All the comments on this page are verbatim copies of e-mails sent to me regarding the response. If I have posted your comment and you would like me to remove it, please contact me.


Date: Thu, 07 Oct 1999 06:46:52 GMT

From: "xavalon xavalon" <>


Hi, I read your response to the MS myth pages and I think there are several wrong statements in your responses.

First, MS NT 4.0 comes with a journaling filesystem. In fact, the NTFS filesystem was always a journaling filesystem and is there since 1993 when NT 3.5 came out. That's why it doesn't need to check after a crash - the journal allowed the filesytem to repair even a large disk in seconds (less then 2 seconds for a 4 GB harddisk or so). It is also 64-bit despite running on a 32 bit OS. 64-bit allowes filesizes of many Terabytes. In W2K this journaling is even enhanced and now allowes to attach a history file (based on the journal) per file (however, there is no tool which allowes you to do so; but in MSDN Journal (the magazine) they explain how to do it).

Second, NT 4.0 got a European C2 rating with NIC. True, it still hasn't got a American C2 rating with NIC (it does have one without NIC). It is under current evalution of C2 with NIC - but this process takes a very long time. And European C2 and American C2 don't difference that much. But remeber Linux cannot get a C2 rating since it lacks too many parts of descretionairy control.


Subject: Linux Myths Response!

Date: Wed, 6 Oct 1999 17:47:01 -0400 (EDT)

From: "Jonathan A. Graham"


I'm not sure if you're the one who maintains the "response to Linux Myths" page. I just thought you might want to expand on the point about:

"The cost of the operating system is only a small percentage of the overall total cost of ownership (TCO). In general Windows NT has proven to have a lower cost of ownership than UNIX. Previous studies have shown that Windows NT has 37 percent lower TCO than UNIX. (This is a comparison to the TCO of a Solaris based system, not Linux.) There is no reason to believe that Linux is significantly different than other versions of UNIX when it comes to TCO."

Although you are completely correct that the study is about Solaris and not UNIX. The argument MS puts forth is that "Linux is no different from Solaris in terms TCO". So this is the issue we must address.

If we examine the article that MS quotes it's easy to see how flimsy MS's argument is.

MS is comparing NT running on x86 hardware vs. Sun on SPARC hardware. If you look at table 1 of the report you see that there is a 54% difference (The single most expensive part of the TCO comparison) between the "Initial systems purchase price" of NT and Solaris/SPARC. At the very least, Linux should get the same "Initial systems purchase price as NT" This, ceteris paribus shortens NT's lead to a mere 23% per user. This does *NOT* include the vast difference in pricing between NT Server and Linux. Therefore there's no corrilation between Linux and Solaris in terms of HW costs. is invalid.

In additon the report states the following:

"Value-added software costs have historically been tied to systems costs; customers typically pay less for software running on less-expensive hardware."

Since NT and Linux run on the same (read: Cheap!) hardware it would seem that any comparison of Linux to Solaris in terms of Software costs would also be invalid.

In short MS comparing TCO of Linux/x86 to Solaris/SPARC is just plain silly.

Subject: Linux Myths


Date: Wed, 6 Oct 1999 20:39:25 +0100

From: "John Everitt" <>


Hi Will and apologies to those BCC'ed.


Up untill last week I was working for a American Company (Arca Systems Inc) that does a lot of eval work and training for various US Security agencies. These are my comments. They are not aligned with that company in any way and are personal. I am aware that some of these statements are sweeping but I beleive them to be ultimately true and having a contact at ZDNet I have to comment because the Myths page really stinks.

You can definitely quote *me* on this one:

Linux is an evolving 'UNIX style' operating system that contains features and bleeding edge compatibility that cannot be beaten by a commercial operating system. Given its age we can all expect foibles and performance issues in some configurations. This will all change. If you look at the development time of Windows NT and juxtapose this with Linux's development time you gain a radically different perspective that is not portrayed on the Linux myths page. Windows NT is older than Linux, it has more features and more problems. Linux is newer than NT, both in concept and development time, but already is a viable competitor. If Linux continues cruising at the same speed will overtake. Here is a few soundbites that do or will apply to Linux:

(1). The XFS filesystem ( will make Linux a much more serious contender. I am really looking forward to this. I like the NTFS filesystem and I want similar functionality i.e. streams etc.

(2). Los Alamos nuclear labs have a Beowulf cluster that has very high performance. This is not yet possible on Windows NT, effective clustering that is. If its good enough for nuclear labs you can be pretty sure it's good enough for business.

(3). The kernel is only in it's second revision. Windows NT is in it's ninth (4 + 5 service pack revisions) and still has VMS legacy source code, Linux did not begin with legacy code and has more room for manoeuvre. Also and particularly from a security standpoint because it can be stripped down and is ripe for CC Evaluation with the right team behind it. Right up to the mid EAL levels. I have worked as part of an evaluation team and can certainly say this is the case. If evaluated it will be taken up by a lot of big organisations like US DOD and US DOE.

(4). MERCED/IA-64/AMD's new processors. Linux can scale very effectively to 64bit platforms because of its Alpha experience. The CPU market is dividing into two camps and Linux can play both.

Some of what Microsoft say is true. However they will be overtaken and this has only been commissioned because someone is scared at Redmond that the monopoly will be ripped from under their feet. If anything this is a sign that Linux is finally big enough to be taken seriously in the corporate market. Microsoft are running scared if they want to take risks like this in the middle of a DOJ trial. Could Microsoft be using its might to crush a GNU/GPL venture? Not very effectively.

XFS provides the following:

Sub-second filesystem recovery after crashes or power failures, 64-bit scalability: millions of terabytes, millions of files, and a million files per directory, High reliability and performance from journaling and other advanced algorithms.

This is out soon. I have a feeling that Microsoft is worried.

Here's a quote: "Configuring Linux security requires an administrator to be an expert in the intricacies of the operating system and how components interact. Misconfigure any part of the operating system and the system could be vulnerable to attack. Windows NT security is easy to set up and administer with tools such as the Security Configuration Editor."

This suggests Microsoft don't feel that you actually need to get an expert to configure your server. I wouldn't go along with this line. You need people like David Litchfield who really understand NT to get a potentially secure NT system. If someone follows the Redhat or SuSE manuals they can have a cheap, fast, functional set up with less risk than a of the shelf NT installation. It is asinine to suggest that a operating system shouldn't be configured by an expert. Is Microsoft predicting the end of the Systems Engineer, Computer Expert and knowledgeable highly skilled IT staff? A scary thought.

And have you ever played with the NT registry? Ever seen Chain Reaction with Keanu Reeves? provides a more realistic perspective than the Microsoft page.


Subject: MS customer testimonials

Date: Wed, 6 Oct 1999 22:07:36 +0300

From: Mikko Hyvarinen



A friend of mine read the MS Linux Myths page & decided to check the apparent validity of their customer testimonials. It turns out that the MX host of is running sendmail, not Exchange ;)

telnet 25 gives:

220 ESMTP Sendmail 8.9.2/8.9.0-M2; Wed, 6 Oct 1999 11:43:09 -0700 (PDT)



Subject: Linux Myths Explained

Date: Wed, 06 Oct 1999 15:16:11 +0000

From: "Karsten M. Self" <>

Organization: Self Analysis

To: <>

General comments, more later, as my copious free time permits. First, good job. Fact fights FUD. I'd suggest a slightly different format for the article. The HTML blockquote tag with a cite style produces a very good form for quoting and commenting on an original work:

This is what we are commenting on

This is our comment on the quoted text. ...and is much less visually complicated than trying to track parenthetical comments. Please *do* check your spelling (no errors noticed) and grammar ("it's" and "its" noted). You're putting this in front of the world. Citations or links for specific counterarguments would be helpful. Throwaway lines "SSL is about security, not speed" make weak arguments. Try to find out what the underlying issues are. If I can serve up secure pages quickly *and* securely, I think I'd prefer it. Good piece. Keep polishing it.

Subject: Linux myth page - possible addition

Date: Wed, 6 Oct 1999 10:19:41 -0500 (CDT)

From: Ken Fowler <>

To: <>

Near the bottom of the page, there is a link to an article about the US Navy's experience with the NT-controlled "smart ship". Good reference. How about equal time for the US Army replacing NT web servers with Macs?


Ken -- Ken Fowler, wa5bru | Real Klingons

klfowler at | don't use .signature files.

Subject: Linuxmyths: TCO

Date: Wed, 6 Oct 1999 18:09:42 +0200

From: "J.H.M. Dassen (Ray)" <>

To: Martin Brooks <>

Hi Martin, I just browsed through . It looks like a very effective reply, but I think a more effective response to the TCO argument is possible, namely: ownership of Linux and NT is an apples and oranges situation.

What does ownership mean?

To be able to own NT the same way you can own Linux, i.e. by being able to fix it if necessary without outside help, you'd have to have the NT source code (plus drivers and apps' source code), which either isn't possible, or requires a Gates-like fortune. One doesn't own NT, one has a copy. Linux you can own (the source gives you the same power as its creator).


Tevens ben ik van mening dat Nederland overdekt dient te worden.

Subject: New entry for the myths response...

Date: Wed, 6 Oct 1999 11:09:49 -0500

From: "Earl, Frank"


Here's the Myth that MS put forward... "There are no commercially proven clustering technologies to provide High Availability for Linux. The Linux community may point to numerous projects and small companies that are aiming to deliver HA functionality. D. H. Brown recently noted that these offerings remain immature and largely unproven in the demanding business world."

I could have sworn that SGI was showcasing some HA technologies at Linux World Expo. Technologies that allowed systems to scale larger than NT's current limit of something like 2-4 machines. I'll bet SGI would be interested in knowing that they're "small" and their stuff is "immature".

Also worth noting is that the D. H. Brown study MS is referring to is based off of information that is a year to a year and a half old, along with being over six months old. Unlike other software, Linux improves in "Internet" time- information such as the D. H. Brown report becomes very stale in only a couple of months' time.

Frank C. Earl

Subject: RE: Linux Myths from M$

Date: Wed, 6 Oct 1999 18:32:59 +0100

From: Martin Spamer

To: "''"


I've just been reading the your response to the M$ Linux Myths and have a couple of contributions you might be able to use.

The so called "Customer Testimonials" are actually M$ cases studies that appear on the Microsoft site, they are NOT Testimonials.

Another large scale NT failure is the NATWEST (UK Bank) ATM project, check out

Check out for general info on NT Failures

Martin Spamer
Senior Software Engineer
Kingston Vision LTD
Phone +44 (0) 1482 602 892
Fax +44 (0) 01482 602 899

Back to the response.