Aquarionics

Telecomuniculture

Okay, so in the course of work (developing games for mobile phones for the next week and a bit) I've been playing with GPRS and watching the news in the mobile world. The ideal is somewhere near Paul's vision, in that Microsoft currently want to design a box that syncs with your alarm clock so it can, for example, tell you that the traffic is murder and you're going to have to leave early. This is an idea I like. Stuart's, however, is an idea I'm scared of, because the future is now.

Bluewater is a FOG shopping centre hiding in the Kent countryside just south of London. It was slightly lacking in mobile coverage, so they employed NTL to help it along. The results are that each customer with a mobile phone can be tracked though the centre, the central system can tell which shops they go in, how long they stay in the centre, and where they go. They can even send text messages to the phone, with special offers they might have missed.

I'm not really paranoid about user tracking. I don't block cookies, and I'm not exactly a difficult person to track down. (Really. According to Google this site is 12 for "Avenell", #1 for Nicholas-Avenell, Aquarion and Aquarionics, and Nine of the top 10 results for Nick Avenell are me, even if I don't go by that name very often any more), And I activly like the idea of a CMS that follows me around a site and recommends other things I might like. I do, however, object to being told it isn't happening, and that it's being done for my privacy and security.

In other fun and exciting news, Category based browsing has been written and enabled, although I still have to do the index for it. Oh, and writings is back, although none of the old content has been put in yet.

Curious

I will continue to find it amusing that Mark's reaction to my ESF proposal (and therefore a link to the spec itself) is archived on Backend.Userland in the RSS examples.

Invalidate

My RSS feed isn't valid because I used the entity "×" in my last entry. So I rewrote the code to de-entity all RSS output, but the × symbol isn't valid in XML either. So, World, How do I include the symbol in my entry without breaking the RSS output?

Shooting Standards

I have to phone the Letting Agency tomorrow, and I’ve no idea why. I got the message about half an hour after the saga was posted, having forgotten that I missed a phone-call whilst on the tube. This is ever-so-slightly worrying, so displacement activity happens.

First, flick your eyes to your left and you shall notice a new suite of icons for the various meta-datas. The icons are variously by Antipixel, who created them originally; Mark, who created more of them for his Raging Platypus parody thing (In which the icon for ESF is labeled “For the Angry and Embittered” – which I like – though if he is using MT, does that mean it supports Epoch time format now? If so, I’ll start writing the ESF Syndication Template :-)), and I created a couple more (Only one of which – blogtree – is currently there). Eventually we’ll have the full suite of Blogicons there, and it’ll take a month to load…

But ho! There appears to be more! Can I have been creating /new/ stuff? Well, yeah. After several hours work and a series of low-grade headaches, I’ve finally managed to grok XFML to the point where it works. Thus you can grab XFML feeds of the site. Now I just have to finish the system which will allow me to connect categories of mine to those of other people who also have XFML feeds, and it’ll be finished.

And what was all the point of that crap? Well, it’s another acronym to place on my under-reconstruction CV, and you can see what it can do in the worlds only XFML Explorer... and that’s sort of the point. “Worlds Only XFML Explorer”, there is no software out there that reads it right now, and no earthly real use for it apart from being able to connect to Mark’s:http://diveintomark.org/xml/diveintomark.xfml so that when we are talking about the same things, I can automagically link to stuff he writes on it. As if our blogspheres crossed over that much anyway. The only topic we appear to share is Cat Pictures, and his are very differant to Mine. Not to mention the fact his XFML file is not updated any more.

As much fun as it is riding the leading edge of the development curve, it isn’t half pointless at times.

Oh, yeah. Pointlessness.

I’ve been attacked.

Well, not me personally, but the Epistula Syndication Format has been accused of being pointless, in an article that misses the point slightly. It doesn’t save bandwidth just because RSS is bloated with XML tags – which it is, but then again so is XFML, XML, and most of the other XML standards out there right now. If you look at them, you’ll note that without the article itself, the content-to-exposition ratio is worse than your average maths exam. ESF merely has an order in which things appear, requiring no exposition once you’ve read the spec. But that isn’t it, as I said. The point is that when you download an ESF feed, you only download an X+6 line file, where X is the number of headlines. There is no place in the standard for the content itself, and no place where (as was in the original RSS spec) you can bend the standard to put the site. When you get an ESF feed you just get the newest headlines. If you want the rest, go for a more bloated standard. It’s not designed to replace RSS, it’s designed as a scaled-down version for people who will read the content on the site that it was published on (In exactly the same way as 90% of all “Trimmed down content” RSS feeds do).

When I summoned it into existance (ESF, Syndication format level 9, Plus four against bandwidth, minus six against RSS Developers), I said “It’s Just Data”. Partly as a ‘publicity stunt’, I agree, but mostly because it is just data. Knowledge is information in context, Information is processed data, Data is flat alphanumerics. XML is Information, ESF is data, RSS is information, Ampetadesk is knowledge. ESF has, in classic geek style, not got the information at hand, but knows where to get it if you need it. If you don’t need it, you didn’t just download it.

And yeah, I know it’s similer to RSS 3.0, but I dislike polluting namespaces, so I didn’t read it until after someone pointed it out as a similer idea.

Really Simple Crossreferencing

A proposal, and request for comments. This isn’t finalized yet, comments would be appriciated.

RSS2 has currently no method for indicating links relivant to an item. For example, this item will eventually have a link to the final version of this spec. A discussion on the nature of a new law bill would have a crossrefernce both to other discussions on a topic (such as Trackback or Pingback, or a manual link) or items relivant to the object itself (such as a BBC News story on the bill, or the text of the bill). It would be useful for RSS user agents to have this information, and for systems such as feedster to be able to tell what was relivant to what. Also, it gives a method by which the various discussion monitoring protocals (Pingback et. al.) can be displayed in an RSS2 feed without a seperate module for each standard.

The format of the link (within the tag) would be:

<qv:link type="Pingback" href="http://www.moronweb.com">Moronweb on the decline of vanilla ice-cream in the modern world</qv:link>

User agents would be encouraged to group crossreferences by type and display the contents of the type attribute (rather than attempt to keep a conclusive list of all possible values of type)

Issues

First, this is yet another addition to the RSS 2 spec, and one that has currently no known use. Is there any point in adding this (or anything else) to the spec whilst so much of the available metadata is being completely ignored by modern user agents?

(Above point made by Phil Ringnalda on the Blogite list where this was originally brought up)

Comments? Suggestions? Is this a waste of time and effort, and will anyone support it if I release it?

Crossreference Theory

This is what I want to happen:

I write an entry on Pingback and how good it is compared to Trackback. I crossreference it (Like the links in the entry below) automatically to the Trackback weblog and Pingback spec using the respective technology, manually link it to pingback v trackback and also file it under the Intertwingularity category here, and give it the added keywords “Pingback”, and “Trackback” So far, so manual.

Epistula then automatically links to the last couple of entries I’ve made in that category over the last few months, as well as any document in the category that appears to be popular (In this case, the ESF spec).

At this point, Epistula has done everything it can to crossreference things internally (It could, optionally, add some things that are relevant to ESF, but deciding what is relevant is a little complex) so we move one step forward and do a websearch (using the “Google >API) for “pingback”, “trackback”, and “pingback trackback” and publishing the first few results, priority given to any link that appears on both or all three of those lists. This would also crossreference the entry to 0xDECAFBAD and some others. Also, I’d do the same thing with Feedster results (Though the keywords would have to be better) and possibly Daypop, once again prioritising things that appear in more places. That has to have a user input, though, since the results aren’t always relevant to the article, though always to the keywords.

Again, so far, so internal, and very much an automatic thing, a global thing, rather than any kind of prioritising towards people whose views I respect and/or care about. Nothing, in other words, that is directly relevant to my opinions or articles. Better would be the ability to join categories together, which is why everyone should export as XFML. That way I could say Simon’s category on Webservices is roughly equivalent to my category on “Intertwingularity”http://www.aquarionics.com/category/intertwingularity and so I if he’s been taking about the same sort of things I have in the last week they’ll be linked. Then I can also decide that anything with the keyword “Floz” is equivalent to Sarabian’s category Web Coding and Design. Should I want to. Well, I could...

The problem is that all this requires other people to have XFML feeds, and they don’t. So do :-P

Please stop hanging around

So this is what we find.

In the last days of March 2003, nobody could have imagined that forces beyond the imagination of average were conspiring to make Aquarion Kael D’Blue’s life more complicated than necessary. Necessary, in this case, being… well, at all would be nice.

The house hasn’t happened. I’m still homeless, still staying with Pol & Supermouse in the wilds of Aylesbury, still trying to get back the deposit for the old place, still recovering the extra rent they took by accident, still trying to resolve the new flat before I wear out my welcome here.

I hate, with a passion unholy, this, because I should have moved in last week, I should have sorted my life out, and even with the crap at the end of last month, I should have moved in by last weekend. To be fair, this isn’t all my fault, but it’s still annoying.

The results of this are severalfold. Not only has it resulted in the creation of Point First d4, quick and painful replacement for the dead 2d10 site – in three days flat, but also a number of minor (mostly internal) additions to Forever (Including the RSS feed, and the ability to add new users, which has apparently been dead since December) but also a bug & suggestion tracking system (Not written, but installed, in the form of Anthill, a PHP Bugzilla-like) for all the various Aquarionic systems (Epistula, Afphrid, Aqcom & PFd4) and I’ve also started on a redesign (You can act all suprised now) which, whatever it ends up being, will be less orange than Simon’s.

In less technological news, I’ve been helping Pol & Supermouse do such things as put up fencing, trellising and posts. Also clearning out server rooms, putting things into lofts, and watching copious amounts of really, really bad TV curtosy of my first detailed exploration of the anchient art of having hundreds of channels of bad TV at my summoning.

In conclusion, Samuri Jack is quite good, but most of them should have been shot at birth. Oh, and Graham Norton is So Very Annoying. Conclusion ends.

Oh, and we appear to have succeeded in invading Iraq, leading to lots of incredibly cinematic pictures of US Troops – ably assisted by grateful residents – destroying statues. I mean, you couldn’t arrange for such cinematic footage. Sorry, I’m being cynical again.

Does that mean we can get back to the important things, like new XML specs, Grad school, Personality Defects and the value of scemantic markup? apart, of course, from those who never left it. I tell you what, I’ve got this great idea for how to communicate between weblogs, and it involves carrier pigeons with the evil bit set...

Commentary

Paul has been adding things to his RSS feed. More specifically, he now has the <comments> attribute. I've followed suit (in the RSS 2 feeds) as of just now. The only reason I hadn't done it before was that I was under the impression that the comment attribute was the url at which you were to add and view comments. Since Epistula (And Klind before it) doesn't work that way, I ignored it.

We also have the slash:comments stuff, which it would be nice if somebody used, and the dc:subject for categories.

And people wonder why I think RSS is overly bloated?

New Toy

Data is fun.

One of the nicest things you can do to a programmer like me is give him (or her) a pile of data in a nice format that is easily extractable, and tell him (or her) to go have fun.

Last night, after a couple of weeks of looking at it, pol installed mod_log_sql which is an Apache module that makes all access logs become a mysql database, so a list of referers becomes simply “select count(referer) as referals, referer from table group by referer order by referals desc”. Hit counters become easy, life becomes good, and I get to do all the automatic “This is popular, if you like $foo, go see $bar” stuff without parseing text files! Woot, and indeed, yay.

Syndication::ESF

Um. Someone has proposed a Perl Module for ESF.

Golly

Scientific progress goes eep

There is now a perl module for ESF.

Yet MT(Moveable Type) still can’t do it.

CDF Files

This entry is going to be very geeky for about four paragraphs, and then be actually useful to non-geeky people

Do you know what’s fun? Syndication. Syndication is always fun. It’s one of those topics that never ceases to enlighten, entertain, and educate. And golly, were every discussion in the blogsphere as calm and rationalised as those on Syndication standards, wouldn’t the world be such a wonderful place? Gosh golly and wow, it would.

It’s obviously in my interest to promote little-used syndication standards. I created one, after all. So I present the newest addition to the wonderful example of standards-based technology that is Epistula: CDF.

What Is CDF, I hear you cry. Well, CDF is a syndication standard invented by Microsoft for their content-push technology (Remeber, Active Desktop? Netcaster? All this vital information being delivered to your desktop? Wasn’t it wonderful. wouldn’t it be the next big thing? Apparently not.) It was, in a very real sense that Dave Winer would probably disagree with, the forerunner to the modern use of RSS. It even has a distinct advangage over RSS in the standards market, as it’s a W3C submitted standard.

So, Epistula now produces CDF feeds for everything it also produces RSS, RSS2 and ESF for (I love my module system). What use is this to you, I hear you ask. Well, if you are an IE user (And have IE as your default browser) you should be able to click on a CDF file (Like this one) and have it as you what you want to call it. Thereafter you will have a menu called “Aquarionics” in your favourites folder containing my last 10 items.

Actually, that doesn’t work with my computer, since IE6 renders it as a XML document, apparently that overrides windows’ ability to deal with it as a file-type. If your system also does this, you might need to do put this into a run dialog and execute it:

rundll32 cdfview.dll,Subscribe http://www.aquarionics.com/meta/all.cdf

But now at least I have one syndication method available that doesn’t require normal users to install extra crap. You can find all the various syndication methods (including feeds by category) in the Syndicate section.

Attachments

• Screenshot of a CDF-powered favourites item

Fixing the right problem

Kevin My ex-sysadmin (to be credited with getting me off IIS-based hosting, leading to the ability to use PHP and do all the wonderful things you see before you) has proven his elite coding skillz once again with the release of MT-Epoch, a Moveable Type plugin which will allow you, yes you, to create Unix Epoch dates in your entries, meaning that for the first time, MT users can create ESF feeds!

Yay for MT’s plugin system :-)

Nechophila

So, Today I had an Interview, and it went well, and I hope to get the job, and stuff.

But on to the more important stuff. Various people have been talking about a new format for syndicating and interacting with weblogs (I’m working on a WYWO for the project) and Sam Ruby released the First cut of the spec. Aquarionics and Epistula, in a shocking display of interoperbility, now supports it. It’s all in the newly recoded Syndication section

Independance Day

Today the US celebrates it’s separation from the Empire that founded it. One day, I hope my country does the same.

So, the Interview thing, then. Wednesday afternoon I got an email saying that the people who I applied to would like to interview me, and would tomorrow be okay. This was interesting. I’ve never had an interview with more than 24hrs notice ever, so that was less of a problem. Casting Summon Parental Bail-out so I had enough cash to get to London for it was annoying, though. Monday morning I wandered down to London, found the place, met ccooke for breakfast in King’s Cross station, wandered over and was interviewed for an hour or so. The company looks great, it’s doing something I’m really interested in, and I think I did quite well. Even if I don’t get the job I’m going to watch them very carefully, and when they go public with it I’ll be the first in line to get involved.

Beyond that, shtum, since if they aren’t saying anything, I’m not going to either.

I _do_ hope I get it though. They were interested in Aqcom and Epistula (And ESF, which is spooky) and I got to rationalise the theroy behind ESF, which was interesting and something I’ll probably write up some point soon, because the artists currently known as The Necho Project are falling down the same holes that led from RSS to ESF.

Today, however, I tripped over Silence of the Lambs: The Musical, which is just weird

The Sin in Syndication

After discovering – via an inbound link from Solitude – that CDF files should not be served as text/xml but as application/cdf instead, and also within moments from gilmae that my CDF files were invalid I’ve fixed both of these. IE Users can now just click any CDF file and make a menu of my last ten items in any category or section appear in your start menu. Neatocool.

Also, magic linking is starting to happen. Eventually this is going to be vector-based searching and crossreferencing, but for now it’s simply the last few items in each category that this entry is in. Due to technical issues this only works on entries added after I switched it on, but I’m working on this…

Living in Syn

Hot topic within the geekoblogsphere this month is – in reverse order – the WOX project and WinerWatch.

I’m going to ignore WinerWatch (which is password protected now).

The WAX project – also known as “PIE” or “nECHO”, but I like “WOX” to stand for “Weblogs over XML” Eventually they’ll think of a better name and a permanent one, ‘till then I’ll call it WOX.

The project, whatever it’s name, is really simple at it’s heart. They are trying to define an XML format for weblogs. Problem is they are making a number of mistakes, and because I don’t trust Wikism they’ll never know I think that. (I was involved in Everything2, one of the first wiki-likes, and then went away for three months. In that time the mood of the site and general consensus was changed, and half my work was deleted. I’m now extremely wary of putting anything into that kind of public editing process) so you get this rant instead :-)

When I wrote XML is the new black I meant it. All-things-to-all-people will be the death of XML. If you look at RSS2 you can see exactly why Dave Winer doesn’t like Funky Feeds (Which a careful calculation has seen means “Anything that uses name spaces”), but his reasoning is different to mine.

My point, and the reason I created ESF last year, is that when you are sending out a version of your site that’ll be collected once every hour or so by anyone who is even vaguely interested in what you say, you want to keep the amount of bandwidth that is being taken up by that feed to an absolute minimum. To a site like Aqcom where most of my visitors are normal browsers this isn’t much of an issue, but for people like Mark or Stuart where a large percentage of their readership browses with aggregators (Last time I saw Kryogenix’s stats (Which were updated in April on the page I found) his XML-feed count was twice his home-page hit-count. RSS Readers account for 1.58% of my readership (IE 49.74%, Moz 22.7%)) this is a bandwidth-breaker. It’s the reason Mark only puts excepts in his feeds. If you feed your entire site, including meta-data, I can’t help think you’re giving too much away.

Syndication means feeding your content out so other people can use it. The current model includes facilities for extending the feed infinitely using name spaces (meaning you can include foaf, ent dc or whatever data you want in your feed) which seems like a neat idea, until you have to support it. Do you know how many XML specifications there are for categories? DC has one, ENT _is_ one, WOX itself has a proposed “metadata” tag for this kind of thing, how is an aggregator meant to be able to tell what it is? The problem with names-spaced XML is that in order to display a page correctly, you have to understand each and every tin-pot format the creator has used, meaning it’s ideal in an enclosed environment where somebody somewhere defines what name spaces the document uses, but loose on the Internet it means that any given aggregator has to keep track of hundreds of specifications if it wants to get all the information it can out of the feed, not to mention the problems of people who pollute the given name of a – and I use this phrase in the loosest possible sense – standard. On top of all this metadata for the entry, you are now putting in metadata for the feed itself, meaning that for every element of data you include, you have to explain it, further bloating the feed.

This is why I think WOX is making the large mistakes. Also, I disagree with the decision that trackbacks and pingbacks are comments, and have to be treated as such, when I don’t.

Updates

Aquaintances now exports a valid OPML file.

This was far more work than it needed to be, because I have been unable to find a reference for a valid OPML file anywhere, blo.gs OPML files got imported by Dave’s Wonderful New Toy as “0 feeds added”, which is odd, because they were in exactly the same format as Dave’s old Blogroll before he redesigned Scripting.com. Grr.

Paul? Does this lower my Winer Scorecard number?

Oh, yeah, the other thing I’ve done today.

Banners are sticky. That is, it seems a shame to lose all these nice banners I spend ages making, so they now stick to the archive. If I can find my archive of all the ones I did last time I did the rotating banner thing, I’ll put those up too, but right now it’s just this weeks and January’s.

And yes, I’ll explain the “Frowny Lightbulb” thing soon. Promise.

Event Share Framework

gilmae alerted me to the fact that someone is creating an RSS 2 extension called Event Share Framework or ESF. This could be interesting.

I’ve just sent them this email:

I’ve just discovered your site, You should probably be aware that there is a syndication standard called ‘ESF’, the Epistula Syndication Format, invented at the height of the ‘RSS needs to be extended’ argument two years ago. It’s more or less obscure, but a high number of people are still generating it (Sam Ruby, for example, at http://www.intertwingly.net/feeds/) and there exists a number of modules and extensions for weblogging systems to use it. The spec is at http://www.aquarionics.com/article/name/esf.
I would – as the creator of an existing syndication standard – prefer it if your syndication standard extension did not share the name, to be honest.

Magic Blue Smoke and Mirrors

Okay, so I melted a floppy drive. Could happen to anyone. Who hasn’t done it? I mean, there they are, solid plastic, sitting in your computer case. Obviously at some point it’s going to melt. It’s just going to.

So this is what happened, This is why I hate technology and am going to forthwidth go live in a bunker. No. A monestary. Get me to a monestary, because if this is what it means to be free, I don’t need it.

(Today’s obscure reference brought to you by Winamp)

(Tails of technological woe start in three paragraphs time, feel free to skip to them)

I had a computer called reef. It was a good computer, the heart of it was the Celeron 333 my parents bought me for my 18th birthday. It’s a grandfather’s axe thing, I’ve replaced the Mobo, memory, hard-drive, graphics card, case, keyboard, mouse, floppy-drive and network port, but it’s still the Celeron my parents bought me.

The Mobo is a little dead, in fact, as at some point 2002 the PS/2 ports on it died, and since it lacks decent USB support (Or anything else) it means I can only access it remotely, which isn’t any problem as it’s a server. It’s main functions were replaced mid last year by Atoll, a 2ghz Athlon doing exactly the same things at 5x the speed, so here I have a spare box doing nothing that I can’t use as a desktop.

I also have a new broadband connection arriving on Tuesday, and the router I used to use with it is an ADSL router, and this is cable. New solution time. I decided to turn reef into an IPCop box (I’ve used IPCop before, I’m used to how it works, and I understand it in a way that I don’t for – for example – Cisco routers) which would stand between our network and the rest of the world. So far, so hoopy. Pol has also lent me a Wireless Router, and I can therefore switch to Wireless and get rid of the unsightly cables that have littered every other house I’ve lived in.

Because I can’t plug a keyboard into reef (prospective firewall, welcome back, woe fans) I decide to put the componants into Maelstrom (my desktop box) install IPCop onto the HDD like that, and then transfer it all to reef and then go. I get as far as turning reef into a reef/maelstrom hybrid and halfway though formatting the old HDD before I realise that maelstrom has a 6 gig + 20 gig HDD, and reef has a 20 gig HDD which it doesn’t need. I therefore put Maelstrom back together, rescue important things from the six gig HDD (onto the secondary HDD) and install IPCop onto that. It’s about this time that I realise I’m going to need another NIC, so I put Maelstrom back together to get online and buy that.

‘cept I’ve just reformatted my Windows drive, so I’m going to have to reinstall Windows (No broadband + No install CDs + Winmodem == No Linux). It’s about this time that I plug the power for the floppy drive one pin to the right of where it should be. I wonder where the smell of TCP is coming from. I start the install process, with my Unattended Install Floppy in the drive (This floppy, as you may remember, makes Windows install itself without any input from me). Setup doesn’t read it. The smell is getting stronger. I note that the BIOS doesn’t see the Floppy drive on reboot. It’s really starting to smell quite bad now. I turn off the computer and check the connections, realising that the floppy-drive power cable is really quite hot now. A short tug and I have a twisted mass of plastic, two pins of the power connector, and a really fucked floppy drive.

Did I mention the floppy drive has my WinXP Serial number on it? It does.

Did I mention that this number’s only other place of existance is in the home directory of the server that I now cannot access because my computer is buggered?

It is.

So, girlfriend’s laptop, serial number, reinstall Windows.

“Drive F: (Secondary hard-drive) isn’t formatted. Would you like to format the drive you put all the important files on three paragraphs ago?”

No.

“It looks like you’re having a bad day. Would you like some help?”

Fuck off.

I ordered some new parts from Dabs. They arrived next day. Yay Dabs.

I fiddled with the cables. I could see my old Hard-Drive. Yay cables.

I installed IPCop with the new NIC. It worked. yay IPCop.

It didn’t boot. Boo IPCop.

I tried to reinstall IPCop, but it wouldn’t boot. Boo IPCop.

I tried to reboot windows, but it wouldn’t boot. Boo Windows.

I tried to boot with a handy Gentoo Live CD, but it wouldn’t boot. Oh bother.

Apparently the secondary hard-drive was still not working. I discovered this by a process of elimination about an hour later.

I installed IPCop. It worked. It booted. I put reef back together. It booted. I put maelstrom back together with the new floppy drive I bought from Dabs. It booted, but performance was crap. I reinstalled Windows. It booted. It worked. I watched Bubblegum Crisis 2040 for several hours. Yay Anime.

Right, part two. Wireless networking.

On one side of my bedroom, I set up the wireless router and plugged it into my hub.

On the other side of my bedroom, I set up maelstrom with the new PCI Wireless NIC I’d bought from Dabs. After a little mucking around and leaning on the reset switch of the router until it forgot it’s old password, I configured the router to work.

(Not work the way I want it to. All I want it to do is act as a hub and forward packets to the router. I don’t want it to filter them &#