Slashdot Mirror
Report From "Get The Facts"
Richard W.M. Jones writes "Huw Lynes wrote an interesting
report from Microsoft's
"Get The Facts" show in London
(earlier
Slashdot story).
Along with the report he provides some
analysis of their apparent strategy, which
includes equating "Shared Source" with "Open Source"
and making out that Linux isn't free."
Microsoft Starts its "Get The Facts" Campaign
So I sat with about 150 other "technical decision makers" in a very plush hotel in Holborn while representatives from Microsoft tried their best to convince me that I should not be considering moving to Linux. To run the discussion Microsoft had employed a fake-tan horror who had clearly escaped from daytime TV. He was by turns chummy and condescending. However being a reasonable man I will not hold Microsoft responsible for his failings.
First up was Phillip Dawson who leads Linux research for analysts Meta Group. He quoted heavily from a Meta analysis which shows that Total Cost of Ownership (TCO) for linux and windows is comparable. This study has been widely reported in IT press but I can't for the life of me find a link to the original. He made some interesting points about where the datacentre is going to be in a few years. His basic thrust was that everyone is moving from proprietary Unix with its expensive platforms to Windows or Linux on x86 platforms and that it this hardware move, rather than linux versus windows, that will drive all the cost savings. Dawson believes that in a few years the only place we will see proprietary Unix is in very large enterprise databases.
After a promising start, Dawson then got into the territory of why Windows makes more sense for enterprises than Linux. He introduced what was to become a running theme for seminar, Linux is not free. It turns out that the TCO statements made earlier were based on the licensing costs of SuSE professional and Red Hat Enterprise versus Windows. They had refused to consider that people might run a business on something that they could download free from the Internet. Later in the Q and A session Dawson got quite aggravated when people pointed out to him that many Linux-based businesses run quite happily on free linux (this was shouted by the scruffy-looking Debian hackers in the back). I can only assume that businesses that are brave enough to save thousands of pounds per unit by moving away from expensive hardware platforms are meant not to care that they can save another couple of hundred pounds on Microsoft licence fees. Later in the presentation he said "Don't compare to the free downloads. They are not free". Precisely what he meant by this escapes me.
One area the Meta study didn't look at was Linux on the desktop. Phil claimed that linux was not ready for the desktop because it lacked administrative tools. He was carrying on in a similar vein when he said "Management tools on Linux are nearly as good as a DOS prompt".
Nick Barley, business and Marketing Director for Microsoft UK took to the stage to baffle us with market-speak. There was lots of talk about strategy and leveraging which I didn't follow. He talked a bit about Microsoft's shared-source program and tried his hardest to make it sound like open-source, mainly by refusing to say Open-source and talking about shared-source instead. Continuing in Phillip Dawson's footsteps he repeated the mantra "Linux is not free" several times. Although he was at his best when talking about business models amongst Linux distributors claiming that "Linux is moving to the same model that Microsoft has been using".
My absolute favourite part of the talk was when Barley started to extol the virtues of Windows because everything in it was made by one manufacturer. A fair point which would have been well taken had he not gone on to draw an idiotic analogy. He asked us to imagine an aeroplane where different components were made by different companies. Apparently he's never heard of Airbus.
Next up was Nick McGrath head of platform strategy for Microsoft UK. The main bulk of his talk was taken up by a demonstration of a document sharing system based on Microsoft Sharepoint. Very boring for those of us running heterogeneous systems that Sharepoint will not run on. McGrath was much more technically clued up than Barley, and seemed to be aware that the audience was not entirely on his side. He made me
Has anyone else noticed that in the metagroup TCO analysis, they compare a windows server running on a two processor intel machine, and a linux server running on (one or two - can't remember) MAINFRAME processors.
I mean - cm'on, perhaps they should have pitted a walmart PC with windows installed vs Linux running on a Cray server... The TCO takes into account the entire purchase of hardware, and in the Mainframe case - you probably looking at 16 processor machine to begin with, which kind'a spikes the price up...
But - the graph looks very convincing - and isn't it what it's all about?
Just a little food for thought...
Essentially Coke was the biggest cola company on the block, until they acknowledged Pepsi as a competitor.
You say this as if they aren't still the biggest on the block. Coke is still (as it has always been) well ahead of Pepsi in both global market share and global market value. Their stock price is higher, and they still ship many more units / yaar then Pepsi. Sure Pepsi may have more flashy ads in the US, but that doesn't mean squat to their international presence. Just do a Google on the cola wars.
This said, if Linux ever got to the point that it was as much of a competitor to MS as Pepsi is to Coke, I'd be damn happy.
And they always conveniently forget to mention the cost of upgrading your Microsoft products. My current employer lost a boatload of money when they tried to move from NT to Win2k on the server, because a last-minute backwards incompatibility threw a spanner into the works. The project had to be called off, effectiely wasting several months of effort by about half the engineering group. You do the math on how much that cost the company, nevermind the actual license cost.
They also don't mention that in many cases, a great deal of the cost is inspired by Microsoft's lock-in. Your data in their products isn't open... you have to pry it out. If your data was in open formats (i.e. actual, for-real XML) then you'd be able to migrate a lot easier. So, it's a cost really imposed by Microsoft, rather than a cost imposed by any alternative solution. The erverse probably isn't true... once in an open format, there's usually not an 'exit cost' associated with moving to another solution.
I jumped in to the "Desktop Linux Consortium" back in the Feb 2003 to offer some thoughts about direction for the forming DLC and the linux desktop in general. If you have any interest in what I said back then:
l c-discuss/2003-February/000002.html
http://www.desktoplinuxconsortium.org/pipermail/d
I think that the crucial missing application and management pieces are staring us all right in the face. It is not enough to have an easy install. It's not enough to have a slick desktop and functional apps. Those are important, certainly, but if we are really doing well at them, why hasn't the momentum shifted?
I've worked IT for fifteen years and the number of systems I've imaged with their OS and software loads dwarfs by 100 to 1 the number of times I've used any OS installer, even if you count the last five years of Install Parties at the Melbourne Florida LUG! The things most developers and non-corporate users think are important don't apply to corporate IT like people outside of IT would think.
The typical larger IT department has to deal with things like corporate software policies, locking user account profiles, automated application and operating system patches/updates and remote helpdesk. How can I enforce the corporate software policy against instant messengers when every distro except debian bundles all the stock KDE applications (including instant messenger apps) in a few giant RPMs? KDE 3.2 will be doing more profile locking features, but what about applications that don't use the KDE libs? What about Gnome?
I know people point to things like Red Carpet and the Red Hat Network for updates (still not 100% in my opinion), but I think corporations will need to be able to build or rebuild apps with different attributes or patches for distribution to corporate clients. SUSE is using 'autobuild' internally and Red Hat wants you to buy a Red Hat Network Proxy, but again, no-one other than Debian provides access to the build architecture to be able to modify certain stock bundled apps like removing parts from larger RPM's like KDE.
Remote helpdesk and other IT-friendly features are available in most distributions at this point, but they aren't really bundled and configured for that role in the context of the distribution. This needs work and attention. VNC is great, but a distro focusing on corporate desktops needs to have that puppy configured for easy remote desktop support by default.
I've spoken at LinuxWorld and other conferences, but every time I try to submit a topic that addresses some of these kinds of issues, I hear crickets and we get 10 more 'How to install Samba' sessions. We need a focus on what all the "Ticket System Cowboys" know about desktop deployments before some of the spectacular Linux desktop announcements turn into craptastic failures.
Just my $0.02.
DaGoodBoy
My God! It's full of Voids!
C:\WINDOWS\system32>strings ftp.exe | grep -A 1 Copyright
@(#) Copyright (c) 1983 The Regents of the University of California.
All rights reserved.
# cat
Damn, my RAM is full of llamas.
who doesn't know of the socket handle leak that MS can't fix because otherwise they'd break 1000's of apps
My sympathy levels for Microsoft engineers skyrocketted after reading this and this, detailing the horrors they have to deal with in the name of compatability.