The only time I don’t want to hibernate my PC is if I’m walking from one meeting to another that’s two minutes away - like a conference where I use the ten minutes between sessions to catch up on my email before rushing off at the last minute to the next presentation. And with the much faster resume time on Windows 7 (15 to 30 seconds) I might remove my complicated hybrid sleep timer (3 minutes into sleep then 7 minutes into hibernate, in case I find a fascinating conversation and linger in the hall) and just hibernate all the time.

So many updates slipstream without forcing a reboot now that I can just keep going until I choose to restart (assuming Office 2010 sorts out its issues with making Word documents left open from the server during hibernation read only, which I’m working around by using Offline Files - although that has its own issue where the files can’t be saved when I’m actually offline; there’s always something with pre-release software).

But  when we first put the RTM code for Windows 7 on it, Simon’s HP EliteBook 2710p kept waking up like a child asking for a glass of water.  If you upgrade a machine from Vista to 7 and you find it won’t stay in hibernation, check the BIOS. Do you have Wake On LAN turned on? If not, check the disk partitions.

Like many OEMs, HP ships the 2710p with a recovery partition; it has a utility for fast booting and looking at email and contacts in a pre-boot environment and it has what you need to get the original version of Windows restored if you ever need it. That means it’s a system partition that the BIOS needs to know about and that means it ought to be marked as active, but then you wouldn’t be able to boot from the Windows 7 system partition. Sometimes that’s not a problem - but sometimes it means that when you try to hibernate, when the system hits a sufficiently deep ACPI power-off level, it all wakes up again because of the recovery partition - which puts you straight back into Windows. If you do an upgrade install to preserve your installed applications, that leaves the original - and now useless - recovery partition in place. You can remove that and add the disk space to the main partition; we’ve seen that fix some hibernation issues, and on a two-year-old notebook the 8-10GB disk space you get back is well worth it.

However, it didn’t fix the hibernation problem on the EliteBook. We had the same problem with our elderly Elonex media center, which has a new lease of life with Windows 7; with the Release Candidate it was so sprightly that about a minute after we hibernated the machine it would just start back up again, and that didn’t have a recovery partition. The EliteBook didn’t have the hibernation issue with RC, so it’s not a bug. In both cases, a clean install of the RTM code fixed the problem instantly - our suspicion is that it’s an interaction between a driver and an RC to RTM upgrade (which, although, possible, certainly isn’t recommended). In practice, you’re not likely to see this issue on any user machines when you roll out Windows 7, though you might find on your own test system. Bite the bullet and do the clean install; it’s going to give you a more reliable system.

BONUS HELP: if you have a completely different hibernation problem, and you’re looking on the Microsoft knowledge base, you might find a new tool called Fix It. When there’s a registry change that needs making, or some other simple-if-you-know-how fix that you wouldn’t want an end user to mess around in the system trying to implement, many KBs now give you a button to press to make the change for you. No copying keys into your own registry fixomatic scripts, no wondering if the advice site you’re getting a .REG file from is really safe to use. This has been quietly building up since last autumn and you can see all the fixes so far at the Fix It blog or keep them to hand for users with the Fixit sidebar gadget. Invaluable!
-Mary

Remember macro viruses? Trojans and bots have taken over from them in the virus top ten, but there could easily still be binary Office documents lurking in your business’s fileservers with unwanted code in them. The XML file formats introduced with Office 2007 mean you know when a document has a macro by the file extension (an XLSX file can’t have code in, an XLSM can) but even though XML files are smaller as well as more secure, not everyone wants to spend the time to convert a backlog of many years. So to protect you from anything worrying, Office 2010 introduces a Protected View that locks documents when you open them, and runs in an isolated, low-integrity  process with a restricted token (rather like combining the protected mode that IE 8 runs in with the secure desktop you see with UAC elevation prompts - Protected View uses the same User Interface Privilege Isolation).

As the Office engineering blog post puts it, “For a malware to actually be able to run in Protected View it will first need to find a way around DEP, ASLR, GS and our new 2010 Office File validation checks.  After all that, the malware would need to find a way to break out of the sandbox.”

The Office team is confident enough in Protected View that opening and previewing attachments from Outlook will get less annoying; you won’t have to say yes, you trust every different type of document to open and preview individually the first time you come across it. It seems like a welcome security measure that will make life easier too. Sadly, as implemented it’s currently a productivity blocker that will be turned off or loathed by every user that comes across it.

On my system at least, every single document I open in Office 2010, binary or XML, from the office network is opened in Protected Mode and tagged as coming from ‘an unsafe location’. That’s supposed to be for documents downloaded from the Internet (”When a file is downloaded from the Internet the Windows Attachment Execution Service places a marker in the file’s alternate data stream to indicate it came from the Internet zone,” says the Office Engineering blog) and I’m kind of offended that Microsoft is telling me that our network isn’t secure - it is Windows Server 2008 we’re running. I’m also losing time on every document, having to click through before I can start editing.

I tried turning Protected View off; you can’t. You can go into the Trust center, ignoring the sign that tells you not to go in there and not to change anything, and tell Office to trust network documents (again, ignoring the warning that a network is a scary place and you shouldn’t be trusting it) but that didn’t fix it. I had to manually add the file shares on the server, mount point by mount point. You can’t just give office the name of your file server and trust the whole thing; Office refuses to mark the root of the server as safe.

This isn’t supposed to happen, says Microsoft. In some cases, the proxy settings are to blame (check out The LIZ and Proxies: the surprising connection for an explanation by Eric Lawrence of the IE team of why proxies are involved in the intranet at all. We don’t use a proxy. Maybe the Local intranet setting in Internet Options isn’t set to ‘Automatically detect’? It is, as it happen. 

Ah, says the Office team; it’s a bug, and they’re working on it. That’s good news; if I only have to put up with this until the beta of Office 2010 this autumn, that’s fair enough - you expect problems when you use a ‘technical preview’ (or alpha code as we used to call it).

But the fact that Office 2010 is relying on Internet Explorer options that may or may not apply if you don’t have Internet Explorer on your system is a little worrying (Firefox doesn’t use security zones, for example). And Simon, who is joined to the domain doesn’t see Protected View on network documents. So the underpinnings of Protected view seem to be a tangle of Internet Explorer, Active Directory and Microsoft network settings; that’s fine for an all-Microsoft business - like Microsoft. It’s less useful for the rest of the world where heterogeneous networks are the norm and security is important - but will always get demoted if it gets in the way of getting your job done. Let’s hope the bug fix does more than just tweak things; Protected View uses a spiffy new architecture inside Windows and it needs to take a clear and manageable approach to defining what a ’safe’ or ‘unsafe’ location actually is, or it’s going to be unpopular and insecure (cue everyone copying documents onto their laptop to edit them without the nagging and leaving them in the pub car park).
-Mary
 

Microsoft announced RTM today (Wednesday 22nd), but that doesn’t mean you can get the code right now. Depending on how you get Windows, you’ll see it anytime from August to October. Courtesy of the Windows Team blog, here are the dates. All dates are for the initial English version; other languages will follow by the end of October at the latest.

When                           Who

Right now                               Gullible downloaders risking malware to get a leaked version on Russian sites that appears to have significant problems.
now to 2 days after RTM      PC manufacturers
’shortly’ after RTM                IT pros who want the evaluation version from Springboard
August 6th                              software and hardware developers who use Microsoft Connect and MSDN
                                                 TechNet and MSDN subscribers
August 7th                              Volume Licence customers with SA
August 16th                            Microsoft Partner Program Gold/Certified Members who use the Microsoft Partner Network (MPN) Portal
August 23rd                           Microsoft Action Pack Subscribers
September 1st                       Volume licence customers without SA
October 22nd                         Consumers buying a boxed copy or a new PC

We got back from the US earlier this week, and I immediately got to work - tweaking the network in the office. Our main server got a 1Gb hard disk upgrade, and while I was transferring files from our aging NAS to it, one of the two gigabit switches on out network failed. Luckily I had just bought a handful of patch ethernet cables at a Frys, and was able to switch all the devices from the failed SOHO 24-port switch onto a much heftier enterprise-grade 24-port that sits in another part of the office.

That done, it was time to dive into another experiment with Windows 7. We’ve got an Elonex Lumina media centre PC in the front room, and it was time to see just how the Windows 7  Media Center ran. The Lumina’s a decent piece of kit, that folds a PC into the innards of a 32″ LCD TV.  However Elonex as was is long gone, and drivers for the Lumina’s hardware are a little thin on the ground (along with the rumoured firmware upgrade that unlocks the LCD panel’s inputs). Even so,  we decided to give Windows 7 a shot.

Sadly there was one big problem - the ATI Radeon 9600 card that drove the big screen wasn’t supported.

Actually that’s not quite true, as I was able to get it to run at 1024×768, with full glass. The only problem was that the screen was actually a 1366×768 display, just right for HD TV, but not good enough for ATI’s minimal support for old cards under new OSes. That’s always been a problem for Windows OS updates, hardware manufacturers who see the upgrade as a point to instantly obsolete old kit. If it’s not HD and not PCI-E, then you’re not going to get much support from ATI for your old All-In-Ones and Radeons.

I spent some time trying to find the old old drivers, but to no avail. The only option was to replace the card. Of course that meant finding an appropriate AGP card that worked with Windows 7. The old graphics card from a dead PC was the first candidate, and it booted and (even) came up in the right resolution. Nvidia’s Windows 7 drivers were identifying and working with old cards quite happily. Then, disaster.

Suddenly the screen froze.

The old card’s fan wasn’t spinning, and the video card had overheated.

Still, we were somewhere on the right track at least, and I popped out this morning to buy an appropriate card. The Nvidia card I ended up with was passively cooled, and so ideal for a quiet machine in a lounge. It was also tiny - much smaller than the Radeon it was replacing. I hooked it up to the internal S-Video and DVI connectors, and turned the Lumina on.

Success.

Everything was working, and at the right screen resolution.

So now we have a television running Windows 7. The Media Center features are clean and clear, with pleasing transitions and plenty of features.

Now, I think it’s time to test it out. What shall it be? Chuck or Futurama?

When Tim O’Reilly coined the term Web 2.0 for what happened after the dotcom boom and crash (dotcon to dotbomb as the 2001 technology downturn is sometimes known), what he didn’t take into account was the stunningly naive assumptions that people in the technology industry are prone to make.
Web 2.0 was a name for the principle that the companies that survived the crash were the ones that had learned to use the Web as a platform. That’s not just Google Docs and Zoho building Web applications that I personally compare to Microsoft Works or at best Excel circa 1998; it’s Amazon and eBay and sites that get better because people use them and add information (whether that’s reviews of historical price trends).
Google Docs and OpenOffice give me the same feeling I had when I saw Mono on a stand at Web 2.0 labelled Open Source Innovation; what’s so innovative about recreating something that somebody else already built? Far too much of open source is conscious or unconscious reaction to Microsoft.
That’s understandable when you hear the rather self-satisfied way people like Microsoft’s Stephen Elop refer to Web 2.0, innovation and the role of Microsoft; he said that he’d found moving from Silicon Valley to Seattle not as different as he expected when he left Macromedia because it had the same “huge concentration of people concentrating on fixing  problems” and when no-one in the audience agreed that “Microsoft is the most interoperable company in the world” he fired right back “I would argue that’s becoming true.”
Microsoft is working hard on interoperability and standards, and it does have a huge concentration of smart, smart people. But it’s also something of a monoculture. With Windows and Office as a multi-billion dollar revenue stream that’s inevitable. Add to that the number of users (around half of whom pay for Office Elop added) and you have if not a barrier to innovation then at least a brake; innovations that come from Microsoft are going to have been worked out carefully enough that they may no longer be new by the time you get them.
Chatting with a friend who’s planning to move back to Windows from Mac because he’s had too many hardware failures, wants a cheap PC and finds he can do many things on Mac but not write fiction, I urged him to try Windows 7 on the netbook he’s trying out. It’s in beta I told him, but it’s very reliable. There are some drivers that aren’t ready yet, but most things just work (thought I do wish I could persuade someone at Microsoft to slip me a copy of build 7068, say) and the performance is better than XP or Vista. And at this point I thought to myself that if Microsoft were to ship Windows 7 right now, based on the beta plus the improvements I’ve been told are in slightly more recent builds,  it would be like Vista - a good product that needs more work from both Microsoft and the ecosystem. This time around, the Windows team is going to hang on until it’s done.
Will we call it innovative when it arrives? Will it be a platform that inspires developers with its potential the way Web 2.0 does? Will it look as cool as Palm’s Web OS - which is also likely to arrive some time between April and September? Windows 7 isn’t a dramatically new idea of what an OS is (and Windows 8 may not be that either), but neither is Web 2.0 (which in many ways grew out of Outlook Web Access, the original Ajax application). Web 2.0 is about discovering how much more you can do with the technology you already have. If Windows 7 can unlock the value in servers and applications and storage that Vista didn’t quite reach, then that should certainly count as innovation.

-Mary

Windows and IE get the features that people use; that’s good if you’re in the majority.

In early betas of Internet Explorer 8, when you typed a search into the address bar rather than the search box, the list of results you got was five URLs from your history and five URLs from your RSS feeds.  Less than 1% of the IE8 beta testers ever clicked on those RSS results and most people said what they wanted was URLs they’d typed in - so the shipping version of IE 8 doesn’t give you RSS results when you search in the address bar, just ten results from your history. ..

Well, actually, it does, but only if you dig into the options and turn RSS results back on. There’s some confusion in the IE team about this. Depending on who you ask you’ll be told it’s relegated to an option and off by default or it’s gone completely because having everything be an option increases complexity and the number of combinations you have to test. The ‘ordering pizza for 150 million people’ analogy that Dean Hachamovitch is applying to simplifying Internet Explorer and Steven Sinosfky is applying to simplifying Windows comes from a boss they once both reported to and it’s true that you can’t have all the features and only the features that suit everyone, and that if everything is an option there will be many users who never figure out the right options for them. (And that the more combinations you have to test, the longer it takes to ship.)

If RSS results from the address bar had gone away, that would be bad news for 1% of IE 8 users, which is quite a large number of people - except, of course, it’s not really 1% of everyone who uses IE. It’s 1% of everyone in the beta test group, Paul Cutsinger of the IE team told me, and “it would drop way down in the mainstream population”. Beta users are different; we’re actually prepared to use beta software for a start. “They’re early adopters, they tend to be power users, they tend to have more tolerance for problems that show up - but they also complain more!” Dev teams at Microsoft have to compare what they see from beta users with what they see in usability labs and in other tests. They have to filter out the biases from internal Microsoft users as well; nearly half of all Windows user have between six and nine windows open at once but most Microsoft employees have 20, 30 or 50 windows open at the same time.

Balancing that out is a difficult problem and I’m not sure Microsoft always gets it right. In the M3 and beta builds of Windows 7, Win+E opens the Explorer pointing at your document Libraries; in RC it will go back to opening My Computer, so that wherever you want to go is a click away. I should start by saying I’m a huge fan of libraries - I’ve been waiting for years for an easy way to search all my documents without having to remember which drive the one I want is most likely to be on. But most people who open Explorer are going to be looking for a document - at least most of the people who will use the final version of Windows 7. Beta users are implicitly more technical folk so a higher percentage of them will be going to a variety of places and so they want My Computer; but users will want documents more often than they’ll want multiple drives. Libraries are the way that we’re ‘intended’ to get at the majority of our files so having them as the default target of Ctrl-E made huge sense to me, both for immediate use and as the way going forward. If Libraries become as widely used as Microsoft must hope, not having them be what you see first every time will come to seem a confusing thing in a few years time.

And that’s the other problem with relying on users voting with their mice, even if you manage to remove the bias of early adopters, technical experts and other oddities. If what you want to do isn’t already a feature, how can Microsoft see in the statistics that it’s what you want to do? And if there’s a visionary feature that may not become part of the way you work until you’ve used it for a while, should Microsoft give up on it because the usage isn’t there at first? At the MIX conference this week, Senior User Experience Designer Stephan Hoefnagels claimed that the taskbar in the 1985 release of Windows 1 predated the Apple Dock by 15 years. If you want to take credit for a feature, you have to have the courage of your convictions and make it prominent - not hide it away behind the old way of working, even if it doesn’t win a popularity contest on the first day.
-Mary

I love the way you can drag a window into the corner of the screen in Windows 7 and have it resize to fit half the screen; it makes it so easy to arrange windows neatly. Unfortunately, it doesn’t work quite so well with multiple monitors. So far I can only get the documents to fit themselves neatly to the left of the left-hand screen and the right of the right-hand screen, which makes sense – the corners are the triggers and logically there are two even if physically there are four.
 
I’ve asked Microsoft if there’s a trick to making this work, but in the mean time, here’s my work around. I drag the first window I want into the left-hand corner on the left-hand screen and it resizes. Then I drag the document I want next to it into the same corner, let it resize on top of it and drag it by hand to the right. It’s already the right size so it’s easy enough – but if I just drag it back to where it was in one move, it goes back to the original size (which in any other circumstances is great), so I have to drag it a little, drop it and them drag it again.

It would be much easier if Windows spotted the end of the taskbar as a logical point to work with as well.

Almost everyone believes there should only be two or three versions of Windows; no two people agree on which two or three versions those should be (to paraphrase Brandon Le Blanc over at the Windows Team Blog).

With the SKU list for Windows 7, as Simon pointed out, any one person will only ever see two or three versions. If you’re a home user you’ll see Home Premium (and wonder what’s premium about it), Starter if you’re looking at the cheapest of netbooks and Ultimate if you look at really over-powered gaming rigs. If you’re a business user you’ll see Professional or Enterprise, depending on the size of company. It’s Windows Me and Windows 2000 all over again. So no, there aren’t too many versions of Windows 7; there are the number almost everyone thinks we should have.

But regardless of whether these specific versions are the right ones for users (I think they probably are, although I still see people calling for Windows Home Runs On All My PCs For The Same Low Low Price, Windows Poor Student, Windows Actually-Poor Teacher, Windows Small Business Because I Don’t Realise That’s What Professional Is, Windows Componentization Because That’s My Theory For ‘Fixing’ Windows and a few others); does it make sense for Microsoft to continue to have more than one version of Windows?

Don’t tell me there should be one version of Windows because there’s one version of OS X. There’s one version of OS X because there aren’t enough business features to need a separate business version.

It is a good thing to have a cheaper version of Windows that doesn’t have business features home users don’t need like joining a domain. It’s a good thing to have irrelevant options, like joining a domain, totally removed from home versions so no one turns them on by accident (or by following a tweaking guide that promises to fix things that weren’t broken in the first place). It’s good to be able to upgrade from Starter to Ultimate when you realise that not being able to use an external monitor with your netbook is annoying, or from Home Premium to Ultimate if you want to use BitLocker (which I still think should be in Windows 7 Professional). It’s good for businesses who buy Microsoft software year in year out to get a loyalty bonus with SA and Enterprise. But does Microsoft make enough money out of the different licences to make all the hard work make business sense?

There’s the management time to decide what’s in the different SKUs, explain to everyone, explain it a couple more times and justify it a whole bunch more times. There’s the programmer time to mark each feature for each build, and the regression testing to make sure each SKU of features works properly together. I’m willing to assume, given the high bar the Windows 7 pre-beta and beta have set that we’re not going to have a backup interface that offers me features that won’t work in my version, for instance.

There’s also the cost of the encryption system that protects all the features that aren’t in the SKU I’ve bought, because they’re all on the hard disk whichever SKU you install (so no, having different SKUs doesn’t save me disk space on features I’ll never use. And given that Windows 7 handily unloads services I’m not using any more and reloads them as needed, I wouldn’t be wasting memory or cpu cycles on them either). There’s the legal costs of pursuing people who have a go at cracking that encryption. There’s pressing DVDs with two labels and printing two sets of boxes and retailers managing two areas of shelf space….

I asked Laurence Painell, the UK Windows marketing manager who deals with OEMs. His answer is that Microsoft doesn’t discuss those kind of commercial questions and that the same protection system is used for Office-ready PCs where you don’t have to install Office for all your users, you can just distribute the keys and it appears as if by magic, so it’s not all on the Windows balance sheet. 

Either the OEMs and customers want multiple versions of Windows enough that all these costs are worth it, or having different SKUs turns out to make Microsoft more money than having one version. I rather suspect it’s the former. I just hope I’m not going to be looking at too many netbooks with Windows 7 Starter on next year, because I’ll be upgrading them straight away…
Mary

Usually I say the fastest way to speed up any PC is to stick some more memory in it, but with cheaper notebooks and netbooks it’s actually more likely that your hard drive is slowing you down.

Take the Toshiba R400 I use as a rather stylish but not very speedy bedroom browsing machine. In its time this PC has had everything from Web editing tools to databases to three different 3G connection tools to the unremovable Nokia PC Suite on it and the parade of utilities and apps has taken its toll. I’ve been having problems for a few months that I was pretty sure SP1 of Vista would fix - I just never seemed to have the time to do the update because the same issues that would stop Outlook and IE from launching had the  update failing when I tried to do it online. If I have to spend an hour downloading and running an offline update, I thought, it’s just as fast to spend half an hour upgrading to Windows 7; if it runs happily on a netbook then it should do just fine on the R400.
 
Apart from the size (12″ screen), the fact that it’s the slimmest convertible Tablet PC you can get and the price that those extra features put on it, I tend to think of the R400 as the first netbook. It looks lovely, it’s white and glossy, it’s very portable, it doesn’t have much battery life and the performance is limited. Not by the Centrino Duo chip, not by the memory (I’ve got 3GB in it) - but by the 4200RPM PATA drive. To fit in everything, Toshiba picked a 1.8″ drive and at the time that meant using the same hard drive that you’d find in an iPod. It’s still impossible to find anything beyond a 5600RPM 120GB SATA drive in the 1.8″ format; Toshiba has announced higher capacities but they’re not on the market for upgrades…

Many netbooks pull the same trick; by comparison the 5600RPM drive in the Lenovo S10 gives it the kind of performance you thought the Atom didn’t have the oomph to deliver.
 
Why does the drive speed make so much difference? Even 1GB of memory is enough for whatever you’re doing at the time to fit in memory but as soon as you switch to another program Windows does a ‘context switch’  - and pages out to hard drive for virtual memory.
 
Upgrading the R400 to Windows 7 beta went almost flawlessly; of course we had to stop and put SP1 on it first because you can’t upgrade without it, so it wasn’t actually a timesaver, but it’s the smoothest OS upgrade I’ve done in a long time. With Vista I’ve always recommended a clean install: 7 coped happily with the rash of apps, including Nokia PC Suite, and even recognises the 3G connection and shows it next to the list of Wi-Fi hotspots. But while things do feel a little nippier, and CPU usage was down around 1% with a couple of apps open in the background, I thought I’d run the Windows Experience Index to check how it rates the R400.

It gets a 2.0. not for the processor or the memory or the graphics; in fact this is the first time I’ve seen a notebook where the overall score is lower than the graphics result. It’s the hard drive transfer rate that brings the machine down - which now reflects exactly what I’ve seen in practice.
 
The Windows team has been talking about WEI and disk performance recently, including why they’ve changed the results from what the same configuration would have been rated as under Vista. It’s not that 7 has worse performance - it tends to be faster on the same hardware - or that your hard drive has magically got worse - it’s always been that bad; the WEI is just better at rating the performance you’ll actually see.
 
The same discussion reveals that a number of first-generation SSDs score very badly on performance under heavy load because they built up a backlog of data transfers and slow down to deal with them - scroll down for an excerpt with the details. Microsoft doesn’t feel that it can name names, which is a shame; unless all those drives are off the market by now, it would be nice to know what to avoid before you’ve bought a machine that gets a surprisingly low hard drive rating in WEI.
-Mary

“With respect to disk scores, as discussed in our recent post on Windows Performance, we’ve been developing a comprehensive performance feedback loop for quite some time. With that loop, we’ve been able to capture thousands of detailed traces covering periods of time where the computer’s current user indicated an application, or Windows, was experiencing severe responsiveness problems. In analyzing these traces we saw a connection to disk I/O and we often found typical 4KB disk reads to take longer than expected, much, much longer in fact (10x to 30x). Instead of taking 10s of milliseconds to complete, we’d often find sequences where individual disk reads took many hundreds of milliseconds to finish. When sequences of these accumulate, higher level application responsiveness can suffer dramatically.

With the problem recognized, we synthesized many of the I/O sequences and undertook a large study on many, many disk drives, including solid state drives. While we did find a good number of drives to be excellent, we unfortunately also found many to have significant challenges under this type of load, which based on telemetry is rather common. In particular, we found the first generation of solid state drives to be broadly challenged when confronted with these commonly seen client I/O sequences.

An example problematic sequence consists of a series of sequential and random I/Os intermixed with one or more flushes. During these sequences, many of the random writes complete in unrealistically short periods of time (say 500 microseconds). Very short I/O completion times indicate caching; the actual work of moving the bits to spinning media, or to flash cells, is postponed. After a period of returning success very quickly, a backlog of deferred work is built up. What happens next is different from drive to drive. Some drives continue to consistently respond to reads as expected, no matter the earlier issued and postponed writes/flushes, which yields good performance and no perceived problems for the person using the PC. Some drives, however, reads are often held off for very lengthy periods as the drives apparently attempt to clear their backlog of work and this results in a perceived “blocking” state or almost a “locked system”. To validate this, on some systems, we replaced poor performing disks with known good disks and observed dramatically improved performance. In a few cases, updating the drive’s firmware was sufficient to very noticeably improve responsiveness.”

Michael Fortin, Engineering the Windows 7 WEI

David Mitchell of Ovum suggests there are ‘profound similarities between Windows Vista and Windows 7′, although he doesn’t actually list them. Up to a point, Lord Copper.
Is Windows 7 the same as Vista?
No.
Really?
No.
Really really?
No.
Is it descended from Vista?
Yes - and Windows Server 2008 R2.
But it’s different?
Yes.
Really?
Yes! You try running Vista on a 1GB Atom netbook…
Is that a bad thing? The descended from Vista part, not the running on a netbook part…
No.
Really?
No.
Really really?
Vista went on for too long, changed too many things, lost sight of its direction, lost two of its three pillars, didn’t have the right drivers at launch and badly needed its SP1 to smooth over the cracks. But even before SP1 it was a huge improvement over XP and laid the architectural foundations for moving forward. 7 moves forward, builds on Vista, improves small subtle things that make a big difference to usability and has enough new features to fill an 80-page briefing - and it avoids all of Vista’s mistakes so far. Sensible and confident is a better look for Windows than glitzy and flamboyant anyway.
What about UAC?
Oh, it won’t annoy you any more. partly that’s because software developers have learned to love limited user; partly it’s because users telling Microsoft that they didn’t want to be warned so often about potential security breaches has struck home.
How about not having any new features in the beta?
Are you complaining that Microsoft was doing so well on Windows 7 that it could put nearly all the new features into the pre-beta and have them work?
Shouldn’t the beta be exciting and make me want to buy the product?
No, the beta should tell you what’s coming and how well the coding is doing.
I don’t want to search for things on the Start menu. I want to have a huge expanding menu that fills my screen that I can organize just the way I want.
That’s not a question, it’s an opinion. I want a 30″ screen I can drive over USB that folds out from the wall on an extending arm. Both take up a lot of room and neither are ideal for the vast majority of people. More usefully, I’d suggest you might want to give the new UI a try before you say it won’t work for you. Do try it on a 7″ netbook as well as a 30″ screen before you decide that the XP UI from 2001 designed for a 15″ display is definitely what’s right for today…

Mary