"...Windows loses on the same box..."

Well, it loses because you can't run a Microsoft solution on one box and be supported in 'production environments'.

With a FOSS solution, support isn't a huge issue. Because of that, what support you can get is often more flexible. So if you're throwing together a DEV or UAT environment, it's often an all-in-one box. Webserver, database, app daemons - everything on one box.

And you can deploy live that too, in your Production environment. You only need to break it apart into multiple boxes when you hit any kind of performance/scaling issue.

But the Microsoft Way is that all these things must be separate machines. And that ends up costing a lot of money, taking up a lot more time (because there are more machines to deal with), and generally being cumbersome.

Microsoft doesn't care that other solutions require less machines. They sell licenses. Not software - licenses. So to them, having a solution which requires three or four machines is actually a plus - more Windows licenses!

Case in point - I was recently involved in rolling out Office Communications Server (OCS), which does instant messaging/voice/video chat.
We already have another system which also does IM/chat/voice. It requires two servers, and even then it only requires two because we want high availability - if one server/site dies, the other takes over.

Microsoft's OCS, to do the same job, requires three servers. Multiply by two to get high availability, and then add in an odd little beastie of a server which "monitors quality of service", and we have a total of... Seven servers.
(The monitoring server can't be highly available, you see. No, don't ask. The answer will just make that dull throbbing pain in your skull worse...)

Two servers vs seven servers.

Microsoft's solution is not obviously better than the two server solution. They both have strengths and weaknesses. Overall, for our requirements, they're on a par with each other.
But in terms of cost, Microsoft's solution requires five more Windows Server licences, and some SQL Server licenses.
(I'm still trying to fathom why an IM server needs SQL Server behind it, when the volume could probably be happily handled by ESE, and the workload itself may well be better suited to ESE anyway. Oh, wait, there's that dull throbbing pain again. Better think about something else.)

Now, I can see some good reasons for splitting things into so many servers. But I can't see good reasons for REQUIRING it. Any technical reasons are, IMO, the result of sloppy engineering. But I don't actually see any technical reasons in 99% of cases. The only real reason for needing 7 servers where competitors can do it on 2 is that they want to sell more licenses.

Look at any backoffice solution from Microsoft, and watch the costs spiral as you suddenly and inexplicably require SQL Server and IIS boxes as separate hardware. And boggle further as every possible element somehow becomes a separate server, which then sits there at 5% utilisation for its entire lifetime.

Sure, you can re-use some of the SQL Servers or IIS servers for multiple products to save costs, but in a corporate network with DMZs for different purposes, you're still looking at a lot of servers for very little gain.

Oh, and yes, we considered virtualisation. Not supported. (I'd imagine that will change in the future, because whilst it reduces data centre bills for power and environmental control, virtualisation doesn't take any money from the licensing pool.)

Microsoft's rigid focus on getting as many licenses into the data centre as possible is directly related to why most young hip developers would rather fire up a LAMP stack. The LAMP stack is cheaper, it's more flexible, and it means you can get coding faster.
Or just take time out to party.

Either way, you can see why it's more attractive.

Yep, they're not cheap.

But they offer static IPs, in both IPv4 and IPv6. They do decent monitoring of your line, and have excellent tools for reporting your bandwidth usage, uptime and so forth. They can be your domain registrar and offer DNS servers.

With the exception of the bandwidth limits, they're pretty much a geek's dream ISP - pay for good service, get good service.

Heck, you can ask their support staff questions via IRC. You can get SMS alerts when your ADSL line is down.

I have two friends that use A&A. They're very happy with them. Most other people I know are on cheap ADSL providers, Virgin Cable or work for an ISP so have a connection through their employer anyway. I don't really hear complaints from the A&A users or those getting their connection from their employer. But everyone else, myself included, has had issues and not felt the support was good.

And no, I don't work for A&A. I get my internet connection through Virgin's cable service, because I had cable already. And I'm not being paid to say this - A&A don't know me from the proverbial Adam.

But A&A are the first company I'd look at if I had to switch to ADSL. I want their service, despite the costs.

It's not about piracy, this is about service. Many A&A customers host their own webservers at home, for example... Do you want to lose your internet connection just because of a bogus complaint about a webserver only you and a few friends use?

That's probably why they're putting these protections in, more than anything else. Arse covering for their customers. ;-)

Who says that the problem is the organisations running IE6?

Most of them would move off it to IE7 if they could. Really they would. It's not much more (or less) difficult than any other large application - test it, package it, drop it to machines. That's an established and controlled cycle.

The problem is not that they don't want to. The problem is that the enterprise software segment has been very shortsighted.

SAP, PeopleSoft, Seibel. ERP, Human Resources, CRM systems.

Things that organisations put in which are strategic, in a way which deserves block capitals that I'm not quite willing to put it in. But trust me, these things are signed off at board level and cost a truckload of money, so there's a management investment in getting them working at all costs - otherwise they'd have to admit that they were wrong!

And once in, they become very hard to remove. Stuff like PeopleSoft is often tied to processes like leave booking, expense claims and payroll, for example.

Upgrading these systems is not a trivial task. It's one that, even if it's a simple and smooth process, has huge risks. Risks that run towards lost days of business, inability to produce corporate accounts, or handle staffing changes and expense claims. So these systems are upgraded at a glacial pace, with the process being rigorously controlled and methodically run.

Guess which systems a company buys that would require IE6?

I'm sure that they all have versions out now that support IE7 and higher. But the companies running these systems are often one or two versions behind, and have real incentives to avoid change. Incentives that don't even factor in the client web browser's name and version.

And don't forget that these systems are not cheap. The upgrade software bill will be huge, before you even start any work. Another reason to delay, especially once management remembers how expensive and painful it was tweaking and customising these systems to match their organisation's workflows and requirements. Do that again? At huge cost? Barely two years after we last suffered through that? NO WAY!

And so this is how it starts. You can then add the fact that developers then have a standard platform of just one browser, and you soon find any in-house development is tested on just one browser. Which compounds the problem.

It started with enterprise software. It's continued by in-house developers. It has nothing to do with the merits of the client in any way.

And good luck not dealing with such organisations. This, sadly, is the state of pretty much all large organisations...

I'm not defending them, by the way. I'm just trying to help you understand why large organisations are stuck in this rut.

Off the top of my head, I'd say that the big changes would be:
1. Russia will get boisterous and attempt to take on some of its smaller satellites.
2. China will make a move for Taiwan, and might get more aggressive with Japan/Korea.
3. At some point, someone will attack Israel.
4. Um...
5. Nope, that's about it.

Number 1 has been happening on and off anyway. I just think that with no threat of U.S. intervention, Russia might throw caution to the wind and go a bit nuts on that front.

China taking Taiwan is kind of predictable, too. Japan and Korea might follow.

Israel is obvious. Someone will see a moment of what they think is weakness, and try to walk in and be an Islamic hero. I'd say that they'll have their arse handed to them on a plate, unless they're remarkably successful, in which case their arse will become a glowing cinder. No country in the world has the will to use its Nukes in self defence like Israel has...
But because it's a religious thing, someone will be dumb enough to try it at some point. Sad, but true.
The absence of the U.S. might just make them a little more eager about it though.

But here's the thing that most Americans don't seem to understand...
NOBODY CARES ABOUT THE U.S. ANYMORE. The U.S. HAS BECOME FAR LESS RELEVANT.

If the U.S. dismantled its military, it wouldn't affect much. The only countries it would really affect are Russia, China and Israel. By extension, it will affect the ex-USSR states, Taiwan, Japan, Korea and whoever's dumb enough to attack Israel.

However, I must ask why you think the U.S. should completely dismantle its military. It's just not necessary.

The U.S. could make huge savings just by admitting that the cold war ended years ago, and that state vs state war is going to be skirmish at best - especially if they keep their ICBMs.

The U.S. has carrier fleets that they're afraid to deploy against pirates, because they know that a speedboat loaded with explosives can take out one of their destroyers.

The U.S. has hordes of tanks that take forever to deploy, require huge supply lines, and yet can be taken out from a rooftop with an RPG.

The U.S. has aircraft that are truly fantastic, amazing bits of kit - but that are hugely expensive and not much more effective than their immediate (much cheaper) competition.

The U.S. military-industrial complex is throwing money away fighting a war that ended two decades ago. What's needed now is helicopter carrier fleets - smaller, faster, more agile. More Marines and more transport and support for them.
More unmanned aircraft and ground support aircraft (like the old A-10 and the AC-130).

Also badly needed is strong military field engineering, with a civilian eye. No U.S. field base should leave an area without giving every nearby village better water supplies, a prefabbed school building, and a courtesy lick of paint. Hearts and minds will secure the bases just as well, if not better, than barbed wire and watchtowers.

And the U.S. needs a "Missile Shield" to protect itself like I need a six-barrelled rotary cannon with laser sights to protect me from flies in the summer.

You could cut the U.S. military budget in half, embark on a major restructuring project, and within five years America would have a far more effective military force than it has right now.

Because right now, if the U.S. was attacked, it wouldn't be able to defend itself. That was proved on September 11th, 2001. The leader of the group that made that attack is still not captured. The senior leadership of that group is hardly dented. The group has made huge territorial gains in Pakistan because of the U.S.'s military ability to handle it. And that group is recruiting more people every day.

Dismantling the military isn't necessary. But realising that the USSR has been dead and buried for almost 20 years would be a nice first step to making it cost effective...

I'd second this, and go further.

Patching.

I think that Steam handling my patching for me is great.

If I had to reinstall Quake III from my original media, I'd then have to go out there onto the internet and find the patches.

Or I could buy it on Steam, and download the last version without hassles.

Same goes for moving machines - Steam makes this easy.

As a rule I dislike DRM. But I won't object to it on principle - some DRM can be quite acceptable because it gets the tradeoffs right, and I believe Steam is a good example of that.

Having followed Damian's CouchDB project since its inception, I'd say it's important to note that he didn't know much about TRADITIONAL databases when he started.

By traditional, he'd mean relational.

But he's trying to build a document-oriented database.

A relational database has about as much to do with a document-oriented database as American Football has to do with Football. There's a little common ground (flat green playing surface, two teams, a football) but even in that common ground we find that the definitions hide differences which make them impossibly interchangable. (Try changing a football for a football and then playing football. Doesn't matter which side you approach that sentence from, you're not going to be happy with the results...)

By building a schema that splits "documents" apart and puts them into records in many seperate tables, an RDBMS gets consistency and efficiency across the whole store. It does so at certain costs - for instance the cost of higher maintenance costs of change, as the schema must be updated and current data massaged to fit.

By building without schemas and allowing each document to contain what it likes, a document-oriented database gets consistency and flexibility for all documents. It does so at certain costs - for instance higher costs of storage and querying performance, as identical fields in documents are reproduced rather than referenced, and must be stored/queried each time.

But both approached have their benefits. The RDBMS gets faster querying and better storage efficiencies. The document-oriented database gets document integrity and the possibility of easier versioning and security functions.

Neither approach is globally right. Each has its use cases. If I had to build a database that tracked ticket sales/seat allocations in realtime for a worldwide airline, I'd pick an RDBMS.
But if I wanted to build a discussion forum or a document management system, I'd be more inclined to pick the document-oriented database system.

To say "it's as rational as rolling your own encryption from scratch without the slightest clue about encryption algorithms or theories" merely shows your own ignorance of the document-oriented approach.

You might as well say that it's as rational as trying to build a boat when he has no knowledge of motorbike maintenance...

Still, your title was right. Laziness rules when it comes to people informing themselves before posting on Slashdot... ;-)

Most people don't buy Windows.

They buy computers with Windows on them.

When you buy a Dell, it comes with a load of software on it. That software could include Firefox if Dell wanted to ship that.

(I'm just using Dell as an example, by the way.)

Now, the outcome of the DoJ antitrust case was that Microsoft's control over the desktop that an OEM ships was reduced. Prior to the DoJ case, Microsoft could actually prevent OEMs from putting any icon they disliked onto the desktop. (Really!)

However, the DoJ case had a weak punishment. Whilst Microsoft can't tell Dell not to install Firefox, and can't stop them from putting the icon on the desktop, that's still not good enough for Dell and other OEMs.

Because the DoJ outcome still has IE bundled with Windows. So if Dell did ship Forefox, IE would still be there. That potentially confuses customers, which can increase your support costs, and makes Firefox (or any non-IE browser) more expensive.

Now, if Dell could choose to create an image that didn't have IE installed, it would change things. They could ship another browser, with potentially less support costs and less confusion for users.

(I'm not going to address the issues with web design, except to agree that many sites are crap!)

They haven't charged end customers for a desktop browser since 2005. Version 8.5 dropped the requirement on the desktop.

Opera do still charge for their Mobile browser - the binary platform-optimised version for mobile phones. But they don't charge for Opera Mini, which is the J2ME version of their mobile phone browser.

Basically, Opera have been moving away from charging the customer unless there's clearly a market for it. Optimising a browser for a phone is difficult and expensive, and many phone companies have done an awful job of it - hence charging for the mobile version. But that's changing, so the mobile version might either die or go free at some point in the future.

Opera are a business, so they have to make money somehow. I'm afraid that this isn't 2001, and business plans in the "... Profit!" model don't actually work.

Most of Opera's revenue doesn't come from end-users. It comes from licensing and customising their browser product for OEMs. For instance, Nintendo paid them to produce the web browser that they use for their Internet Channel on the Wii...

Opera is complaining that Microsoft are illegally abusing their monopoly by bundling one product with another. It's the same complain Netscape made in the US, the same complaint that was upheld in the US, and the same complaint that the US failed to meaningfully punish.

I'm posting this from Opera (9.6) on Linux (Ubuntu 8.04), so haven't the foggiest what your point was there...