Rupert Goodwins' Diary

A week topped and tailed by Microsoft blunders, with a tasty filling of Viking porn, viral vicars and vapid VoIP. Get stuck in!
Written by Rupert Goodwins, Contributor

Monday 31/07/2006

You've seen the YouTube clip. You've read the commentary. You've even bought the T-shirt. And your aunt has laughed her head off. Microsoft, never a company lacking in hubris, decided to mock the gods of demo and committed the number one sin — showing off speech recognition to an important audience.

Lesson one: never show off speech recognition to an important audience. It never works. This has been an accepted fact of life since 1984, when Apricot launched its stylish, innovative and futile Portable to a chortling world. That had speech recognition: it didn't work. Microsoft shows no sign of believing this lesson, as it says it's fixed the bugs and will be showing off Vista speech recognition again soon. It may even be right — but a hundred flawless exhibitions of virtuoso word capture won't undo the damage.

Lesson two: be honest about what happened. Ten out of 10 for Microsoft here, as the developers in charge of the speech components swiftly coughed to the problem (at least, we think they coughed. Vista reported the sound as "Rough dogs delete my aunt"). Rob Chambers and Larry Osterman dish the dirt: part of the sound system cranked up the gain on the microphone during the silence between words to the point where the words themselves distorted. There was code to prevent that happening, but sometimes it didn't work.

There's still a good question as to why they chose to demonstrate a system with a known intermittent bug — especially since a later build was available with the problem fixed — but I've been there, I've convinced myself that because something worked three times in a row in private it'd be fine that one more time in public. I've enjoyed the full facial albumen. They have my sympathy.

Lesson three: have a word with your PR company. While Microsoft has learned lesson two, its PR company, Waggener Edstrom, is still living in the past. The TV presenter on the video clip was happy to report that Microsoft "was not happy we showed you that video, and blames ambient noise for the failure. But as you can hear, it was quiet until the product didn't work and everyone started laughing. Live television is rough. Welcome to our world". Ho ho. But of course, it wasn't Microsoft being unhappy, it was WaggEd trying to spin the story. We've been on the end of that spin machine, and it's counter-productive. Perhaps it used to work with the non-technical media, but no more — after all this time, the huge gap between the neo-Stalinist stance of Microsoft Official Reality and what people can see for themselves has percolated into the noodle of the least critical observer.

Microsoft knows this in its heart. That's why is why Rob Chambers and Larry Osterman are free to tell the truth and in so doing help to heal the decades of disdain in which the company has held its customers. Even Steve Ballmer feels able to say where the Vista development effort went wrong. WaggEd, bless it, is still flying over the jungle blasting propaganda from its helicopters. It looks foolish. It makes its client look foolish. It's out of touch.

You don't want that in your PR company. Perhaps someone in Microsoft could recognise this.


Tuesday 1/08/2006

Curious goings-on at Amazon. The company is getting worried that its rate of growth is slowing, so it's been investing in lots of things such as free delivery schemes, a bigger range of products including groceries, and fixing problems caused by erstwhile partners pulling out. This involves spending lots of money for gains that will take a while to come through, so naturally enough the analysts are seeing this as a terrible state of affairs.

So far, so normal. But there's been some collateral damage. Last year, Amazon bought a company called mobipocket, which publishes e-books for PDAs. Fair enough: the company needs an e-book strategy, even if the medium is currently commercially insignificant. But now it seems as if the company is losing all non-Amazon published e-books from its lists. It's also kicking off the saucy stuff, with mobipocket announcing that it would no longer carry "pornography" and, with immediate effect, it was withdrawing all content that it deemed porn.

Which is curious, as e-book porn was the only success story in the medium. To be more accurate, slashfic was the success — and the purveyors and consumers of slashfic are foaming at the mouth at having their one commercial channel cut off.

I won't dwell overly on slashfic. It's quasi-legal porn written by straight women for straight women, all about alpha males having gay sex: the quasi-legal side is that a lot of these alpha males are famous celebrities and thus have disturbingly non-fictional lawyers to help them protect their virtue. This makes it hard (yes, thank you at the back) to sell the stuff through normal channels. Self-publishing is fine, but marketing and commercials are hard. Amazon e-books take care of both: you're in.

No longer. It seems odd that Amazon/mobipocket should choose to complain about the pornographic nature of the content, rather than the intellectual property issues. There's plenty of non-infringing slashfic ("How am I going to get my Viking porn now?" asks one outraged afficionado), and Amazon itself has no problems selling sex toys and the sort of mainstream fiction that qualifies as erotica chiefly because there are no naked people on the front cover. As far as I know, the issue of e-book porn hasn't been raised by even the most rabid right-wing fundamentalist family-obsessed anti-fun lobby. And while slashfic is something of a specialised taste, I've long been of the opinion that it's a vast underexploited market that could rapidly rise to commercial prominence if somebody would just take a stand (that's quite enough of that).

So why cut off the only source of revenue for e-books, just as pornography is becoming more acceptable in the mainstream than ever? Is this another manifestation of the peculiar American puritanism that expresses itself in zillion-dollar fines for a second of TV nipple, or some peculiar kind of corporate homophobia?

Perhaps one of the slashfic ladies wrote up a Bezos/Gates fantasy. That would make as much sense as anything. But no, thanks, I don't need to see the evidence.


Wednesday 2/08/2008

Vonage is also investing heavily in marketing — and also getting it in the neck from the market — although as far as I know the company has no stance on teh ghey pr0n. In this case, though, it's not clear how the company will turn the corner and establish itself as a permanent, profitable service provider.

The same's true of all public VoIP telephone companies. I use Skype, and darn useful it is too — but although I've used most of the services from time to time I doubt the company's had 20 euros out of me per year. The major reason they don't get more is that I'm unconvinced of the long-term financial viability of the service and I'm not going to rely on a proprietary system that doesn't give me the freedom to move number and service provider. It might seem contradictory, but I won't stay put unless I can move. There's no reason why my telephone service shouldn't be like a Web site — able to move hosts or even be self-hosted — with a single unique identifier that moves with it.

Such standards exist, but even open source VoIP systems such as Project Gizmo are built around the idea that for incoming calls you buy a number from them that's more theirs than yours. You can look through the FAQs as long as you like, but the one Q you might expect to be FA'd — "If I have to change providers, do I keep my number?" — is never addressed. I've used Project Gizmo. I've used Google Talk. I've used Yahoo IM voice calling: none wants to admit I might not be theirs exclusively and forever. That's not a healthy way to run a relationship, outside marriage.

It's true that telcos still tend towards such ideas, preferring to think of you as their property. You still can't move your home BT number when you change location — even if you're sticking with BT. It's not convenient for them. And while you can move your mobile number between providers in the UK it's amazing the resistance you can encounter when trying to exercise what is now a legal right.

The Internet way is a much better way. The first VoIP provider who'll deliver a package of services which includes an unambiguous, documented method for leaving them while keeping my number, will get my business. Of course, that means they'll have to keep working to stay ahead of the pack, so any such company will have to be innovative, smart and agile — but what else should I expect?

VoIP, ENUM and number portability. When that's fixed, we'll talk — and when we talk, VoIP providers will make money.


Thursday 3/08/2006

Of all the deep, deep joys of IT journalism, none compares to a good old-fashioned vicar story. And that's what we have today, following a complaint from Church House Publishing — the propaganda arm of the Church of England — that they weren't getting no satisfaction from Symantec. Norton Antivirus was falsely identifying a component of their Virtual Liturgy package as a virus, and suggesting that the users delete it: they did, and subsequently found that the package didn't work. So they complained to the publishers at Church House, who in turn complained to Symantec, who in turn... didn't seem to do anything.

We got hold of Symantec, and the problem was fixed. Actually, Symantec says that the problem had been fixed last month and the publishers hadn't bothered to get in contact to find out. The publishers say that they couldn't get through to anyone at Symantec no matter how hard they tried.

We can't easily find out whose version of events is accurate, but it does highlight one problem in AV software: the people who end up doing the support are rarely those who can fix things. In this case Church House Publishing ended up fielding all the calls from frustrated clerics, for a mistake made by Symantec. Reason says that Symantec should take on that burden — even if just by giving CHP an email address or URL to distribute to affected parties — but when the company seemed unable even to tell the publishers that the darn problem had been fixed, it's clearly not going to manage anything like an acceptable level of user service. I wonder whether AV companies are worth the hassle. One of the unexpected benefits of switching to Ubuntu on my elderly laptop has been that stuff runs a lot faster; not necessarily because Linux is more efficient than Windows (although it is), but because there's no fat old, cycle-stealing, interface-clogging AV software imperiously squatting on top of my stack. What do you get for your money? False positives, their own security problems, incompatibilities and ineffectiveness — modern malware is tested to make sure it gets past the big four before release.

But is that enough to go open source? I think it is. One of the big no-nos about open source that the commercial boys push about is that you don't get support. What sort of support did the vicars get for their money? The community had to provide: it did. If we're going to have to support ourselves, then we might as well support the sort of software that gives back to the community as much as it gets. I don't like the feeling that we have to pay as well. Another no-no about open source is that it doesn't have any AV software. Guess what? If there's a need for it, it will evolve — and because it will evolve in an open environment with no secrets, it will avoid all the nonsense that happens because Symantec has to double-guess Windows and its applications.

Closed software is dangerous software: it cannot evolve, nor can it contribute to the evolution of its environment, and AV software is the best example of this. We know evolution works better than design in producing robust, flexible, working solutions to problems. Time to act on that knowledge.


Friday 4/08/2006

Meanwhile, Microsoft's efforts to become a player in the online services game stumble on, with a relaunch of MSN Spaces as Windows Live Spaces having plenty of teething problems. This proves a number of things — firstly, nobody has any idea what MSN is for any more, any more than what the Live strategy is. Second, Microsoft hasn't twigged that the Windows brand is not what you want in an online service: in fact, its Windows addiction is stronger than ever. IE7 has now been renamed Windows Internet Explorer 7 For Windows XP and Windows Internet Explorer 7 On Windows Vista. But then, Microsoft can't do names. It couldn't name a cruise ship if it was given a bottle of champagne on a ribbon.

All of which is a bit sad, because one of the side effects of the anti-viral vicars has been to provide astonishing proof that public perception has moved on from software being something you run on a computer. It's now all Web sites. As journalist Andrew Brown bloggishly points out, when The Times gets the story quite so wrong, the only explanation is that they now think that computing is the Web, and the Web is computing.

Does it matter that Microsoft wants to make Windows the default brand for its web services? It does. It's not that the stuff won't run on non-Windows platforms — even Microsoft's learned that one. Nor is it that people will see Windows and unconsciously stay away: in theory, there's nothing wrong with using the brand awareness they have on their desktops to keep them signed up when they go online.

It's all a matter of community. Lots of smart people don't use Windows. They won't hang out in places with Windows dressing. They know that supporting alternatives is important, and that there's no point in lending their online moxie to something they don't like. That's why you want to make sure your site is Firefox-happy: it's not that you'll be shutting out 10 to 20 percent of your potential audience, but that's exactly the 10 to 20 percent that you want. Creating an online community that will discourage that sort of user isn't the smartest thing to do. They'll go elsewhere, and they'll be followed.

All of which adds up to a company dragging a long way behind the online curve. And it wants so badly to be cool.

Want cool? Monday. Apple. No nonsense. Think they'll be announcing a me-too MySpace? No chance.


Editorial standards