Wednesday, 26 September 2007

Excel 2007 bug: result of 100000, should be around 65535

If you use Excel 2007 in any serious capacity, you really should know about this issue David Gainer blogged about in the early hours (my time, at least) this morning:
Calculation Issue Update
The short version of the story is that if you have any calculations with a result of around 65535, they may show an incorrect result of 100000.
I expect a few people on the product team at Microsoft will have a very uncomfortable few days. Hopefully David will keep us updated through his blog.

Tuesday, 25 September 2007

Why I like PCs

I hate to jump on a bandwagon, but Stephen Fry's epic post about iPhones, SmartPhones, and technology in general got me thinking. (Everyone and their dog has referred to his blog recently, it seems, including me now. And worse, I've shamelessly referenced it and I'm not even going to talk about mobile devices.) The age-old rivalry between PC and Mac fans, Fry being a proud Mac devotee, is something that's always baffled me. And when I say rivalry, I really mean hatred for each other's choice of product. If I knew people who drove Ford Focuses (Focii?) who absolutely detested Volkswagen Golfs then maybe I could understand it, but I don't.

I've always liked PCs for the same reason I like the English Language. I learned to use each of them before their alternatives and they're familiar to me; we go back a long way. I’ve had a lot of practise so I can now use both reasonably competently, and they serve me more than adequately. I believe there are parallels in the way the PC and the English Language arrived at where they are today. They've grown, absorbing components from anywhere and everywhere, evolving as they needed to because no one organisation had control – at least in recent times. Standards have appeared, but mostly because conventions were adopted and became widespread, then were documented and made official - not the reverse. I would argue that both are better for being products of chaos and evolution, not control and design.

There's little doubt that a bog-standard Mac is more aesthetically pleasing than a bog-standard PC or laptop. But if I could have an upgrade to ‘pretty’ for my PC for £50 I don't think I'd bother. It sits under my desk, for practical reasons, so it would be a waste of some designer's efforts, not to mention my cash, for a pretty PC not to be seen. As long as it isn't conspicuously ugly, I'm quite happy. Nondescript suits me just fine. To borrow Stephen's analogy about concrete buildings, they don’t bother me if they’re the most practical solution to a practical problem. Unless they’re truly hideous, they usually only become offensive when placed next to a thing of beauty, which is why my PC doesn’t sit next to a vase of lilies.

In fact, I just wrote that last paragraph completely forgetting the small and (to my mind) attractive glossy black AOpen machine I recently built as a server. It sits on a bookcase in my lounge. I did pay a bit extra for the case, but mostly because it was small and had quiet, integrated cooling. There probably were some butt-ugly alternatives which I didn't even consider - so perhaps there are times when design is important. But I still chose the nicest case available for the technology I wanted, not the other way round. It's just a bonus that I didn't have to find a place to hide it.

Then there’s the idea that choosing a Mac or PC, Windows or Linux, says something about who I am, as if my laptop is somehow a symbol that other commuters on the train may use to judge me. If my choice of computer makes some kind of image statement, then I hope the statement is just that I don't care enough about my image for it to dictate my choice of computer.

As for the aesthetics of the desktop operating system, I quite like Windows Vista. Is that controversial? The whole Glass thing is easy on the eye, and it’s very useable. (I say that, but I frequently use a remote connection to my main machine which immediately disables all the flashy graphics to conserve bandwidth. I barely even notice a difference.) I even remember being impressed by Windows XP when I first saw it, at least after I switched off the Teletubby Hill desktop image and the Fisher-Price inspired Start Menu. I expect the Vista ‘look’ will get tired quicker than any MacOS just like XP did but mainly because of its ubiquity. And even then it won't bother me - it doesn't have to be a work of art or even visually pleasing, it just has to be reasonably intuitive to use without being downright ugly. Windows 95 looked sleek and sophisticated compared to Windows 3.1 - it just shows how our expectations change.

As with the English language, another reason Windows feels natural to me is because its history and culture don't need to be explained - I know that CTRL+Insert does the same thing as CTRL+C because it's a legacy from DOS days. (I choose that example because I'm a leftie-mouser and those keys are convenient for my right hand, so I’m happy that trick has been preserved.) I expect I could learn to like a Linux or MacOS desktop, but it would take a long time for them to stop feeling somehow foreign. My TiVo is a Linux box and I love it dearly as a PVR, but if I need to telnet to it to perform some housekeeping (yes, of course it's hacked and networked), I need step-by-step instructions for all but the most basic tasks. It’s like trying to speak Russian from a phrase book when the only words I’ve committed to memory are “hello”, “goodbye”, “two beers, please” and “thank you”.

Ultimately, I guess my point is that technology isn't religion. Nobody has to choose one computer or OS and forsake all others. We don't have to engage in some holy war over who is right or wrong. If your circumstances require you to use a piece of technology you hate, I’ve been there. It sucks, and I sympathise, even if the very thing you hate is the one I would choose.

Monday, 10 September 2007

Daylight Saving 3: When 'summer' moves

In my part of the world we've hardly had anything resembling a summer... I'd love to know where ours got moved to.

Seriously though... in my previous post I showed how Outlook appointments appear to move when people in different regions 'do Daylight Saving Time (DST) differently. Now we get onto the really fun stuff... when legislature keeps us on our toes by moving DST about.

In 2005, the US Government passed a bill to make DST last longer starting from spring 2007. Various other countries followed suit, and in total 33 time zones changed. Software vendors had to issue patches to all their systems that were DST-aware, Microsoft included. Patches for Windows were released to cope with the change so everyone's system clock should automatically update on the correct weekend.

Because of what we already know about how Outlook stores appointments, this caused meetings to appear to move about. An appointment would have been affected if all the following were true:

  • The appointment was scheduled from one of the affected time zones (i.e. one whose DST rules changed)
  • The appointment was created on a machine with the unpatched version of the time zone information (i.e. one where the meeting was created before the Windows Updates were applied or possibly before they existed)
  • The meeting occurred in one of the 'delta' periods, i.e. the dates between when the clocks would change under the old and new rules (so the spring delta ran from the second Sunday in March to the first Sunday in April)

Here's an example of how an appointment would appear to move just by applying the Windows DST patch to a user's machine:

Appointment time is stored in UTC when the appointment is first created, i.e. before the update was applied to the user's machine:

Appointment time is displayed in Local Time, calculated at the time it is viewed. So this is how it will appear after the update is applied to the user's machine:

Local time in NY




Adjust for Daylight Saving


Adjust for Daylight Saving


Adjust for GMT-5 time zone


Adjust for GMT+5.5 time zone




Local time in NY


*according to unpatched settings, 15th March is not within Daylight Saving Time

*according to patched settings, 15th March is within Daylight Saving Time

(If those tables don't make any sense, check out this post for an explanation.)

Microsoft released some 'rebasing' tools to help users 'fix' the appointments that appeared to have moved. These were mostly sufficient for the average home user, but for large organizations they had the following drawbacks:

  • The tool could only reliably tell if a recurring appointment had the incorrect DST information and needed fixing. (This is because single-instance appointments don't store time zone data at all.)
  • The user-side tool wasn't easy to automate because of the risk of 'fixing' a single-instance appointment that was already correct.
  • The first iteration of the server-side tool aimed at network administrators was slow, unstable, and required constant monitoring. For an environment of more than a few thousand users it just wasn't a realistic prospect.
  • The next iteration of the server-side tool was released very late - only about a week before the first delta period began, if I recall correctly. Many network administrators felt there wasn't sufficient time to test it, and in any case had already been forced to make alternative arrangements for their users.

Which left a lot of folks looking after large enterprise environments in a bit of a mess.

The organization I was working with at the time has a huge number of users, many of whom collaborate with colleagues across continents on a daily basis. These people are very dependent on their Outlook calendars and take it very personally if someone messes around with their data. Stability was their highest priority, so with the timescales involved we decided the safest thing we could do was provide recommendations to our users of how they could manually fix their own appointments. Our IT support teams out at the sharp end were armed with the client-side rebasing tools to use for coordinators and other people who schedule a massive amount of appointments. We decided it was best for those folks to have some personal help to be on the safe side. Maybe we were too cautious, who knows... without splitting the organisation in two and trying a different approach on the rest of them, it's impossible to say.

Next time I'll share some assorted DST facts that didn't make it into the first three posts.

Creative Commons License This work by TechieBird is licensed under a Creative Commons Attribution-No Derivative Works 2.0 UK: England & Wales License.