MicroSoft pwns u

Uncle Entity

Sat Jul 26 11:44:12 -0700 2008

manage

Methinks they could have embraced Apache without kicking down a hundred grand a year to become gold (or whatever) members of the Apache Foundation (or whatever).

OpenBSD has done a variation on this by maintaining their own version of Apache after an issue with upstream acceptance of security patches (or whatever) with the Apache devs.

The real question seems to be why they are giving away a big wad of cash when they could just take without giving back like the Bezerkely TCP/IP stack.

"Back in the day," what day?

MicroSoft pwns u

Eadwacer

Sat Jul 26 11:59:56 -0700 2008

manage

It's a big wad of cash to chumps like me and you. It's rounding error to MS. The purpose is to make us believe that they really, truly have changed, because, you know ... they gave this big wad of cash.

"Lennie, give Mr Apache here $100K out of petty cash."

MicroSoft pwns u

Uncle Entity

Sat Jul 26 12:21:21 -0700 2008

manage

Good point.

I wonder how many developer hours that will buy them?

MicroSoft pwns u

GuyFawkes

Sat Jul 26 14:16:26 -0700 2008

manage

MS can almost certainly donate 100k and make a profit on the deal by getting 150k worth of tax breaks for it...

MicroSoft pwns u

Thomas Lord

Sat Jul 26 20:34:30 -0700 2008

manage

100k and make a profit on the deal by getting 150k

Doubtful. They'll probably save about 35K on taxes, net expenditure of 65K. They probably spent over 100K just deciding to spend 100K, too.

Their cost on that little transaction is probably between 150K and 250K (not counting senior exec / board member minutes which would raise the amount a lot). And their direct financial return is, yeah, probably about 35K on taxes.

Personally, I doubt they much did this for propaganda reasons, per se. I don't suspect they give a crap what Bruce et al. say. I think they did it for a very simple reason: the right reason.

I think they did it so that they could be sure there'd be someone in or around the core of the Apache project who would be happy to return a phone call in an hour of need.

-t

MicroSoft pwns u

Aaron Farr

Mon Jul 28 08:07:15 -0700 2008

manage

None.

The sponsorship money only gets Microsoft three things:

• A logo on the Apache "Thank You" web page (points if you can find it on the apache site).
• A joint press release
• Public thank you's from the Foundation at events

In other words, not much.

And more importantly, the Apache sponsorship provides _no_ special access to the code.  But hey, it's not like the code isn't already available under a liberal open source license.  If they just wanted developer hours on the project, they should have used the money to directly hire someone.

MicroSoft pwns u

Aaron Farr

Mon Jul 28 08:04:21 -0700 2008

manage

Exactly.

And besides, Microsoft already bundles a web server with each version of Windows.  You know, their own.

The article author clearly didn't think this conspiracy theory out very well.

MicroSoft pwns u

tert

Sat Jul 26 12:49:05 -0700 2008

manage

I don't get where the complaint or problem is in this post.

The Berkley TCP/IP stack still works just as well now as it did then. Nothing was stolen from BSD. If anything it works much better now due to the improvements made on it thanks to the much wider audience, and the fact that it becomes much more useful as the audience grows.

I'm glad my linux and my bsd box can just plug in to the same router thats sold for windows users and have everything Just Work. Who would be better off if we were still using Microsoft's protocols? Do you actually prefer areas where we're stuck playing follow the moving target (samba, circa win2000/xp release)?

I feel the same way about Apache. If they want to use Apache and the end user gets a better experience as a result, all the better. If anything Microsoft's support might actually be the push Apache needs to stay relevant with Lighttpd taking its spotlight lately.

MicroSoft pwns u

GuyFawkes

Sat Jul 26 13:58:37 -0700 2008

manage

I don't get where the complaint or problem is in this post.

No problem, and no complaint, just observation.

As you say "it works much better now due to the improvements made on it thanks to a much wider audience", but you leave off saying "and you can thank Microsoft for that much wider audience"

So the improvements that you like are just passengers on the coat tails of Microsoft achieving world domainace.

So now you are hoping MS will push apache to stay relevant, and it will, but again it will be on the coat tails of MS achieving dominance, but this time not by integrating a browser into the OS, but by integrating a server into the OS too.

Some perspective

mdb31

Sat Jul 26 12:57:37 -0700 2008

manage

Back in the day, Microsoft had the truly horrible Netbeui and Netbios, NBF proprietary protocol lock in

Ooh, what a wonderfully revisionist history. First of all, Microsoft wasn't responsible for developing any of that: IBM was. Also, NetBIOS was as close to an industry standard as you had at the time: it's basically an API for IEEE 802.2. Microsoft was entirely responsible for inflicting NBF on a sizable customer base in Windows for Workgroups and the earliest releases of Windows 95.

Given the state of TCP/IP at the time (which still had serious competition from the properly committee-developed ISO protocols at the time in large parts of the world), that could be considered a pretty practical decision.

That didn't work very well so they had a go at reverse engineering IPX (Novell)

Again, you're making this sound more evil/incompetent/clueless than it really was. IPX was the dominant networking protocol in corporate networks at the time, and its developer, Novell was unwilling and/or unable to deliver a stable Windows-compatible implementation.

Despite actually well-intended efforts on Microsofts behalf to ensure compatibility, the classic IPX DOS real mode drivers never worked well with protected mode Windows, and the (very, very late) Novell Windows drivers were very unstable as well. The main, if not only, reason for MSIPX's existence was to offer corporate customers the ability to run Windows, sort of the MSFT flagship product at the time, on their networks.

So they (MS) bought Spider Systems licenced product which was based on the Berkeley stack

No, no, no, it wasn't: it was in fact based on an industry standard at the time, STREAMS. Microsoft did port the user-mode API (sockets interface) from BSD, as well as some utilities, such as FTP. To this day, the Slashdot crowd still enjoys running 'strings' on FTP.EXE and claiming that the Windows TCP/IP stack is BSD-based, but this has been debunked so may times (including quite authoritavely at www.kuro5hin.org) that it isn't even funny anymore.

Ladies & Gentlemen I give you Web 2.0, the new and improved thin client cum cloud computing model where all you need to do anything is a browser and a fat pipe.

Yeah, which totally isn't anybody elses strategy -- every heard of Google?

Other than uninformed "M$" bashing, what exactly was the point of your article?

Some perspective

GuyFawkes

Sat Jul 26 14:15:33 -0700 2008

manage

Ooh, what a wonderfully revisionist history.

Not really, but allow me some artistic licence in the interests of setting the scene.

Again, you're making this sound more evil/incompetent/clueless than it really was.

On the contrary, I am painting is a cold, hard, ruthless business at work, with great efficiency.

No, no, no, it wasn't: it was in fact based on an industry standard at the time

Like I said, artistic licence, to emphasis the important point, which kuroshin also makes, that to this day the MS TCP/IP stack is what it is because the Berkeley licence was not GPL.

Yeah, which totally isn't anybody elses strategy -- every heard of Google?

Yes I have, what is their desktop share compared to say, oh, how about MicroSoft?

So no point even asking about integrating a browser into that desktop OS, and then a server, is there?

So if MS controls the worlds desktop OS and further still said desktop OS has an integrated browser and an integrated server, tell me again how significant Google is, and think about it before you answer.

Other than uninformed "M$" bashing, what exactly was the point of your article?

Ohh, uninformed eh, I like that, I remember going around to my mates with a CD with the new Microsoft IE v1.0 browser on it, and he stuck it into his 486 to upgrade "Chicago"....

Completely trashed his system, oh how we (I) laughed, how ridiculous to think that this IE junk could ever compete with Netscape, and we went on to Altavista and Webcrawler and found universal agreement, IE was junk and would never amount to anything.

Some perspective

Nobody Real

Sun Jul 27 00:37:58 -0700 2008

manage

Ok, so let me get this straight.  You use "artistic license", otherwise known as "BS" to make your point, then you expect people to take you seriously?

Your facts are wrong, your history is wrong, and the motives you attribute to things are wrong, and you base your conclusion on these wrong things, is it any wonder your conclusions are wrong too?

Here's another piece of information that blows your theory out of the water.  The BSD code Microsoft used in the first 2 NT releases was *NOT* open source BSDL code.  It was proprietary licensed BSD code.

Spider Software *paid* UCB for the right to use, and resell their network stack.  This is easily verified by looking at the BSD copyrights embedded in the code, they predate the first open source release of the BSD Unix network code (Networking Release 1).  Prior to NR1, the Open Source release of BSD did not include a network stack.  What this means is the only way they could have got that code (legally) was by licensing it from UCB.  Another company that licensed this code was a small-time startup company called Sun.

So no, Microsoft did not use any open source code in their network stack.  They used code paid for and licensed by the University of California at Berkeley.

And the other guy is right, the only reason MS created an IPX stack was because MS and Novell were in the middle of a knock-down drag out fight in which Ray Noorda was deliberately using their near network monopoly to attempt to force MS to not release a server operating system.  Because NT Server competed with Netware, they refused to support IPX on Windows NT, so Microsoft was forced to create their own IPX stack to allow their customers to interoperate with Netware (note: The IPX stack was only meant for client use, not server use.  This alone is proof your point is wrong).

Some perspective

GuyFawkes

Sun Jul 27 03:24:57 -0700 2008

manage

Your facts are wrong, your history is wrong, and the motives you attribute to things are wrong, and you base your conclusion on these wrong things, is it any wonder your conclusions are wrong too?

You see, you're not listening, or reading, you already have an opinion and that opinion is that I don't know what the fuck I'm talking about.

I can list all the facts from Lenny Klienrock and Lickliders "galactic network" in 1961 through CCITT all the way to the present day, but then you have completely buried the fundamental point.

The fundamental point that you and missing because you are so busy being intellectually correct and superior pointing out the things that I have omitted, because they have nothing to do with the fundamental point, is right there in the story.

Internet Explorer v1.0 released in august 1995, and IE was as big a thing as "Chicago", and the non GPL Berkeley stack made it possible, novell networking didn't, netbeui didn't, and everything else is an irellevant distraction, which is why I invoked artistic licence and left it out.

Type "C:\" into IE and do the same in Firefox.

In MSIE it is ___***UTTERLY***___ different from Firefox, just as in my whimsical MSIF it will be UTTERLY different, and just as IE changed forever the nature of the WWW so MSIF will utterly change forever the nature of Web 2.0.

Microsoft set the standards for the content we have today on the WWW, not NCSA, and MSIF will do the same for the future.

But hey, know yourself out, tell me I am a clueless luser who really has no idea about US DoD & Gov in 1988 mandating that OSI was where it was at and TCP/IP would be phased out due to lack of compatible hardware platforms

I've never heard of Dave Clark - we reject kings presidents and voting, we believe in rough consensus and running code.

ISODE, TP4, CLNP, GOSIP, nope, never heard of them.

Dec 7th 1995, Internet Strategy Day, nope, never heard of it.

Some perspective

Nobody Real

Sun Jul 27 12:16:21 -0700 2008

manage

Great.  So what is your point?  Fabricating history then drawing conclusions based on your fabrications is all it appears to be.

Maybe you've been around since the dawn of eternity, that doesn't change the fact that both your "history" and your "conclusions" are wrong.

By the way, typing C:\ into IE7 (or IE8) doesn't do what you think it does.  Microsoft removed that integration.

MicroSoft pwns u

Kai Howells

Sat Jul 26 16:16:35 -0700 2008

manage

 If this is a GPL vs BSD licensing argument, I'm all for the BSD license.

Without the BSD license, we couldn't have Mac OS X in it's current form. In a recent press release, Ballmer specifically named Apple and Google as their two main threats. Bring it on.

BSD is for people who love Unix. Linux is for people who hate Microsoft...

MicroSoft pwns u

Michael Smith

Sun Jul 27 01:47:57 -0700 2008

manage

Without the BSD license, we couldn't have Mac OS X in it's current form.

Oh I don't know. I think Apple could put their own shell on top of Linux just as easily. Offhand I can't see that they have a reason to put their own stuff in the kernel and not GPL it.

They took NextStep to pay Steve to come back and run the company.

bruce

Thomas Lord

Sat Jul 26 20:22:05 -0700 2008

manage

This is one reason why Bruce's recent comparison of MSFT to Nazis bugs me. Our "Chamberlin" moment occurred with the solemnification of the Open Source Definition. The rest is consequence. The OSD tried to fudge away the centrality of the "copyleft" concept and so, IN NO WAY "says the same thing" at the free software movement.

-t

bruce

Uncle Entity

Sat Jul 26 21:12:12 -0700 2008

manage

Nazis?

I thought he meant that the FOSS community was a bunch of pinko commies that were pandering to the capitalist swine like Stalin was doing in that picture.

Oh, my bad...

bruce

Thomas Lord

Sat Jul 26 22:04:06 -0700 2008

manage

He sort of muddled there. Yeah, so, maybe MSFT is Stalin not Hitler. Somebody is supposedly Chamberlin in there. Pot / Kettle / Black is all I'm sayin'.

-t

MicroSoft pwns u

Anthony Steele

Sun Jul 27 05:53:30 -0700 2008

manage

So Microsoft 7 ships with what used to be once the Apache web server for Web 2.0

Where the analogy breaks down is that Microsoft already has ISS (http://en.wikipedia.org/wiki/Internet_Information_Services), which isn't that bad. IIS has active development momentum, with new better versions coming out every year or two (IIS7 is the new hotness). Suggesting that it's suddenly going to go away in favour of Apache is just silly.

Why would they just use Apache instead of IIS, which has had a lot of work put into it to serve asp.net web applications? While people have thier favourtites in the ISS or apache debate, neither of them is out of the running.

MicroSoft pwns u

GuyFawkes

Sun Jul 27 07:02:13 -0700 2008

manage

Why would they just use Apache instead of IIS, which has had a lot of work put into it to serve asp.net web applications?

Cos IIS is resource hungry and likes a machine to itself, but you can pop LAMP on a winders box with really minimal footprint and minimal resource usage.

Plus, I wouldn't fancy running IIS on a MIPS box.

MicroSoft pwns u

eydelber

Sun Jul 27 09:00:38 -0700 2008

manage

Another option is that they get rid of IIS for Apache, throw in mod_mono and support Mono on all platforms (Miguel works at Novell, funny enough, there's Novell again). Then they can win back mindshare of developers, have a leightweight/extensible stack, fully open source framework/ECMA specification, and provide a migration path for existing developers (see IronPython, IronRuby, Phalanger, IKVM, etc.)....

~ Kevin

MicroSoft pwns u

Anthony Steele

Sun Jul 27 09:34:40 -0700 2008

manage

Another option is that they get rid of IIS

That certainly is an option for Microsoft. So is MS giving up on the whole "software" thing and making dishwashers instead. It seems about as likely to me.

MicroSoft pwns u

eydelber

Sun Jul 27 09:56:16 -0700 2008

manage

I agree it's not likely _now_, but it'll be more and more likely as they keep hemorrhaging developers to the open source/Amazon EC2 stack. Their only option is .NET, and if they go "all out" to dual support Apache+mod_mono and make Mono production quality, they can get developers back.

So, in general, I do agree with you that they'd never get rid of IIS (backwards compat, corporate customers with "mission critical" ISAPI extensions, blah blah), but I think it will be more and more likely that they'll dual-support Mono on top of Apache and on top of "open" stacks, since .NET itself is open.

MicroSoft pwns u

Anthony Steele

Sun Jul 27 09:32:34 -0700 2008

manage

I wouldn't know if there's any merit to that statement about resources, I haven't compared the two, but either will work for light loads, and for heavy loads all kinds of factors like caching, database responsiveness, application design and so on come into play.

However with regards to your concerns, windows server 2008 is all about the virtualisation - running multiple VMs under the hypervisor with less overhead, so IIS can have a (virtual) machine to itself. IIS 7 is all about the modularity, so if all you want to do is serve static pages, or run PHP and mySQL on IIS, you don't have to have the overhead of the other parts.

It will remain to be seen how successful this strategy is, but the point is that it is a strategy for advancing Microsoft's interests, it is the strategy that Microsoft is using, and Apache is nowhere to be found in it. 

MicroSoft pwns u

Thomas Lord

Sun Jul 27 11:00:10 -0700 2008

manage

Incidentally, while I agree that MSFT will continue trying to own the cloud (they started years ago) I don't think that they're well positioned to get far.

Reasons: (1) Enterprise customers at big firms increasingly demand browser-independence and standard-conforming web services. That trend will increase as Firefox adoption grows and as the W3C standards that are in the pipeline make it into the field. (2) Enterprise customers at big firms are actually increasingly skiddish about new forms of "lock-in". (3) Google and Amazon have a big head start in owning and operating clouds and in defining the platform they'll run. MSFT is too late to the game. (4) Good database technology *and innovation* is essential to the emerging "web 2.0" platform - most likely database technology currently owned by Mark Logic, or Oracle, or IBM or maybe one or two others. MSFT doesn't own any of that and would have a hard time doing the "extend" part of embrace and extend even if they started with an open source (non-copyleft) winner like Oracle's Berkeley DB XML. (5) Google is already well ahead in working out what goes on a "slender" (not quite "thin") client. Google has two serious advantages in that game: too much in-house talent looking for stuff to do PLUS, inevitably, Google will release all of their clients under GPL or something BSD-ish. (GPL is better for them but they might fear to tread there).

What will probably happen is that MSFT's plays in this space will prolong the life of some of their existing base and probably also find some odd niches, here and there, but it's not a growth area for them. (Aside: I'm not sure they have any big growth areas left in them.)

Remember that they first started trying to "own web 2.0" years ago, before the term "web 2.0" had been coined. MSFT's original plan was to leverage their strong position on the desktop by building "web features" into applications. That way, the thinking went, browsers and standards don't matter so much -- to really get the benefit of the web, you'd better be using MSFT apps. That was a total failure for the simple reason that the "mass consumer"'s platform drives the economy of web technology. At first that meant browser + ActiveX + flash. Now it mostly means browser + flash. In a few years it will mean just a browser, albeit probably with support for local storage, etc.

Netscape basically killed MSFT when they stuck Javascript in the browser. It was just barely enough, at first, to beat back things like ActiveX and thus gave the W3C standards time to grow. In that sense, standardization of Javascript (ECMAScript) was the nail in the coffin.

-t

MicroSoft pwns u

Bruce Perens

Sun Jul 27 13:21:17 -0700 2008

manage

I'm going to continue my usual policy of using GPL (and more recently GPL3 and AGPL3).