Rupert Goodwins' Diary

Monday 17/7/2006

We get a lot of phone calls that go something like this: "Hi! I'm from Pushit PR! Did you know that 70 percent of UK companies have no trained macaws on their premises, even though they are very good at configuring firewalls? It's a major security oversight. There's a survey from Markit Research Inc of 300 companies which shows this."

"Really? Who paid for the research?"

"Er, the Macaw Breeders Association of Great Britain..."

Ding. Thanks, but no thanks. Now and again there's a good story attached, but by and large you get what you pay for, whether it's market research or analyst reports.

Things are a bit more complex with BT's claims to be the number one ISP, courtesy of a report from Epitiro, a company that does Internet performance monitoring. Although BT is a client and therefore does have a commercial relationship with Epitiro, the same's true of many other ISPs — and the figures on which BT bases its claim are from Epitiro's central monitoring system on which it bases its business. So the figures have a good chance of being correct.

But BT's not commonly on anyone's list of favourite ISPs. Everyone — and sometimes, I think that really does mean everyone — has a billing, customer-service or fault story connected with the company, and that's been the way for ages. It's particularly fun, I'm told, to try and move your DSL connection when you move house, and to synchronise the process with the moving of the phone service. The sheer number of things that can and do go wrong, most of which seem to be connected with some world-class cluelessness by BT salespeople, is impressive. Things like changing the name of the owner of a phone-line trigger a broadband disconnection, or an offer to "sort out the broadband" quietly moves you from your current ISP to BT-Yahoo.

Another friend of mine always checks his phone bill, and for around two years has found at least one error per quarter. It's got to the point that they just wearily wave the refund through when he makes his quarterly call. Would he recommend BT to anyone? You get a hollow laugh if you ask.

It doesn't matter whether BT Internet does best on speed tests. It does matter if they're a pleasant company to deal with, that knows what it's doing and provides a reliable service. The core tech might be fine, but if it takes forever to get it installed or a fault fixed that's no help.

As a result, BT will find that its claims to be super-duper falls on stony ground. We just don't believe it, not because we think that BT bought the result it wanted but because the information is almost irrelevant. Everyone knows better. Even the macaws.

Tuesday 18/7/2006

There's a wireless band coming up for consideration; it's allocated to 3G overflow, but the regulators are wondering whether to repurpose it for more general-purpose use. Unsurprisingly, the GSM Association isn't keen.

The argument is simply stated: the mobile phone companies do not want competition from new technologies. As there's no band available at the moment for a mobile phone competitor, the incumbents have a free run — but there will be some more frequencies coming up for reuse shortly. If those bands are 'neutral' — not tied to any particular technology — then new ideas and potentially disruptive products can appear; as with Wi-Fi; if the band is licensed with strict standards for its users, as GSM and 3G, then there are fewer surprises.

The GSM operators' concerns aren't stated as such — they're just worried about a fragmented market keeping prices high and therefore not serving customers well. Bless. They do have a point, though; the very regulated approach produced a stunning success in the GSM worldwide network — but then, the hands-off approach produced a stunning success with the chain of WiFi standards.

Regulated bands are troublesome when the service to which they're allocated is a commercial failure. Then you end up with acres of underused spectrum — if you're really unlucky, with a few users who can't be moved off to a different service They're also bad for efficiency as they encourage the continuation of old technology, which is why the switch over to digital TV is taking the best part of a decade.

What the regulators and network operators know, though, is that in time all this will stop mattering. Software defined or cognitive radio renders the question moot: when you have a common circuit that can modify itself in software to any radio specification, then brand new standards can be introduced at any time and with minimal disruption. It's even proposed that bunches of cognitive radios can evolve their own standards by monitoring band conditions, maximising throughput without interfering with other signals, and pass each other code fragments to more efficiently implement their decisions.

There are a lot of similarities between that approach and what actually happens on the wired Internet. New protocols and services are constantly being invented and distributed — Skype took practically no time to get onto millions of computers, for example — with no central regulating or controlling authority. It works very well, and wireless regulators are consciously trying to find ways to make the radio spectrum more like the Internet — as the regulators themselves admit, they're not the world's most accurate predictors of the future, so the best deal they can offer is frequently just to get out of the way.

The question is: does the freewheeling, hyperevolving Internet work because it is so free, or because everyone on it is using an agreed, non-mutable, rather elderly standard called IP? Both, of course, so perhaps the most useful way to regulate the radio is just to posit the basic protocols by which radios can sniff each other's bottoms before deciding how to talk to each other. That's work to be done.

Bear in mind, though, that while we get there, there'll be some ferocious rearguard action from the established operators. Just see it for what it is.

Wednesday 19/7/2006

There is a worried look in many an IT support person's eyes today. Microsoft has bought Winternals, the company behind the Sysinternals free utilities on which many people depend. Will these utilities continue to be developed? For now, says Mark Russinovich, programming folk hero and one of the two men behind the company. But then? The FAQ that Microsoft has posted regarding the acquisition doesn't mention this most frequently asked question of all; I suspect I'm not alone in grabbing everything I can from the site before the living hell of Windows Genuine Advantage gets going. One tool in particular, TCPView, is in daily use here just to keep an eye on what's making contact with what over the Web.

It's good news for Russinovich, of course, who, for helping millions of people sort out the hideously under-documented internals of Windows, deserves to have a statue erected. However much he got for the move, it wasn't enough. But we're losing a very valuable independent mind — Russinovich was the person who analysed and broke the Sony rootkit story. Although he and Microsoft have always been friendly, he was more than able to say when the company had got things wrong and to offer ideas to work around the problems. When Vista comes out, it'll be riddled with all manner of "security"-related features, through which Microsoft maintains or increases its monitoring and control of computers and the software that runs on them, and one of the people best equipped to watch such things on our behalf will now have his hands tied. That's bad news for us.

And why now? Again, it has to be connected with Vista, currently slouching towards its Christmastide release. Microsoft has always been poor to the point of lackadaisical in the provision of low-level support tools for its operating system: when things go wrong, the chances of working out what happened using just the utilities provided are slim indeed, no matter how knowledgeable you are about what goes on under the bonnet.

If Vista is going to come with a proper set of administrative diagnostic tools, then that's very good — but it's rather late in the day to be thinking of that. Unless, of course, Russinovich and co have been developing this already and Microsoft liked what it saw so much it decided to buy the company. The other thought is that there's some low-level Vista madness that's in need of genius healing, and the job lot was too tempting to pass up.

Whatever the reason and whatever the outcome, a little bit of diversity has been lost from the Windows technology world. Here's hoping that the dreaded effect Microsoft normally has on acquired companies and their products — rendering both invisible — holds off on this one. We can't afford it.

Thursday 20/7/2006

I don't know how they do it. Apple sold eight million iPods last quarter, into a market I was sure would be showing signs of saturation already. And this without having much new to flog — where is the iPhone (on hold, apparently) or the wireless iPod? It is good to see that they're thinking about making the iPod speak the track selection, an addition I've long wanted. When you have 8,000-odd tracks on shuffle, there's a lot to remember.

But even better news is that Mac sales are up 12 percent. The transition to a new processor is always a tricky business, with the chance that customers will wait and see how well things work and that developers will not want to make the various investments necessary to keep up. It's even trickier when you're transitioning to the same architecture as your major rival, enabling all sorts of curious cross-fertilisations that you may not really want but might be out of your control.

Apple's pulled it off, with only the occasional blast from the tuba of stupidity. Leather cases, anyone? It's still making very desirable equipment, maintaining its differentiation and keeping its army of rabid fans happily on-side — even if some are publicly defecting to Ubuntu. It might seem churlish, when faced with such success in such a competitive market, to accuse the company of coasting, but we have high expectations of continual Apple cleverness and a rather prejudiced wish for the company to succeed. I want to see a proper digital home strategy — I couldn't care less about that normally, but Apple has the gumption to make it work so well even curmudgeons like me will want a part of it. I want to see a decent attempt at enterprise computing. I want to see Apple buy Moog and produce the world's gnarliest synthesiser. With sparking electrodes.

Sorry, sorry, got carried away there. But enterprise would be good: I can't make any case for work to buy me a Mac, and while I can deploy Linux on my own and integrate it with the work systems I'm not going to shell out tons of dosh to do the same experiment with Macintosh. Much as I'd like to. Get it sorted, Steve.

Oh, yes, Ubuntu. Ups and downs this week: Gaim has stopped letting me start group conferences on Yahoo IM, which is annoying, and twice now the machine has descended into massive disk thrashing and glacially slow response. The first time I had to hit the big red switch: the second time I managed to pull up a text console despite the GUI being effectively dead, find the problem (the clock applet had decided to use 1.3GB of virtual memory, which I suspect is not entirely how it is supposed to work) and recover from it. So I'm getting there — when I finally understand the file system, I'll be happy.

Doesn't take much when you're a geek, really.

Friday 21/7/2006

On Tuesday next, the UK IT website scene will be joined by IT Pro, a newcomer from Dennis Publishing. We're all keen to see what it looks like: Dennis tends to know what it's doing, but that's OK — so do we. A bit of competition is nothing to be scared of. Grows the sector, and all that.

That's not necessarily the universal opinion. IT Pro is staffed by escapees from VNU, most notably editor Chris Green who used to be technical editor for Computing magazine before fleeing into the arms of Felix. This got written up in tech PR tip sheet The Fullrunner, which mentioned Mr Green's previous work with the cloggies at VNU — at one point using the word 'substantial'. Good lord, no, said the new editor of Computing's Web site. Nothing of the sort. In fact, so insubstantial was Green's input that we're not even going to bother to replace him.

Catty or what? The saga deepens, though — callers to the Computing offices trying to get hold of Green are being told "he's left the industry", which seems a bit harsh when we haven't even seen the Web site yet. And in any case, it's easy to check what he's up to because there's a Wikipedia page about him.

Ah yes, the good old Wiki. If you dig back through the editing history of that page, you'll see that some anonymous people have been very rude about the chap, impugning his honesty and comparing him to an anatomical item possessed by the gender which rarely plays rugby. Other edits make light of his career history, suggesting that he may have spent more time in retail ambient replenishment (that's shelf-stacking in supermarkets to you and me) than at the keyboard.

Shocking, and of course all the silliness was smartly removed. Wikipedia is a place that believes in keeping track of the truth, though, so it keeps a copy of the IP address from which the edit was made — not that this is always helpful. Usually any vandalism takes place from Internet cafes, via anonymising proxies or through some of the many other options for disguising the naughty. If one were to muck about with another website — to leave satirical blog entries, say, that take the mickey out of a publisher's online strategy — one would not want to be traced too easily.

This basic precaution escaped our cyber-graffiti squad. Type their IP address 62.140.213.2 into a network utility such as traceroute or reverse DNS lookup and just look at what comes clogging back.

Bit careless, chaps. Perhaps you should reconsider filling that technical editor position.