EDIT: moreover, the GUI designer broke the holy mantra of "If it works, and it works well, don't fucking change it".

That's a stupid "mantra". Every successful new invention is preceded by failed attempts. Most advantages of a new technology only become apparent after it has been implemented. The first human to send a smoke signal was probably very proud of his invention. Smoke signals worked well. And now we have the internet.

Its a shame I cant upvote this :) so you can have a gif instead

For non-gamers, a computer from 10 years ago is perfectly adequate.

...browsers...

Shitty web sites full of crap never perform well, regardless of hardware. I have a high-end hex-core desktop machine and it still struggles with incompetently designed web pages. Web site designers insist on loading their pages with dozens of ads, each running as much javascript, flash, or streaming videos as they can get away with.

I assign the blame firmly on the web development community here. There's nothing anyone in the hardware or operating system environment can do to fix stupid decisions made by people that write software layered on top of them.

Ok, now I can die HAPPY.

Build 10074

The fundamental problem Microsoft have is that there is now a significant section of the population who don't need Windows at all.

Previously, if you had minimal computing requirements (email, web, store your digital photos, write the odd word* doc), you bought a low-to-medium PC with Windows whatever on it.

These days, those needs are easily filled by tablets, phones, etc.

Of course, there's still a massive market for Windows PCs among enterprise and power users, but as more and more enterprise software goes on the web, Windows becomes less of a requirement for businesses. The majority of business users could easily get by on a chromebook.

* and yes, it was a word doc, because that was the defacto standard)

I've been using the Windows 10 previews since the first surfaced, and use them almost daily, but not constantly throughout the day. My job also involves regular use of other Windows versions, both desktop and server, from 2000 up to 8.1/2012 R2.

I haven't used the latest build (10074) yet, but I will do so tomorrow.

To be honest, I think the Windows 10 UI is shaping up to be another misstep. Now, I hated Windows 8 on release, but since 8.1 and particularly with the monthly updates towards the end of last year, it's improved immensely and I find that the Start Screen is a very effective application launcher. It's actually difficult to go back to 7 once you get used to it, and I now exclusively use a fully updated 8.1 on my own desktops. Never looked back.

Windows 10's UI on the other hand seems to be trying to combine both approaches and not succeeding too brilliantly at either. It's Start Menu equivalent is missing the functionality that made the old Start Menu really great, and by ramming in the Tiles the amount of on-screen real-estate available for collecting and organising the programs you really need to use most often is compromised. It's Start Screen equivalent loses it's effectiveness by again compromising on the available real-estate and by being too "busy".

This is all subject to change as more preview builds emerge, of course, but with the projected release date, what we're seeing now must surely be pretty damn close to what Microsoft are actually going to ship. On the other hand, news like the re-emergence of Aero Glass in build 10074 does indicate that late changes are still possible, and it's always possible that once I get down to using it every day my opinion may change.

Overall I think they're shipping too early. I think that taking another few months to polish the experience, instead of rushing to put out a first-cut that's inevitably going to be rough around the edges, would have benefitted things. I expect that the first few months of updates will deliver what should have been shipped, but I wonder if the damage will have already been done by then. People have very high expectations of Windows 10, especially after the bad rep that the initial releases of 8 got, and in it's current state I just can't see it delivering.

It's also the case that Microsoft are very heavily pushing integration between desktop and mobile in this version, but yet the current previews of Windows Phone are even rawer than the desktop previews - there's just no way that's going to ship until much later this year. So the promised integration is just not going to be there from the outset.

So I wonder why they're shipping so early. The original indicators were that they would ship later this year - say Q4-ish - but current indicators are that it's going to be July. The only explanation I can think of is that something pushed them, and I wonder if it was the upcoming Vulkan vs D3D12 battle. Shipping in Q4 would mean that we'd probably have a Vulkan spec and the first drivers already out (I expect them at Siggraph) some months beforehand, which would give Vulkan a headstart. Whatever it was, it's going to be a shaky first few months for Windows 10 unless things really come together very quickly over the next month or so.

Of course, there's still a massive market for Windows PCs among enterprise and power users, but as more and more enterprise software goes on the web, Windows becomes less of a requirement for businesses. The majority of business users could easily get by on a chromebook.

This is something that's frequently enough claimed, but it doesn't actually work out in practice.

The gotcha is that every enterprise has legacy software, and much of that legacy software is heavily tied to the desktop. It may be dependent on old VB versions, old Java versions, there's even still a lot of Lotus (spit) Notes (spit) out there.

In amongst that mess there are going to be applications that are absolutely essential to the running of the enterprise. It may be old, it may be cruddy, it may be held together with sellotape and rubber bands, but it works and it must keep working. You cannot break that software and you cannot move to a platform on which it no longer works.

The obvious ideal world solution is The Great Rewrite, but most enterprises never do that. They may no longer have the source code, even if they do have it they may no longer have staff that fully understand the source code, and in any case The Great Rewrite is a major project that diverts staff and funds from the day-to-day core business of the enterprise.

So the cruddy old legacy software stays as-is, the IT guys perform unsung feats of heroism to keep it working, and the shiny future of everything being web-based and capable of being run on a Chromebook never happens (or best case only happens in startups, who will eventually become the enterprises with 20-year-old cruddy software - this time web-based - in 20 years time).

Looks like they bring more desktop customization to the new GUI in build 10074, however tons of settings are still in a sort of limbo, you can find the only with a search in the old control panel. There are 2 month left to improve GUI and fix bugs...

Looks like they bring more desktop customization to the new GUI in build 10074, however tons of settings are still in a sort of limbo, you can find the only with a search in the old control panel. There are 2 month left to improve GUI and fix bugs...

You are realising that you are complaining about a piece of test software right. Lots of the UI stuff can quite easily change before release.

Looks like they bring more desktop customization to the new GUI in build 10074, however tons of settings are still in a sort of limbo, you can find the only with a search in the old control panel. There are 2 month left to improve GUI and fix bugs...

You are realising that you are complaining about a piece of test software right. Lots of the UI stuff can quite easily change before release.

Yes, I do. But do you realize that the same complaints come at every Windows major release because there are still pieces of the NT 5.0 era (and even some '95-Me era!) around the full OS breaking GUI consistency and user experience?

2 moths are left, July is the month of pre-RTMs (OEMs and driver summons, some very minor and very quick bug fixes) and RTM. From RTM to GA there are only bug fixes that you will download at day-on or similar.

I should highlight that even the most recent publicly released build is probably somewhere between a week to a month (if not more) behind the current state of the code.