Re: open source
Just because I don't like to leave a good point firmly made, for the benefit of all those Windows-lovin' Microsofty sycophants who love inventing false statistics and spin, if not to convince others then at least, however bravely, to convince themselves that their bullcrap is in any way true and would, if known by Microsoft PR, at once become the basis for the next marketing ploy to get more brainless gorps to buy into their bloatware:
<quote>
"The Internet is *BUILT* on open source. Proprietary systems such as Exchange, IIS
and their ilk are in the (very small) minority."
Err, you forgot Windows.
Given that the vast majority use this proprietary system to access the Internet your
argument is flawed as Windows is in the very large majority.
</quote>
Someone already pointed out that you're wrong about web and mail services, so there's no need to repeat the numbers. Mail, in particular, is almost never serviced by exchange servers connected directly to the net; it usually gets use *inside* the corporate boundary for features besides high-performance SMTP and mail processing (at which it is *particularly* bad, no matter how much hardware you throw at it). And Apache is still Open Source, whether it runs on Windows or not, so talking about Open Source Software deployment in terms of OS choice is meaningless. Apache *is* the leader right now, no doubt about it, although I personally love thttpd which for now only runs on *nix.
And now, for Windows on desktops. Alright then, we'll play your little game, matey-boy. To connect to the Internet in any useful way, you'll be needing implementations of two rather essential protocols in the modern networking age, TCP (RFC 793) and IP (RFC 791), among others atop, near and beneath, also standardised by the consensus process, like ICMP, UDP and SMTP. The TCP/IP protocols in particular are legendarily tricky to implement and have been extended many times in many ways and go back a long enough way to make using a known-to-be-good codebase that implements them well and reliably worthwhile. And guess what, matey-boy? The Berkeley TCP/IP stack, an open source implementation of TCP/IP used originally in UC Berkeley Software Distribution (BSD), still runs to this very day in some (usually modified) way in most operating systems, open and closed, embedded and desktop, experimental and full-fledged. Including, of course, that most omnipotent of all operating systems (if you can call it that ...) most dear to our collective hearts: Microsoft (R) Windows.
Even this very fact was flung out to hideous extent by Microsoft's very own code thief (can't find the refs now, sorry!) as being an unreasonable assertion due to the way the code got into their grubby little hands in a variety of laughable licensing deals before it was finally agreed to put a stop to the fuss made by the righteous by adding the few lines necessary to a deeply-recessed Readme file on every CD-ROM of Windows sold. (You should find it on an XP CD in one of the numerous release notes files, anyway. Don't know if/how the situation changes with Vista, as they've "Rebranded" TCP/IP yet again and are telling us how much better it is than the last one - but all I see is that IPv4 and IPv6 are now in a unified stack and a few more TCP extensions are now on by default.) And it would, of course, be the BSD people who got the ball rolling in the first place, noting how Microsoft could get away with a small but significant license violation for so many years and take the credit for an excellent, high-performance stack developed anywhere but at Microsoft (to this day, only NetBSD performs better at optimised configurations - way to go, guys!). Of course, the code originally got nicked by Spyglass who then licensed it to Microsoft without telling them, but that doesn't excuse Microsoft for their terrible crimes against superlative software engineering; they were hardly open about it and many other evil doings in the name of pseudo-standards conformance (think Kerberos ... think IPSec). Nor, of course, does it change the very terrible fact that OPEN SOURCE SOFTWARE IS USED TO CONNECT ANY WINDOWS MACHINE TO THE INTERNET, AND THAT WITHOUT IT THE MACHINE IS JUST AN EXPENSIVE PAPERWEIGHT. (No, wait a minute - the operating system on the machine is an expensive paperweight - remove it and the machine is useful again, OS or no OS.)
If you want to verify this for yourself, seek c:\windows\system32\ftp.exe, a part of the Berkeley TCP/IP stack that implements an interface for the File Transfer Protocol (another hairy one running on top of TCP), for this string:
"Copyright (c) 1983 The Regents of the University of California.
All rights reserved."
and, as the popular saying went, ye shall find.
Ah, well. Next time, perhaps. Meantime, nearly 100% of all computers are using Open Source if they want to be on the net. That's most consumers taken care of.
Cheers,
Sabahattin
PS: sorry for the heated, patronising nature and length of the post, it is a regrettable necessity. Besides, it's Saturday and I'm bored.