Guy Kewney's Diary

See, I live by copyright. If everybody on the planet can re-broadcast my words, then how can I charge advertisers and readers for them? And here I am, on the Web. So logically, I must be in favour of copyright, right?

Well, up to a point, yes. But the world is the way it is; if I want it to become the way I want it to be, then I have to change it. And before I change it, I have to know what is actually going on. Which involves some awkward truths.

First amongst these awkward truths is "the hazard of using the sidewalk" as they say in the US. You can be so beautiful that your face is exclusively assigned to a make-up company, and nobody else can use your face in advertisements: but walk out into the street, and anybody can take a picture of that face and use it in a news story. They can even put it on the front cover of their magazine.

Lawyers say that by appearing on the pavement, you are accepting the hazards that go with it; and being photographed, interviewed, filmed or shot at are all hazards of walking the streets.

The same, probably, applies to the Internet. The Web, and HTML, is set up to allow linking. To say: "He linked to my site without my permission" is like complaining to the police that "she looked at me when I went outside." It's an obvious, accepted hazard of putting information on the Web.

Totalnews has settled. Some idiots appear to think this means that the legal aspects are settled, too. Not yet, they aren't. I have to spend most of the day talking to QCs about copyright; they agree with me.

Tuesday

It's Rockwell's turn. Do you want faster Web access? (pause for chorus of approval) - well, of course. Will a 56kbps modem give you faster access? "Of course!" screams the mob of modem builders who buy their modem chips from Rockwell.

Hmm. Of course, this is catch-up time. US Robotics may have announced its X2 technology slightly after the K56 design was announced, but it hit the headlines far, far harder. And it lined up all the big UK Internet providers.

Hitting back, hard, Rockwell shows its own assets: it owns the standards body. This takes some background. What both sides are doing is to feed digital data into the phone line at the exchange, even though it is an analogue line. The exchange line takes 64 kilobits worth; some bits are used for signalling and billing; 56 K is about right as the top speed. You need a special modem to receive these digital signals.

And of course, you don't want USR X2 modems to do this. No, you want Motorola, or Hayes, or...

To make this point, the corporation hires a hotel in Piccadilly, and gets some contractor outside to dig a hole through the mains power cable. The Big Return March is conducted by candle-light, which is at least an improvement on PowerPoint presentations, though not as good as an insomnia cure, alas. And it then gets Motorola and Ascend and Psion Dacom and Diamond and so on, to stand up and say how much they approve of the Rockwell K56Flex standard, and how they've tested it, and it works really, really well.

And the important thing is that there will be a standard: it will be agreed by the ITU Study Group 16, sometime in October. By the way, an interesting coincidence: all the members of SG-16 are on the K56flex consortium..."We're not saying that the ITU standard will be K56flex. But..."

We ask whether the tests include tests on USR X2 technology. They say that would be unfair. Instead of being unfair, they say, we should look at throughput. Connect speeds aren't important; getting "CONNECT 56000" is a joke if you drop half the packets.

True. Are they saying that USR modems drop half the packets?

"I have some data here from PC Week, independent tests. It says that all K56flex modems performed about the same," says the chap from Diamond. "Oh, dear, looks like my slides haven't been saved properly."

On return to the office, I check the PC Week site. Sure enough, he's right. All K56flex modems are about the same - as are the figures for USR X2. Odd. That wasn't the impression the Diamond guy gave at all.

Wednesday

Breakfast with IBM, talking about its new ThinkPads. Should have been with Netscape, but couldn't get there (had to take the car to be serviced. Naturally, it started raining immediately I left the umbrella in the car).

It seems to me that IBM has rather been caught, flat-footed, by the notebook scandal. While Toshiba is rushing out sub and sub-sub notebooks, IBM's are all huge, heavy, and conventional. We ask when the first IBM ThinkPad will be shipped with Universal Serial Bus -not till October.

This is probably a sad situation, given that Compaq has been busily making life easy for its competitors. The 4100 model Compaq laptop has had keyboard problems. Gartner Group reckons the "dead on arrival" figures for some Compaq machines (portables only) has been as high as 25 per cent - a figure Compaq says is "ridiculous". But Gartner has told its subscribers not to buy the Compaqs.

This is not, in fact, news. Corporate buyers have been belly-aching about Compaq's unresponsiveness for months; Dell's Latitude machines are selling in huge numbers, and the numbers are growing fast; and some people have literally taken consignments of Compaq notebooks and dumped them - physically - on the doorsteps of their suppliers. Cisco's engineering sales force has now switched, I gather, entirely, to Toshiba.

Dripping wet, but well fed, I fail to make it to the Voice-Type launch of a continuous speech product.

The evening is spent in the company of programmers. We should probably not penetrate that veil; there might be children reading.

Thursday

Turns out that the IBM voice recognition product was rather as I suspected: better than the old one, but inclined to make baffling misinterpretations of what people say. Trouble is: when you get a spelling mistake into something you type, it's going to be picked up by the spelling checker program. But voice-recognition programs make correctly-spelled mistakes.

Also, you need a *minimum* of a 166MHz Pentium, plus 32 meg of ram. Dragon Dictate is similar in requirements for Naturally Speaking. As an expensive way of producing baffling gibberish, I feel it lacks the fluency of typing...

The fun part of the day is the bit where Gartner says it has "benchmarked" the IT excellence of several corporations, and the best -- measured by objective standpoints - is?

[FX: fanfare!]

National Westminster Bank.

Credibility of objective benchmarks instantly wiped out, as several ex-NatWest customers in the office roll around on the carpet laughing. Dear me, that was a good one. Yes, I'm one of those ex-NatWest customers...

Friday

Michael Dell, on his way through London to New York, and back to Austin, Texas, stops for lunch. A nice, intimate meal, with four of his top UK execs, and a half dozen other hax. He says that a "ManagedPC" is the correct answer to the people who want a Network Computer with no floppy. He disappears to go and talk to NatWest.

I did try to warn him.

Microsoft says that you should not get Netscape Navigator.

It asks: "Do you want to pay $59 to upgrade your browser not once (to get Communicator alone) but twice (when Netcaster ships) when the free and more comprehensive Internet Explorer 4.0 is just around the corner?"

My fellow Fellow, Peter Jackson, casts his mind back a decade. "Who was that small California company which started advertising low-cost long-distance? Remember how AT&T went to all its potential customers and said: 'Don't buy that, wait for ours, which will be much better' and then never provided the low-cost option?"

Tip to Microsoft: go check the documents. I don't remember the exact details. Was it $1.2 billion dollars AT&T had to pay when the other company went bust, and the directors sued?

Meanwhile, the "just around the corner" clock is ticking. It is Friday June 13.

Peter Jackson snorts. He is, he tells me, off to Seattle for a combined Internet Explorer 4.0 and Windows NT 5.0 briefing. "Only I've just discovered the NT part of the presentation has been cancelled."

Goodness, gracious. We can only speculate why. Here's a hint: a programmer phones me, saying: "I had a problem with Access programming. Tech support at Microsoft said to send them the file. I did. They said it got corrupted. I sent it again. It got corrupted again."

So the programmer rang: "What's wrong with my file?"

Tech support: "It's all rubbish, just ASCII characters."

Prgrammer: "That's not corrupt! That's a UUencoded file! You have to UUdecode it!"

Tech support: "What's that? Why can't you just send it as an attachment?"

A nice man from BackWeb rings to tell me about "push" technology. His stuff is included, he says, in Internet Explorer 4.0, "to be launched in summer".

I'm amazed. "You're expecting IE 4.0 this summer?"

"I didn't say which summer, did I?"

Monday looks like being fun. I've now seen the Psion Series 5. It's nice, it has infrared links (just like Siena and Series 3C models) and if you buy one, you can't send data to owners of Siena or Series 3C machines.

The infra-red transceiver is identical. It's just that the link layer is different. Oh, and the file structures, too; the new Series 5 machines don't swap data with Series 3. Of course, you can import them to a PC, save them as plain text or worksheets, and then export them to the other system.

Apparently, this doesn't matter. "We're aiming the Series 5 at a different sort of market from the Series 3," they say.

Oh, right...