
I can see that you've never looked at the NT source. You see, NT is actually a _lot_ better than Linux. More stable, more features, better all around. But it has a unique design feature. Main memory is on the disk, and it swaps to RAM. I'm not really sure _why_ this is done, but most of the times when NT crashes it's because of bad sectors on the disk and other shoddy hw manufacturing.
Why, if MS would put main memory in RAM, it'd smoke Linux's ass. It's unlikely to happen though, because MS firmly believes that it's a feature, not a bug.
-- This and all my posts are in the public domain. I am a lawyer. I am not your lawyer, and this is not legal advice.
The Linux SWAP file is limited to 128 MB RAM. In addition, Linux does not support many of the modern operating system features that Windows NT 4.0 has pioneered such as asynchronous I/O...
The 128 MB limit is long gone away, and IIRC ext2 is asynchronous by default.
Linux security is all-or-nothing. Administrators cannot delegate administrative privileges: a user who needs any administrative capability must be made a full administrator, which compromises best security practices. In contrast, Windows NT allows an administrator to delegate privileges at an exceptionally fine-grained level.
Someone please tell them about sudo!
They do however have a valid point on Journaling FS and fine grained kernel locks which are both in development. I won't even start talking about performance/stability and user friendlyness :)
--
Weasel
I'm curious - why aren't any of the big-name Linux companies (Red Hat, I'm looking in your direction) spreading their own version of FUD back at Microsoft?
Think of it like this: Some small company is looking to network their machines. They've heard of NT, they've heard of Red Hat, but they don't really know the difference other than who's made which. Now, anyone can just go to M$'s site and find that document and say "Gee, Linux has got some problems...hmm, I wonder what they'll say is wrong with NT." Now, I just went to Red Hat's site, and I didn't find a damn thing about why Linux is better than NT. So why, faced with 5 "facts" as to what's wrong with Linux, vs. apparently nothing wrong with NT, would some new purchaser go with Linux?
I don't like M$'s lies any more than the rest of you, but I think we need to keep in mind that if you don't counter the FUD with something (and something very public and noticeable), then the public isn't going to "find out exactly how full of it they are."
After repeatedly seeing on Slashdot news about FUD from MS, and a lot of enthusiastic reply from our community, I felt that something has gone slightly wrong with us - I feel that we are not countering FUDs effectively. The followings are my thoughts.
/. readers to capture the minds of the helpless-FUD-targets? From an outsider's point of view we look like nothing but a whining mob.
1. Audience
-----------
How many times have you read an article or a reply on Slashdot and thought, "damn, wouldn't it be nice if it were on CNet / Yahoo / Wired"? While Slashdot has gained some attention of the general public, get real - by posting a reply on Slashdot, we are only talking to ourselves. It does NOT help counter the effects of the FUD and MS knows very well that they have better FUD targets than us.
Spread the words to where it counts. e.g. Friends on your icq lists. Co-workers. Your boss. Anyone savvy enough to know about computers but not enough to distinguish facts and FUDs. Suggested action: get some words on redhat.com, linux.com, and maybe wired and yahoo. These are the sites where the curious or the helpless FUD targets are.
2. Reader-friendliness
----------------------
As an extension of the above point, our counters are too unfriendly to our audience (or helpless-FUD-target-IT-managers). Look at the jargons. "FUD" itself is a good one. And the long, way-too-technical counters.
Can't we summarize our points into some short, precise, easy-to-read sentences just like MS carefully does? Short does not necessarily mean vague, and with an open effort, we CAN out-FUD the FUDs. (Yes! It may be surprising, but the effect of FUD *can* be generated even with true statements!)
3. Professionalism
------------------
How many of us have read the advocacy-HOWTO?
I've see far too much zealotry, foul languages, etc. on our anti-FUD replies. While the intention is adorable, an incorrect implementation may result in counter-effect.
MS's FUD technology is lightyears ahead of ours. Study the article. Count the number of formal sentences that they use. Count the number of euphamisms. Realize the inexistence of any bluntant, rude rebuttal to Linux (yes, euphamisms again). We have a lot to learn.
4. Effort
---------
How powerful are many small, personal voices from
Is it possible for us to start an open advocacy/anti-FUD group to gather ideas, make counter-benchmark benchmarks, and make our voices known? How about making some public poll on "the number of crashes you have in NT in a week" vs. that of Linux, and make it a news on Wired?
Facts that are trivial to us may not be so to the helpless-FUD-targets. We must find some good ways to get our messages onto the other side.
Just my $2%
the other site I put that on (my normal site) doesn't have too much bandwidth (the machine is just fine however) so I mirrored my rant on enlightenment.org
see http://enlightenment.org/rant.html
--
Geoff Harrison (http://mandrake.net)
Senior Software Engineer - VA Linux Labs (http://www.valinux.com)
Geoff "Mandrake" Harrison
Some Random UI Hacker