[nycphp-talk] IIS a dysfunctional kludge?
Paul A Houle
paul at devonianfarm.com
Fri Jan 15 10:09:31 EST 2010
Greg Rundlett (freephile) wrote:
> Actually it's 54%, but who's counting ;-)  Still, that's twice the
> nearest competitor and so you're logic (go with the leader) stands.
> A patchy server  still leads the pack.
I think it comes down to platform choice.
People who choose Unix tend to choose Apache; although it's not the
fastest web server, it's mature and it's feature rich in many ways
(you'll take URL rewriting of my cold dead hands.)
If you're running Windows, you get IIS at no extra cost. Today you
can get a good PHP implementation for it, and if you like ASP.NET (it
has its charms) it integrates easily without a lot of trouble. IIS can
certainly give you headaches involving file permissions, but it really
isn't that bad; IIS 7 is moving more and more towards a
configuration-file driven approach that works like Apache.
When I'm paying for servers on my own dime, I use Linux and
Apache; I still think it's easier to admin a Linux machine that's
"hands off" than it is admin Windows. That said, Windows has come a
~long~ way in the last 10 years: I do some sysadmin work on Windows
servers at work and it's not that bad.
Looking back since 1993, when I built my first Linux box, I'm
amazed at how much Windows has improved and at how, in many ways,
Linux has stagnated and gotten worse. Linux finally has a decent
kernel, the 2.6 line, that's reliable and scalable on multi-CPU
machines, but "Linux on the desktop" seems to be going backwards.
For instance, KDE was actually impressive when I first saw it in
'97. They keep adding more features, but it seems to get slower and
less usable. GNOME lost all credibility with me when they went with
"Spatial Nautilus", which makes it entirely unusable for the way I want
to use it. The same way General Motors has a captive audience of police
departments and rental car companies that gives them no incentive to
develop cars consumers want, Linux distro/desktop developers aren't
making any serious effort to make a product that appeals to people who
aren't already "true believers" in Linux already.
More information about the talk