Slashdot Mirror
Is Linux or Windows Easier To Install?
Mark Cappel writes: "Joe Barr, a LinuxWorld.com columnist, compares Linux and Windows installations. He expected Windows to be faster and easier since Microsoft has been at it for 21 years. (DOS 1.0 was released 21 years ago today.) It turns out Red Hat is quicker and less manually intensive."
then shouldn't the article be comparing RedHat and Windows installs?
If he is only testing with Redhat, it seems unfair to lump all of those installs as "faster" than Windows, based on the performance of only one type.
Just something to think about.
"PC Load Letter? What the $@#% does that mean?!"
Does anyone else think this review would have been more fair if he had used a retail win2k pro disc instead of using the Sony system restore cd's?
How's my typing? Call 1-800-eta-shut
This test has one serious fault in it that I can see. The tester didn't use a stock W2K disk on a clean system, he used a Sony restore disk, which is a lot different than what Windows 2000 would normally be installed as. A lot of that time installing Windows could be attributed to the restore disk installing all of the myriad programs that come with new computers
Sure, I truly believe that Linux can come out on top with new installs. But do we really need to bias test results in our favor, and then expect corporate users to take us seriously? If Linux users want to show the superiority of the OS, they need to present fair, unbiased tests that are indicative of real-life situations, instead of twisting tests around in subtle ways.
Want Slashdot headlines on your site? Try SlashHead
I'm a Lunix-loving looser (tm Trolls, Inc.), but I think this could in no way be called a proper test.
A better test:
2 identical stock computers, fairly recent but not top-of-the-line hardware
1 copy of Windows XP
1 copy of the latest version of Red Hat, Mandrake, or other selected distribution
2 clueless users
2 clueful users
1 administrator to wipe the machines after each test
1 instruction manual per OS
No gurus
4 runs - one with the cluebies doing Windows, one with Linux, and one of each with the clued-in pair.
Neither user can help the other; both are isolated
We know Joe's a Linux advocate. Let's have a real test.
Someday, you're going to die. Get over it.
I'm sure the Slashdot groupies would get a laugh if Microsoft compared Windows XP to Redhat 6.0.
This is like comparing a 1.5 ghz Athlon and a 1.5 GHZ Pentium 4. You don't... There's no point. Stop comparing apples and oranges people.
Flamebait, maybe. So what, reply. Prove me wrong.
/^[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,4}$/i
I dislike MS as much as anyone else, but come on! This installation competition thing is like comparing apples to BMWs...
First off, a Linux newbie would have absolutely NO clue about half of the stuff Mr. Barr did for the Red Hat installation. Clearly Mr. Barr is a seasoned Linux guy and can breeze through partitioning, network configuration, boot manager selection, package selection, etc. Try any of that on a Linux newbie ("...What's DHCP? And what the hell is this GRUB thing it's asking me about? I'm calling tech support...").
I agree that the Windows installation is slow, has too many reboots, and is not fool-proof as far as hardware detection goes. However, the installation of all Windows products except for the so called "enterprise" editions is set up for people who don't know all that much about hardware. The old 80-20 rule kicks in here: if 80% of the folks are covered by the installation, that can justify the remaining 20% who need hand holding. I still have not encountered a Linux installation that does not assume prior knowledge of technical acronyms, Linux-isms, and common package names (how many new Linux users do you think have any clue that Samba offers Windows network connectivity? How many Linux installations present Samba as a "Windows networking" option and not as "Samba"? None that I know of, that's how many).
As a pro-Linux, pro-BSD, pro-open-source guy, I'm giving this comparison two thumbs down. Sorry, Joe...
Then it was Red Hat's turn. I inserted the first installation CD and rebooted Windows. I chose to manually partition the disk using fdisk. First, I deleted the partition I had originally created for Linux. Then I created a 256-megabyte swap partition and gave the rest of the drive to Red Hat, choosing the ext3 journaling filesystem.
Ok ok, let's stop right here at the first paragraph. So, he already had his drive partitioned from a previous install (meaning he didn't have to mess with fips, partition magic, etc.) and he used fdisk to partition. And exactly how is this easier than a Windows install?
Granted, I've used Linux for years, and fdisk isn't difficult for me to use, but having to use fdisk raises the difficulty of an install considerably. I know that RedHat doesn't require the use of fdisk in their install, but this reviewer should have known better.
I make it a point to try out the various latest Linux installations on a spare machine here just to see how far they've come, and when one compares Redhat to something like SuSE or Mandrake, it still lags behind. RedHat is competing in the Windows NT/2k/XP Workstation/Server market, and isn't apparently too interested in the home desktop market, and their installer reflects this. There are still many questions asked throughout a Redhat install that would require some sort of background in Linux to answer.
Something like SuSE's install would work better for such a comparison, as it best combines ease of use with configurability. The SuSE install tries to autodetect and autoconfigure everything the best it can, and then presents you with a summary of everything it has done, along with the option to change anything if you want to. The new Linux user would probably just click the "Next" and accept these defaults, while the experienced Linux user still has the option to change anything he wants.