next »

[Cleo]

An interesting benchmark between Ubuntu, Vista and Windows 7 is available on tuxradar.com.

[Nightowl]

Note that, as mentioned in the comments for that article, most of the benchmarks there are fairly irrelevant.

I mean, installation time? That happens, well, once in a blue moon for GNU/Linux. Partially relevant, I suppose, for operating systems such as WinXP etc.

Installation clicks? Who could care less about that? /me has exactly 0 mouse clicks for installation. Keystrokes and mouse clicks together would perhaps be more relevant.

Boot time, sure. That is relevant. But, did anyone note the fact that Ubuntu probably gets uptimes that are several times longer than Windows Vista/7? I know that my computer, which is running Debian Sid (Sid = unstable), regularly gets uptimes of 10 days at a time. (Assuming there's no kernel upgrades available, etc.)

Disk space used. As pointed out in the comments, most of the time, the average home user doesn't really care about this, considering the price of a 250 GB HDD nowadays. The people who do care fall into two categories: sysadmins running servers, and people who can't afford to buy a HDD. If you can't afford a HDD, how are you going to afford Windows? It's been a while since I've seen a computer that has less than a 20 GB HDD that isn't mine. Netbooks excepted.

I/O testing. One wonders if the tester compared speeds of the DOS copy command and the Linux cp command. If those are different, then I'd begin to wonder. The speed of the GUI tools I've run into in the past is usually slower than those of their command-line equivalents.

The Richards benchmark. Fine, great. A true benchmark . . . can't argue with that, right? Windows 7 is faster, according to that benchmark. Perhaps MS has implemented some kind of instruction pre-guessing or similar. Unlikely. More likely is that this is simply Ubuntu bloat. (No offense to Ubuntu, but it dramatically slows my computer compared to Debian. Out-of-the-box GNOME, similar packages installed (extras on Debian to make up for the auto-installed ones in Ubuntu).)

ext3 vs. ext4, while an interesting comparison, has no place in this benchmark . . .

Anyhow, it's a decent comparison, all things considered.

[Lippy]

Yeah, this is definitely something that needs to be taken with a pinch of salt, though this rule still exists with any benchmark really. I personally rate OSes by their overall experience. That is ease of use, how fast it feels, stability, etc.

I've experienced all three benchmarked OSes first-hand. Ubuntu wasn't bad, though it felt rough around the edges compared to Windows. I'm going to be trying Debian in a bit though. One of the problems I had with it is 3D acceleration support is pretty bad, and we are entering into a time where 3D computing is becoming more and more prevalent. It's not just something games use nowadays. I dual-booted Ubuntu with Vista at one point on my now crippled laptop (graphics card decided to give up the ghost last week), and the differences in CPU consumption even when idle wasn't much. Ubuntu used a bit less memory than Vista though and overall felt a little faster, though both caused the laptop to overheat something chronic. Windows 7 however runs a dream on the laptop. I think it outdoes both in terms of performance. The laptop actually felt pretty healthy again for the first time in months. That is until the graphics card bit the dust.

Vista on my main PC is a different story really. It completely flies on there. I've even got the x64 version installed on here and it just works out of the box. I have it dual-booted with Windows 7 x64, though unlike the case of the laptop, I've seen very little improvement in performance over Vista. However, this computer is capable of running Vista like lightning, so it's not really a surprise.

In my personal opinion, I think Windows is the better operating system on good hardware. Yes, even Vista. It is reliable, easy to use, and easily has the best application compatibility, and is the only sensible operating system for serious gaming. The poor driver support in Linux still hurts it the most I think. When it works, it works well, but more often than not I've had some issues with it. Some I can fix manually, others I have to live with. To me, that's a showstopper on a high-end system.

However, this completely changes with netbooks, even when Windows 7 enters the market. There you either have the choice of Home Premium which will probably cost half as much as the hardware itself, or an extremely crippled Starter edition for a much lower price. I have to say, the crippled Starter edition is no match for Linux. The 3 program limit is a huge deal-breaker for me. The other problem is hard drive space, especially as SSDs are much more dominant on these computers. Also, application compatibility matters much less on netbooks. You will likely be using a browser and a few IM applications (or just Pidgin), and perhaps an office suite such as OpenOffice.org if you want to do some work on it. This is easily catered for on Linux, so I think it beats Windows by quite a fair margin in this area of the market. Unfortunately, I have to say the distro installed by default on the Eee PC is an insult to Linux. Xandros didn't even have GCC included, and it had no desktop, just a dashboard instead which is very hard to navigate. It's as if the OEMs are trying to sell the worst distro of Linux possible just so they can give it a bad name and kill it off quickly.