All posts made in Apr. 2000:

The new Ikea ads are among the best adverts I have ever seen. Watch them.

Beautiful. Nothing finished bar the guestbook, but still utterly beautiful. [via Bud]

Microsoft's new TV ad reminds me of every party political I've ever seen. It's been focus-grouped to death (but then again, because it's low-key perhaps it'll work).

Updated the XML syndication thing in response to various comments. If you can't be arsed to read it, proposal summary follows:
There exist a bunch of hosts each knowing the address of all the others such that they make a network. Node A introduces a new piece of data, identified by origin URL and a timestamp. A message is sent out to other hosts announcing this data, spread virally (that is, all hosts that receive this message send it out to hosts they know). Each message (because the message has multiplied now) has a path back to whoever holds the data.
Node B wants the data. It sends a request back along the path to host A which dispatches the data. Every node along the path keeps a copy of the data for some amount of time. Should it be on the path of another request, it doesn't propagate the request further, it just returns the data.
An alternative method would be to broadcast all data over the network, each node sending out new data as it receives it. Whether the node hangs on to the data for the local users is up to it. This would be easier I think.
The explanation above is a hell of a lot clearer than the document I wrote, but that's because it doesn't include all my fumbling and rationale.

I thought the world deserved to see the disgusting poster adorning my flat's door (photo stolen from

Hey, my syndication thing got a mention on Haddock! Now I wish I'd given it a better title; Haddock's 'XML and syndication and all that bollocks, and how to do it better' is so much better.
So, have I had any feedback? Yes, and some of it makes sense even to me. Nifty ideas: broadcast data through a network, each node only keeping the data if it wants it. By default no storage. This could be really cool. I like the idea of the Gnutella protocol but I think the level of traffic would be intense. Besides, there are different priorities here. I need to think some more.

Why does the phrase 'good beach guide' make me think of Punternet?

I never got into the whole MUD or chatroom thing, but even so Cycosmos looks neat. You get to chose your avatar and anything. Okay so it's hardly The Lawnmower Man, but if it's either shitty interfaces or having a rogue AI take over the world then what would you chose?

Good content at Just read everything. I don't find enough sites like this; I think I'm on the wrong hub.

This Elian satire goes a bit too far.

The guys who aren't releasing Mulholland Drive (David Lynch) are guilty of crimes against humanity.

I'm not quite sure how anything written is Visual Basic can be complex enough to be artificial intelligence, but the Electronic Judge dishing out justice on the streets is pretty cool (and at least its not a jury who reached a decision by flipping a coin).

Right, I've been thinking about the syndication of XML/RSS issue and I'm going to propose a new syndication system. Please read, send me comments &c., and please understand that it's only that long because I'm trying to sort it out in my head too.

Star Wars, with Lego actors - and believe me this link is so good I am officially allowed to talk shit for the next two months (I think that's the going rate for good links). Oh, if you can't download it email me and I'll try to help out with my own copy.

Fatal Facts: 'The cases here described were selected as being representative of fatalities caused by improper work practices.'

I still can't believe that really exists.

Okay, this xml and rss shit is all very well but it's not going to work. Say there's a live rss of headlines, updating every few minutes, that I want to feed to this page. Obviously I don't want to request the RSS every time I get a pageview, so I'll cache it every five minutes, say. But then what if everyone's doing it? The idea of rss is to easily syndicate content, but as it stands it's simply not scalable.
What we need is a distributed cache. Um. I'm still figuring out how that should work (push, I reckon, and a hierarchic flow. But there needs to be information flow upstream too, which is where I'm stuck).

Balls! I forgot my juicer from home so I can't get vitamins, and I forgot Becquerel (my spider plant) so I'm going to have loads of rogue radiation in my room. I'm going to turn into a fatty mutant lump. Bah.

StartsHere have 133 RSS channels. Many channels are generated and unavailable elsewhere, and all channels are open.

Weblog XML feeds listed at

"RSS and You" (RSS/Perl article) at

10 most recently updated Blogger blogs [RSS] found via xmltree.

XML newsfeeds from

For ages I've read riothero as 'raIOT@ro' [machine-readable phonetics], but actually it's just 'riot hero'. Oh.

Hey, you there, you dotcom people: You're paltry and transitory, you hear? Yeah, the Archibishop of Canterbury says so. Not me though. I'm dotorg and that's fine.

Bollocks. I'm back at university. Feel for me.

From February: A man dies in a cross-country crucifix-carrying race to find "Jesus Warriors".

Um - a reconstruction of the crucifixion in Wales.

Rejoice. T'Pau have returned. I saw them a couple of years ago, and man they haven't even begun to leave the 1980s.

No wonder Rover are having difficulties when they produce adverts like this leaflet which at best could be described as Giegeresque and at worst -- well, I wouldn't like to say.

If you've read the book; Dirk Gentley's I-Ching Calculator [thanks Daniel Milnes].

Nick Sweeney ('Not A Lawyer') manages to explain all of yesterdays comments about Godfrey vs. Demon in a single paragraph.
1996 Defamation Act, section 1: That's the only defence that Demon could even hope to use, and the judge threw it out, which is why they had to settle before amassing any more costs. Basically, you become a "publisher" from the moment you're told it's there, and you're expected to take action to deal with it. It would take new statute law explicitly defining an ISP as a carrier to change this (or, ironically, a favourable judgement after the RIP Bill becomes law, given that it would classify an ISP as a "public telecommunication system").

The original photo of the Earth from space, the one that everyone's seen, is a powerful image. Now NASA's brought it up to date with Blue Marble 2000 which just isn't as impressive. It's a computer model for a start: Artifically colourised; shadows computer added (with magnified vertical relief); Moon added. It's not the pixels, it's what's behind the pixels that matters. Oh well. It's only to be expected - their website is just one huge imagemap and the content is in PDF.

Genuine online petshop or thinly veiled BDSM store?

I am not a lawyer. In fact, the most I know about unnecessarily long legal documents is that all the nested <ol>s make them a pain to mark up - but I'm going to give this a go none-the-less.
The 1996 Defamation Act gives the definition of the publisher and possible defences. I believe the relevent passages are in Section 1, Responsibility for Publication (which the judge threw out; see later in this post). I'm not going to reproduce lengthy passages here, but they're not very hard to read. I'm going to make comments based on my understanding of Godfrey vs. Demon Internet.
1.1b, c: Grounds for defence If an ISP has been warned about defamatory content and has power to delete it (Demon was warned and does have that power) then they must prove they are not the publisher.
1.3c, e: Exceptions to the definition of a publisher These statements hinge on the fact that its the message that matters, not the media, and that a communications provider has no control over the person making the defamatory statement. I'm not sure why (c) is not a defence because it seems to me that Demon are effectively duplicating the message (in the Usenet cache). But: Since ISPs want to be carriers they want to be covered by (e) and claim it is not possible to control transmissions - unfortunately it is because they have a cache. For (c) to be applicable though you'd have to agree that a cache is an electronic medium (does an electronic medium have to be discrete?), and also that it wasn't the ISP who were providing the communications network by which the contents of the cache were being retrieved by the end user, it was the end user's telco.
However, as much as I consider the Godfrey vs. Demon case it really doesn't help for hosted websites who move their servers due to the fact that the ISPs could be considered publishers - although I'm convinced that 1.3c would excempt on that score. It can't be the case that simply because an ISP has the power to delete a website it is a publisher because otherwise you would be guilty if you saw a person handing out defamatory leaflets and you didn't attempt to stop them.
This post is already far too long, but I'd like to chuck a couple of other articles into the mix: The reform of British defamation law is an essay that came fairly high up in Google, so it must be good. An article about defamation law with a New Zealand perspective is interesting because it explains the judge's argument. Apparently although an ISP isn't a publisher by the section 1 definition, it is in the broader sense. The case depends on the fact that Demon was given notice and did not respond.
So what to do? There needs to be legal clarification, but I don't know which of precedent or an Act will be necessary. The legal costs for such a case are too high for an ISP to try for a test case for the greater good. I'm confident that eventually the law will reflect the obvious nature of the internet, but it's going to take a few years. How to get to that point? Wait, I guess, for the next generation of MPs and judges. By the way, please enlighten me if I've made you choke on your cornflakes.

Okay, UK censorship pisses me off. But what pisses me off more are people who think that this problem has come about because we're stupid (cf todays Scripting News).
The UK is new to this whole internet game. In the US these battles have already been fought - and they weren't pretty. Don't be condescending just because we've yet to figure some things out.
If you look at the reports of the Godfrey case on the BBC or at the Independent, or at Demon's press release then what you'll find is that there are people who know that the settlement (because it didn't go to court, which was a shame) was not a good reflection of justice.
UK law is made up of precedent. There isn't yet good precedent for saying that Demon would have been liable for what was on their servers; all we can say is that they lacked the cash to take it to conclusion. There is a whole generation of internet native speakers to whom the correct thing is not something to argue about, but common sense, and they'll be making the laws soon.
This is why I'm not too worried about this so-called censorship, because I believe it'll be sorted out soon enough (except RIP which really should be stopped). And this is also why I get annoyed with people who legitimise the don't-get-its like Godfrey by moving their servers instead of forcing the matter.

The Garden of Eden is a fun (and full) personal site. I especially like Eve Asks which, in a rare subversion of the WWW, has loads of very interesting, very readable content. Physics and free will has got me thinking.

Cool philosophical difference between academics and non-academics. Bloody pretentious Oxford students. <g>

Lingua::Ispell is a Perl module "encapsulating access to the Ispell program" (last updated April 2000). Aspell has a Perl interface "in the works."

Hmm. There's an odd bug in my email thing that occassionally causes the headers to be posted on this page as well as the email body. Sorry about that; I'll try and squash it.

Exciting! David Adams at Stuffed Dog has published a Speller xmlrpc RFC. Looking forward to that.
Something that for some reason hadn't occurred to me is that he's be using a standard backend. ispell and aspell are both open spellcheckers, mainly for unices. I wonder if a Perl module to use these already exists or whether David's writing his own?

I was insanely jealous when Teeth magazine started because their tech was so cool. I'm a little more controlled over Deepleap because its the same people and I've already been jealous once.
Deepleap is cool, and potentially even cooler. I think the real innovation is the backend which decides what services are useful for some selected text or a current page; the services themselves are just a bonus. I'd very much like to build on that. Hopefully on their not-yet-there developer's page it'll emerge that the protocols (promised to be open) behind it are simple and only require the Deepleap client (and expert system); what doesn't bode well is that the metainfo is kept locally. Cache the xml metainfo by all means, but determine who's metainfo'd up by checking on client requests. The database is distributed in xml.

First the mullet, now sideburns. The internet is turning into a bad haircut shrine. What does this mean?

My Freecom Traveller [image only; stupid Javascript won't let me link to individual pages] drive looks great, but does it work? And it's taken me over a month to get the returns department of Camelot to return my calls and take it back (I'd link, but they don't have a website. Wha?). I'll get a new one tomorrow, apparently - but they won't fool me, I know what tomorrow means. Hopefully I'll have it in time for Christmas.

Argh, and then you get people whose objection to windfarms is based on wanting to protect the countryside. What do these people think is going to happen if we carry on burning fossil fuels, given the whole world is developing now, not just the West and Pacific Rim? What we need a dispersed (low density) power generation system: Wind.

£100 million for the coal industry? Stupid. I can understand it in terms of injecting money into the economy (Keynsian cycles), but what this is certainly not going to do is help the industry. When will the government learn that is you're going to put money into struggling businesses then you should put it in to sectors near the beginning of their cycle, not near the end? This would have been cash better spent on wind farms, which have a cost comparable with coal, don't make noise, and look beautiful.

I've been looking into how to write content for the web. The consensus appears to be:

  • "Inverted pyramid structure" - conclusion first, then details, then background. For the catering section of our new pub website, that translates into: What the catering company is; exactly what does it offer, where and how; an example previous job, including quote and result.
  • Single topic paragraphs, with the first sentence summarising the entire topic. Readers tend to scan by reading only headings and first sentences of each paragraph.
  • Be succinct.
  • Use meaningful headlines and many of them.
  • Be completely honest (ie avoid "marketese").
  • Use strong verbs and the active voice.
  • Keep each paragraph independent. You can't assume people will read in a linear fashion.

These hints culled from:

I'm demonstrating to Andrew how I can update this page via email.

And so it came to pass that ntl gave me unmetered internet access, and managed to only fuck up a little bit. But don't think I'm not grateful.

Someone please buy me this clock. ( is a pretty shitty site, but it has some good stuff.)

If you have a WAP mobile, there's a test of WAP Dirk up. I spent my day learning ASP (which is new to me) with an xmlrpc ASP implementation grabbing the information from Dirk. There were a couple of issues with the client, but David Carter-Tod (who maintains it) says he isn't paying it much attention since there's a COM-based client, and Microsoft are into SOAP instead.
If you think you can do better, mail me and I'll let you know the xmlrpc interface for Dirk. Or you can use it to feed information into your sites. Why you'd want to do that I really don't know, but it's there if you so desire.

Something I was thinking about last night: We're in a time when the rate of inventions is limited by the size of the pool of capable people and how much time they have, rather than individual's innate inventiveness. What gets invented next depends of the evolutionary fitness of current inventions and so we evolve instead of leaping with paradigm shifts.
Is this the way evolution (of species, of ideas) works; an edge-of-bell-curve step (punctuated equilibrium) followed by competition to evolve to a stable system? Or are the two mechanisms of change of comparable magnitude and we happen to be in one rather than the other?
[or: Is it that great leaps are by necessity rarer? Or: Is it that the car industry (say) which is characterised by a lack of paradigm shifts does indeed have a high rate of invention, just like what's being done on the WWW, but that I can see one and not the other?]
So I have some questions: If we just can't take those obvious next steps fast enough to run out, if we can always see where to go next, is there anything to be proud of? And is what's being achieved now technologically on top of the WWW foundation important if there is indeed going to be a paradigm shift again one day, and is what is really important the types of communities and interactions which occur in this sphere which will remain the same as long as there is interconnectedness and as long as there is humanity?

From: Matt Webb <> To: <> Message-ID: <> Mime-version: 1.0 Content-type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit X-UIDL: d3d96417d8980e7a1694ed4894ea9fae weblog I just filed away all those old messages in my Inbox so now it's empty. A good feeling! I'm posting this via email. A better feeling!

Hey, I went to the Millennium Dome on Saturday.
I can't describe how excellent it was (although I'm going to give it a go). I've never been anywhere like it. The Dome was so Now - museums always put you in the context of history; a world Expo is always really futuristic, but the Dome is such a product of our time it's fantastic. And gutsy too! There was a show in the middle of the day, in the middle of the dome, and it was contempary dance. And really good too.
Okay, let me try to explain this: The Dome is huge. My eyes played tricks when I walked in and it didn't look that big (you can't tell the size from outside because the roof curved so much its much foreshortened). But then I looked up when I moved and the roof didn't seem to move by perspective as much as it should have done. That was when I realised that things were a whole lot further away than I realised. Right, so you've got all this space. The space doesn't mean anything until something defines it, gives it boundaries. The main show took place mainly in the air, and two people dancing together suspended in this vast space, above and in front of everyone there - well, it was beautiful, it really was.
The exhibits impressed me too. There was so much optimism. Each Zone has a theme, like Work or Journey, and it tries to represent what that theme means for us, today. By what it chooses to tell us, you can deduce the philosophy of the Dome, and it's positive: socially inclusive, multicultural, environmental. And what isn't it? It's not empire bound or nationalistic. The Rest Zone was simply a large cavernous room playing gentle ambient music (computer generated to last 1000 years) while strong coloured lights illuminate the ceiling and walls. The public sit in the room. People at the door ask you to be quiet as you go in (that would never happen at Disney). It shows such faith in people that the organisers have devoted an area to this kind of exhibit. I'll be going again, before it closes at the end of the year, mainly to see the show again. Wow.

Now if this works (and I think it should) then this entry has been posted via email.
I spent this morning putting a system together. Every 20 minutes a script checks for email: If the email is intended for my weblog, then it formats a request in xmlrpc and fires it off at a gateway on my server. This other script talks to Blogger and posts & publishes. Using xmlrpc should let me use David Adams' Speller when that's available, and also to fairly trivially change to post to EditThisPage sites aswell.
I wonder how many other people have done something like this?

This is very much a test entry.

I haven't seen much commentary about the Microsoft decision. There's a great Slashdot thread about it (comment threshold at 5).

I try not to take links from other weblogs and put them here, but sometimes they're just too good: Assembler (DHTML wonderstuff) [via kottke]; Constructor (Java dynamics and physics) [via everywhere]; and Eels (I can't even begin to explain) [via Haddock].

To do: Find out how to use Javascript and stylesheets to toggle the visibility of a row in a table.

Weird. I've been working on a website for my Mum's pub and it isn't finished yet (not even close), but we've got a domain and there's some design and random Latin content up there. Somebody just phoned to book bed and breakfast and said that he found the pub on the net. Now that is so much more effective than all my persuasions that yes the web really does work.

Where did I read "the URL is the command line"? There's a Scripting News DG message about it, but I'm convinced I've seen an article about it. Why do I mention it? I think a better way of using Dirk would be through the path, like Metababy. But for that I need Apache Rewrites (which have to be compiled in), or Aliases (which have to go in the main .conf file (which I don't have access to). So I'm looking for pointers.

Okay: "Filmed before a live studio audience". Before they die?

And they didn't even know each other! Why do I never sit next to people like this when I'm on planes? I guess it's because I don't travel club class. So that's what goes on behind the curtain.

Dirk2 got finished today, and it's up. Well, I say finished, and I say up, but the fact of the matter is I would have liked a bit more time testing it and tightening up the stylesheet but in an enormous coincidence (or was it synchronicity?) Dirk1 decided to die today aswell. A thing happened. Shit, you know a thing. Can't think of the word. Well, a thing happened, a race thing. Race condition! Well, that happened and Dirk broke, and Dirk2 was there and took over. There. Bloody things happening. Still, worse things happen at sea.

And then I actually checked No Wonder after I linked yesterday, and it turns out it's changed into a place to bid for tech support from a place where it was given free. Okay, so I'm wrong. Do you want to micro sue me?

WAP micropayments [Sunday Times]. Would you pay to view the latest adverts to rent a flat? Or for someone to reserve you a hotel room when you arrive in a new city?
I don't think WAP micropayments will work. WWW has shown that there's always another way of making money - and demographics will only get more valuable. Competition will drive micropayments out. Anyway, WAP is going to cost: mobile providers will charge for anything that goes through their WAP gateway; any of their own services will probably be free. This is not good - but competition isn't big in the mobile arena, and it hasn't turned in the direction of WAP or unmetered calls yet.
Oh, and as for the rent-a-flat thing: would you rather place an advert where everyone will see it, or where only people who have chosen to pay will see it? Well then.
Micropayments in general? Yes. There will be some things in the end we will have to pay for. What? When there is a need for a human interface to process the information, but with free services like No Wonder what will they be? I don't think they will be services though: service companies can sell demographics. No, we'll go back to products. I would pay (small amounts) of money to view BBC News, for example, and that's something someone else can provide.
How far will 'information wants to be free' go? All the way, hopefully, but then we'll all have to find another way to make money.

Is this the ugliest car in the whole world? Or is it a dolphin with wheels?