Yes, it seems everybody has a performance test. Each one is done by people who wish to have certain performance parameters highlighted. Though I don’t doubt the honesty of the people doing the tests, it is easy to show results that prove what is needing to be proved. Given enough time, almost anything can be proven through some convoluted means.

It all depends upon the bias you bring to the test. Those who want to show that something is better and faster can show the tests where that product does better in them. Tests where the new, or much vaunted, product does worse than that compared to are simply left out. It’s a long standing way of doing things (If you doubt this, look at the tests that Ed Bott, Microsoft reporter and writer of pro-Microsoft books, and columns, uses. The tests will always show that the highlighted item is better; anything that messes with the perfection of the product highlighted is simply left out.)

Friday, Betanews released tests that show Windows XP SP3 running browsers faster than the RTM code of Windows 7. It does not matter that the margin is only 13%, the point is that there is proof of something that many people (including me) have been saying all along. Windows XP is faster, in at least some situations, than Windows 7. (The only place where I have seen Windows 7 be faster, consistently, is in the setup, and it has been explained why this is so. It would be trivial for Microsoft to change the setup method for XP – but we know that will never happen, as it’s a dead product, from their perspective.)

Let’s think about this for a moment, logically. Windows XP can be used on machines that have as little as 256 MB of main memory. Microsoft says it is so, so this is not simply a case where someone decided to remove memory after an installation, to see if the system would continue to run (this is the method that Ed Bott, on ZDNet used to accomplish the feat with Windows 7, and we only have his word on the satisfactory results. What we don’t have is the proof of Windows 7 proceeding from blank hard drive to complete install with only 256 MB of memory.).  On the other hand, the minimum needed for Windows 7 is listed as 1 GB  of main memory, so the question that I come up with is, “Why isn’t Windows 7 always much faster, considering the ‘advances’ that Microsoft states?”

I wonder how much difference can be attributed to the simple change in the video requirements. The necessary memory of 128 MB as a video buffer, and the DX9 requirement, might make the difference in many of the tests we have seen. (Yet the Windows XP systems, not fully optimized for the graphics hardware, manage to be slightly faster, if not much faster.)

Most of the time I read an article, the bias of the writer, or the publication, or site, is immediately noticed. I can’t, however, really understand what the point of this article is, given the pro-Windows 7 bias of the Betanews site. It almost seems as though it could be an attempt at impartiality, but that ship has already sailed.

It’s simply too bad we could not see a performance test with the graphics routines of Windows 7 slapped into Windows XP, and then run on equal hardware… I think some interesting results would come out.



makes you wonder about your fellow man, eh?


Digg This