May 2007 Archives

Microsoft Surface.  There was a leaked demo of this a while back, but now it's official.  I think the idea will be best implemented in an environment where there will be no danger of, say, a cat walking across the table.  Of course, if they're smart, they'll have some kind of panic button that you can smack to neutralize the surface when something like that happens.  I'm also a little leery of the "context switching" issue -- if I have this and a keyboard, wouldn't it get wearying reaching back and forth?  (Then again, this isn't too much of an issue with a mouse and a keyboard, so I suppose it won't be all that bad.)

Still, I have a sick mind when it comes to contemplating what can go wrong with anything this new and shiny.  In 50 years, I'd bet we'll have a three-dimensional holographic spatial display system with gestural controls ... and my pets will probably end up walking through it and out the other side with cat macros hanging off their fur.  With every great technological advance comes the power for the most amazingly boneheaded silliness.

Back in 1994, Daniel Wiese (now of Microsoft Research), Simson Garfinkel and Steven Strassman wrote a trenchantly funny book called The UNIX-Haters Handbook.  Much of the material in it was culled from discussions on a mailing list of the same name (UNIX-HATERS), and the book has since been released for free with the author's permission as a PDF document.

A lot of the humor in the book doesn't come specifically from slagging UNIX, though, but from the general sense of frustration that computer users of all stripes experience when dealing with something that has been seemingly built to make their work impossible.  I couldn't count the number of times Windows has done things that I found counterintuitive, or seems to have been built in some inane way.  (As things have moved forward, though, many of those problems have been dealt with -- although sometimes at the expense of even bigger problems.  But that's another story...)

I've been living and working with Windows Vista for about six months now, through both good times and bad, and on a variety of different machines: my desktop, a couple of notebooks, a few Virtual PCs, and so on.  I've come to like it even when at first I thought some things were very strange, but that doesn't mean there's room for improvement.  Here's a grab bag of things that stand out.

• Reliability and Performance Monitor.  Once you start using this tool, you get hooked on it.  As I mentioned before, it's a great way to find out what might be thrashing your hard disk or hogging your CPU.  Use it in good health and bad.
• UAC.  We hate it, but it's probably here to stay, like Product Activation (yecch and double yecch).  Most of the people who actually like UAC, I've found, are the people who have seen it do what it was intended to do: alert the user to an unauthorized escalation of privileges.  After it trapped one program trying to do just that, I decided UAC was more than worth having handy.  I am reminded of the time I ended up owning a dog, and was rather dubious about keeping him -- until the night he frightened off someone trying to sneak into my backyard.
• Indexed Search.  Entirely too useful, especially with Outlook, although I wish other programs could make more aggressive use of it.  I suspect in time they will.  One problem I did have with Outlook 2007 initially involved it getting stuck and not indexing everything, but after I deleted and recreated the index the problem went away.  Another thing you need to be careful of is to only set up the directories that you need indexed.  Indexing a whole drive, apps and all, is a waste of time and performance unless you have a very specific reason for it.
• Boot management.  It sucks, but you can work around that.  For the few machines where I needed to do multiboot with Vista, I set up Boot-It Next Generation and used that to manage everything.  Trying to work with Vista's boot manager to do anything other than create different boot scenarios for Vista itself (and even that's a pain) is, as someone else once said, like trying to build a bookshelf out of mashed potatoes.
• Windows Photo Gallery.  This wins the "Vista Tool I Never Thought I'd Use" Award.  Like a lot of other people I've got tons of digital photos, and after dumping them into WPG and messing with the program a bit I realized that I had pretty much the only tool I needed to make sense out of them all right there in front of me.  I still use IRFANVIEW for casual image viewing and editing, but for organization and tagging WPG is terrific.
• Speed.  On the same hardware, all other things being equal, Vista is at least as fast or faster than XP SP2.  It may not happen immediately, because Vista needs to learn a bit about your work habits first -- i.e., figure out what programs you're running and set up pre-caching appropriately.  (I know some people are reporting that Vista "churns constantly," and I think part of that is observer bias: they see the HD being active, when it's indexing or defragging, and they think that means the system is suddenly busy.  I've found that if you start doing something, the churn turns out to be entirely passive, because it's been pegged as a low-priority process.)  One thing I'm going to be curious about is testing a RTM build of Vista and the inevitable SP1 for speed on the same hardware, and see how they do.
• Defrag.  I got more mail about Vista's changed defrag app than I did just about anything else I wrote about in Vista (save possibly the backup tool -- see below).  For the most part, people hate it, although a) I like the fact that you don't need to run defrag manually anymore, and b) I'm still of the opinion that the graphical statistics that most defrag programs give you are not going to give you the whole picture -- and in fact may give you wholly misleading information that doesn't really tell you what needs to be defragged and why.  If you want a good third-party defrag program, though, I'll recommend JKDefrag, which has both command-line and graphical modes.  (One caveat: JKDefrag does not honor the prefetch directory's layout information, so after you run it any prefetch data created by Vista will have to be recreated from scratch.  I understand the author is aware of this and is working on a way to accomodate it in future releases as soon as he figures out how to do that.)
• ReadyBoost.  I still don't have a decent USB drive to test this out with yet, but I'm wildly curious especially given the crazy mix of programs I run all the time.  I plan on going shopping next month for one that is certified to work well with Vista.  I'm curious to see how it performs on 512MB, 1GB and 2GB systems.
• Backup/restore.  In a word: BLECH!  The backup application is definitely brain-dead -- but I've stuck with it despite myself and found a few interesting ways to conduct end-runs around its dumb design.  One which I plan to write about soon is how to exclude directories from a full-drive backup without moving them somewhere else (although this trick only works on the C: drive).  More on that later.  Also, the full-system backup only does complete images, as far as I can tell; it doesn't perform differential/incremental backups at all. which further limits its utility.  (Stop me if I'm wrong with this.)
• Hardware support.  64-bit support for legacy hardware is a no-go, so I wasn't able to use 64-bit Vista at all.  Then again, 64-bit Vista on the desktop isn't as critical as I thought it might be, if only because the benefits of 64-bit really don't come forward unless you're running with more 4GB of memory and using applications that demand that kind of load.  That probably explains MS's not phasing out 32-bit Windows on the desktop (rumors to the contrary): it's still quite useful.  I don't see me going to 64-bit anytime soon unless the need really emerges -- and unless a new scanner and printer become part of the picture, it ain't happening.
• Notebooks.  The better your individual notebook manufacturer has provided support for Vista, the better you'll be.  Support for notebook power management with only the default / stock drivers seems to be all over the map -- and as far as Aero goes, it isn't even supported on my notebook's chipset, so that's a moot point.  (Blame Intel.)

Am I happy with Vista?  I could be happier, but right now I'm happy enough that I use it as my default OS, and I suspect that's the most important thing.  Now let's see where they go from here.

A friend of mine is a Realtor, and recently acquired a new smartphone to try and stay on top of her ever-increasing mountain of work (cue sarcastic snort here).  On a whim, I dredged out my new phone to compare against her -- and mine was nothing, really, compared to hers; it was just one of the free Samsung models offered through my cell company (T-Mobile).  But it worked, and for what I needed it for, it worked well: it has a hands-free mode that's better than the actual speakerphone on my desk (!), and it's far lighter if a bit bigger than the model it replaced.

Her phone sported everything you could think of -- full QWERTY keyboard, digital dual-band support, you name it.  Well, almost everything.  Turns out that a week after she'd bought the thing (out of her own pocket, too), she'd dropped it in the bathtub and ruined the keyboard.  The last time I dropped a phone in water (my earlier Samsung, also a freebie), I disconnected everything, let it dry out for 72 hours, and fired it up. It worked perfectly.

Now I remember why, whenever I see someone else's bit of $400 digital bling, I get jealous and then immediately check myself.  It's not really jealousy.  It's me remembering that I'm one of the clumsiest, most unintentionally butter-fingered fools on God's green earth, and that if I trusted myself to anything that expensive in my hands for more than a month I'd end up regretting it.  And so far, I've been right.

(...which explains why the notebook I'm typing this on has a 2-year damage-protection warranty...)

I love technology.  I just know that any technology that I need to exercise undue caution while carrying around with me is probably a bad match for me.  To this day I still don't know how I managed to demolish my old notebook's display doing nothing more than putting it in a bag and walking up three flights of stairs without bumping into anything or getting broadsided.  (I blame elves.)

I also know and am grateful for the fact that on the whole things have gotten that much more rugged in the last couple of years.  The only reason I didn't shell out for one of the Toughbook notebooks was the killer pricetag -- otherwise, hey, a notebook that can survive everything from a rainstorm to squirting out of my fingers and flying across the room like someone smeared it with cooking oil for a prank?  Don't laugh; I've done that -- fortunately the notebook landed on my bed with a clearance of only a few inches, and I did not end up shelling out for a horrific repair job.

But on the whole, when I shop for personal electronics, I'm guided by one criterion more than anything else: How likely am I to mess this up?  If I buy a flip phone, is the hinge strong enough to survive me stupidly sitting down on it when I leave it open on my chair (something I've done not once but TWICE)?  And so on.

If I spend nothing on a phone save the cost of my contract and use it for four years, that's a better bargain to me than spending even $75 (or $150, or $200) and worrying that I've gotten stuck with something that will simply be unable to survive six months banging around in my particular pocket ... or falling into my particular bathtub.

Like everyone else, I've been following Microsoft's proclamations about how Linux and other open-source products violate a whole clutch of their patents ... but they won't say which ones.  At least not out loud.  But they won't sue anyone for it, so don't worry.

Does Microsoft enjoy sending mixed messages?  That's about the only coherent conclusion I can draw from all of this -- but then again, we live in an age where it is possible for a company to say "We are totally committed to X" on Monday and then on Tuesday turn around and claim their definition of "commitment" was wholly mutable, which is why they're not committed to X anymore.

Microsoft's Port 25 blog has a post about the patent foofaraw, which adds virtually nothing to the issue.  But one of the linked analyses is very revealing:

Microsoft probably can't sue many of the companies it needs to worry about most, because existing cross licenses with those companies would prevent it. Note that Microsoft hasn't said a word about its patents being infringed by AIX or Solaris, for example.   These cross licenses would presumably protect Linux distributions offered by the same vendors as well.

The agreements that Microsoft has already signed with customers and distributors may assign little, if any, value to the patents. It's possible that the value actually went in the other direction, with Microsoft paying more to get the other party to agree to include public mention of open source patent licensing at all.

So what's really going on here?  My guess is that MS is, not very competently, trying to defuse its image of being a patent bully while at the same time trying to make sure that anything they licensed from other people (which would explain why the whole thing is hush-hush) doesn't get subjected to the wrong sort of scrutiny.  I suspect one of the things MS is licensing from its open-source partners are things that could not be patented anyway, or which if they were patented would draw fire if those patents were subjected to scrutiny -- so why open that whole can of worms to begin with?

Or maybe they just want everyone to think they're dumber and clumsier than they really are.  That would be a comfort, wouldn't it?

Long before Outlook 2007's indexed search and contact manager, I had Eudora -- probably the first standalone email client I ever used apart from Outlook itself and CompuServe's unbelivably evil Windows client application.  I eventually outgrew Eudora (much to my chagrin), but one of the things that stayed with me as a matter of habit was the use of plaintext for email.

To this day, I'm still a plaintext snob when it comes to email, and it's one of those decisions I plan to stick with as long as it is humanly possible.  Many of the reasons I could quote are probably tiresomely familiar to you by now:

• Email was never meant to become a carrier for rich text and graphics (and spyware, and ads, and...)
• The less there is in any given email, the better.
• HTML is more of an eyesore for some people (like myself) than a prettifier.
• If you think the riot of ">" marks that you get from a multi-response email is bad, wait till you see how it looks in HTML -- where each respondent can potentially have a different font, quote delineator, text color, and so on.

The only concession I've made away from this stance is to use UTF-8 as my standard text encoding rather than ASCII.  This is mostly because I do find that I have to rely on some characters that aren't available in ASCII (for instance, when corresponding with people in non-English character sets), but other than that it's straight 8-bit for me.

In a weird way, this is part of why I've not been as aggravated by the rush to XML as many other people have (as a general data-storage format).  It may not be the prettiest thing in the world, but it's a heck of a lot less thorny at bottom than anything in a binary format, and it can be conquered with any number of free editing tools out there.

I suspect, under it all, there's a high degree of nostalgia at work.  8-bit text (and even UTF-8) is about as clean and uncluttered as you're likely to get in this world.

(While I'm at it, I should plug a tool that I've been using as my plaintext editor of choice: Notepad2.  It supports both ASCII and UTF-8 interchangeably, and has a slew of little features that make it just right for casual editing of text files.)

Judging from the punditry that's being offered up in various circles about various news items -- Dell reinstating XP preloads; one company offering Macs that dual boot between XP and OS X, etc. -- it's starting to sound like Vista is Microsoft's biggest flop since Bob.  Well, maybe not that big a flop, but not the roaring, raging, total smash hit that they hoped it would be?

I'm going to stand by something I've thought for a while now: The initial reception to XP was about as frosty as this was.  The words "memory hog" were bandied around, too.  (Folks it's been seven years since then -- can we please assume the basic hardware profile of a computer has gone up a bit since?  And that the cost of a PC that ran XP then is about on the same level as a PC that runs Vista now?)  I know people who rejected XP with their new PCs the way they now reject Vista, and talk about skipping out on Windows entirely to Linux or the Mac.

This last threat is perfectly valid, though.  When I reviewed Ubuntu Linux I found a system that was a lot closer to prime time than I expected.  I still don't think it's going to eclipse Windows as completely as many would like it to, but the fact of its existence will be a pressure point.  I don't think Microsoft is going to vanish, but I don't think they're going to enjoy the 90+% market share on the desktop that they typically have anymore.  And I think that's a more reasonable assessment than to assume that Vista is dead / dying (look, Netcraft confirms it!)

The patent issue is another story, though ... one I'll weigh in on separately, but which I don't like the sight of one bit.

Mike Neil, a Microsoft general manager, made a post over at the TechNet blogs about what will and will not be going into the next version of Windows Server's virtualization.  Among the things that are not being included this time around: live migration, resource hot-add functionality, and a support limit of 16 cores per physical machine.  As he puts it earlier in the article: "Shipping is a feature, too."  Better a good product now than a great one at some undefined time, and after who knows what kind of agony.  (One feature that people will be thrilled to know is in there, however: full support for VSS, so a guest machine can be backed up while running.  I'm wildly curious about how they implemented this: is a snapshot of the system state data also taken and backed up with the computer?)

The murderous gestation period for Vista was something many people pounced on as a sign that Microsoft's product lifecycle strategy was seriously out of hand.  I still think part of it was that Microsoft wanted to talk badly about the things that were going into Vista -- but at the same time, they weren't sure how much of that could make the final cut, so they would up talking about things that simply couldn't show up as discussed.  It's something of a trap: the way to build excitement about a product is to talk about it, but if you talk about it and then you don't ship what you talk about, you disappoint people.

Would Microsoft be better off if they adopted a more Apple-like approach, where new features are held so tightly under wraps that no one knows about them under mere weeks before launch?  I doubt it -- probably because they're serving a radically different market than Apple.  If they talk about something that's going into a future version of Exchange, for instance, they want to talk about it so that the people who make buying decisions about Exchange can weigh that against other factors (the cost of upgrading in terms of time and resources, and not just money, for one).  Apple's still not a terribly business-oriented company in the sense that the way they develop products and put them out into the market is meant to appeal specifically to workplaces.  This isn't to say that a company can't install a slew of Macs and do well (many do); it's just that Apple's products as a whole don't seem designed to cater as explicitly to business markets.  This is changing, I'll admit, but right now there are still plenty of signs it's not like that yet -- the whole iPhone-vs.-Windows-Mobile-devices debate is a good example.

So, while I can't decry Microsoft for talking about what they'd like to do, I wish they would cast their language about upcoming products a little more explicitly in that regard.  It would not make them seem any the less ambitious -- at least, not in my eyes.

One relatively undiscussed feature of Vista, which I've touched on here in the past, is the Reliability and Performance Monitor, a quick-and-dirty way to find out exactly what's using what resources in your machine.  A common complaint about Vista is that the hard drives are always churning, which is in turn often attributed to the indexing service.  Sometimes that is indeed the case, but what if it's some other piece of software, and you want to find out whether or not the indexer is really the culprit?

To launch the R&PM, just hit Start and type perf in the Search box; it should be one of the first things that comes up.  You'll need to say yes to a UAC prompt, and then you'll be presented with the R&PM main interface.  Click Disk to expand the disk-access panel, and look for SearchIndexer in that list (you can click on the "Image" column to sort by process name to speed things up).

In this particular example, I've sorted by the "Write" column (which is expressed in bytes per minute).  The SearchIndexer.exe process is indeed at the top of that list, but it's only written 500K or so in the past minute of time (which is a miniscule amount of disk activity).  The "Response Time" column is also a strong clue as to how much of a hog that particular process is: 10 milliseconds is relatively small.  So any disk-bashing that might be taking place wouldn't be the fault of the indexing service.

My way to determine what's bashing the disk is to sort by both Read and Write columns (and Response Time), and see what comes to the top of the list after a minute or so.  If the results are an application, then you can look into any possible problems with that app.

By now most of the other tech writers out there have linked to the story about PC World editor Harry McCracken resigning over what were characterized as "disagreements with the magazine's publisher regarding stories criticial of advertisers".  Most of what I've had to say about this has been an echo of many other people's statements, which I can sum up like so:

• When the ad people start calling the shots at a publication, to the detriment of the editorial, it isn't a secret for very long.  It usually doesn't require any degree of detective work on the part of the readers to spot it, either -- it's pretty self-evident.
• I've had my ear chewed more than a few times in the past by companes who felt I had published unfairly critical evaluations of their products.  Believe me: they get over it fast.  They know they can't live without the broadest possible coverage, and they never stay mad for long.  I've known of people who got fired from working for a given publication for violating an NDA, however, but no major software or hardware company can afford to hold grudges for too long.
• I strongly suspect there's more to this story than just the deep-sixing of an article that was mildly critical of Applie, and that it might have been the only last straw after a whole series of camel-breakers.  It takes quite a bit to separate a multi-year veteran of a given publication from his work -- like, for instance, a new boss whose sympathies for Apple are worn openly on the sleeve.

After my last screed about applet junk I received an email from reader Dean Adams, who has created a batch script that performs cleanup on a QuickTime installation by removing everything that doesn't absolutely have to be there.  It's interesting enough that I thought I would share it here, although be warned that it is supplied AS-IS with no warranty.