Then I saw the license statement.

This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 2.0 UK: England & Wales License.

And right there at the top of each ontology's source code is the line

<dc:rights>Crown Copyright 2008</dc:rights>

The license confirms my misgivings about the Ordnance Survey, in spades. Why? Because non-commercial licenses are not free. By using the OS vocabularies, you taint all your data. It is no longer open data (part of the wider field of open knowledge). And if someone uses it, thinking that it is, in a commercial project, they open themselves up to attack from the Ordnance Survey's lawyers (who have proven themselves time and again to be unsympathetic, confused or worse).

What a pity. Clearly a lot of work has been put into the project, but as currently licensed, these ontologies are data plague. Keep your information well away.

Would it be too hard for some of the authors of these add-ons to get together (possibly even mediated by someone from Mozilla) and agree on what their Tools menu entries look like? I think having them all in an Add-on options sub-menu would be a start.

As the contented owner of an OpenID identity, I've come to appreciate the facility of being able to log into multiple websites with a single username and password. Who wouldn't? The legacy model of having multiple user names across different sites - am I hex, downlode, earlemartin, or earle? - to remember, and worse still, multiple passwords*, is looking increasingly outdated and confusing. Sites that don't give you any option but to enter your details (and often way too many details at that), and then not letting you transfer any of that data out, are disparagingly becoming known as identity silos.

An identity silo has been defined (interestingly, by the Information Commissioner's Office of the British Government) as:

An identity that is used to represent an individual in that person's dealings with a particular application and is not integrated with other identities that the person has.

So I decided the other day to make the rounds of some projects that I use on a regular basis, and ask them what they think about OpenID.

First up was LiveJournal. LiveJournal is where OpenID began, and an early first place to allow OpenID logins in some form. However, it is still impossible to link an OpenID identity with a LiveJournal account. In this post to the lj_dev community I ask what's happening. Official response comes there none; and some interesting replies point to much earlier (two years earlier, in fact) discussions in which, allegedly, promises were made to go further (and volunteer patches received) but nothing was done. Conclusion: LiveJournal no longer cares.

Next I stopped by the Wikimedia Foundation's wikitech-l mailing list for MediaWiki developers. The situation there appears to be that once they've concluded their work on unified login for all Wikimedia projects - which is close; as an admin on Wikipedia I've beta-tested it and it seems good - they can look into it. Which is great.

A query to FriendFeed's newsgroup got an enthusiastic user response, but nothing from the developers. The same for this thread in the Ohloh programming site's forums.

Asking the del.icio.us mailing list got no response at all.

Get Satisfaction are working on it (they already have OAuth support, which is good). Brightkite will "probably" do it, but they don't have a timescale for implementing it yet. Twitter have yet to reply a month after someone asked them in this GS thread.

I also asked the wiki-standards mailing list, which is comprised of wiki engine software authors, if they were looking to implementing support, and got a generally positive response.

So that's what I did. You can help out, by asking the people whose software you use if they're planning to implement support. Be sure also to stop by OpenID, please! and vote for your favorite applications.

*Or even worse than that, having a standard password for web stuff (come on, most people do) that varies in some small way across sites due to local restrictions on possible passwords. Ugh.

When Mac OS X arrived, in 2001, I immediately disliked the glossy new interface that replaced the crisp, almost cartoonish GUI of the old OS. More than disliked, actually; I experienced an immediate, visceral repulsion from a number of facets of the new look. As a result, I stayed away from Macs for a long time, until, in fact, this month. However, my current place of employment contains a large and, for the most part, inscrutable systems department, who for reasons of their own decided to give me an iMac. It took me only a few minutes to discover a range of irritations within it, and start looking for fixes. Here's what I found. Software is free unless otherwise mentioned.

Essential fixes

A non-Apple keyboard and mouse

Essential. Even though I was a Mac user for years, and have no problem with the key layout, there's something about the key spacing that's just wrong. I found myself making way more typing errors than I do usually (which isn't many, I'm a good typist). So I got a typical Windows USB keyboard instead. Setting it up to work properly is not as easy as you'd expect, considering Apple's otherwise rather good hardware support. Luckily, Phil Gyford has produced a handy guide to using a British/UK Windows keyboard with an Apple Mac in OS X, including a keyboard layout file you can install.

In addition, Apple's mouse (with the cringe-makingly awful name of Mighty Mouse - pass the sick bag) is abysmal.

• Instead of having buttons, it does left- and right-clicking by detecting when you press on the left or right side of the mouse. Sounds great in practice, but for me it caused unexpected "right-clicks" on a regular basis.
• They've also replaced the scroll wheel with a sort of mini trackball thing. It's horrible to use - too small to get a decent roll distance, and with an itchy, tickly physical feedback instead of the ratcheted rolling sensation of most modern scroll wheels. Plus, the "360°" aspect of the ball is totally useless in 99% of applications (I'm being generous, here; I have yet to actually see anything that supports it).
• It has a shiny, slippery surface. Call me crazy, but I like using objects with a texture.
• The pill shape just doesn't suit my hand at all.
• Finally, it has these "squeeze buttons" on the side. You have to press them pretty hard for a button. I didn't use them/it - they both act as one button - at all except to see what happened when I did (some unnecessary zoomy thing appeared).

Solution: I bought a cheap but perfectly good mouse. My comfort level just shot up.

FruitMenu

The new Apple Human Interface Guidelines say: "The Apple menu provides items that are available to users at all times, regardless of which application is active. The Apple menu’s contents are defined by the system and cannot be modified by users or developers." Well, in the old days you could put whatever you liked in it, because that was useful. This handy "haxie" from Unsanity restores that behavior. For me, having all the control panels - I'm sorry, "System Preferences" available as a sub-menu justifies the $10 cost of this software alone. I also put my Applications folder in it.

TinkerTool

I don't like the Dock much. It took about a day before I turned off the zoomy stuff. Shrinking the icons down helped a bit, as did sticking it on the right-hand side of the screen and turning on auto-hide. But it still insisted on being vertically centered, and I'm a corner kind of guy. Happily, TinkerTool, a sort of PowerToys for the Mac, allows you to stick the thing wherever the hell you want it to be.

WindowShade X

Another feature that the new OS X lost was WindowShade. In the classic System, double-clicking a window's title bar would cause it to "roll up" into it. It was a very useful way of temporarily moving windows aside without actually moving them, or letting you "stack" lots of windows on the screen at once. In OS X, you can only hide windows away on the Dock. No thanks. WindowShade X, another $10 Unsanity haxie, brings that back, as well as a great "minimize in place" feature that lets you turn your windows into live thumbnail versions of themselves with a keypress.

ShadowKiller

Another Unsanity haxie - free this time. Kills those bloody awful drop shadows (on three sides of windows and even menus - what the hell?) stone dead.

Enhancements

Growl

Displays application event notifications in various attractive but low-key ways.

Mail.appetizer

Displays excerpts from new mail arriving in Mail.app in Growl notifications and lets you read it, delete it or mark it as read on the spot.

Hey Folders!

Colors labeled folders' icons as well as their names, as it used to look pre-Mac OS X. (Can you spot a theme in my suggestions developing?)

VirtueDesktops

Virtual desktops for Mac OS X. Not being developed any more as virtual desktops ("Spaces") will be a feature in Mac OS X 10.5, but useful until then.

Essential applications

Firefox

Obviously. The downside is a slightly awkward set of key shortcuts (for example, Command-shift-H for History instead of Command-H, since the latter triggers the global "hide application" shortcut), but I can live with it. Safari just isn't good enough, and Camino, which is essentially Firefox with a more Mac-like appearance, is crippled by the fact that as a native Mac application, regular Firefox extensions (many of which are essential, and I may write an article on these later) can't work with it. You might as well buy a sports car with a V8 engine and rip out half of the cylinders. I'll stick with the real thing, thanks.

iTerm

A terminal emulator and big improvement over the default Terminal.app included with the system. For one thing, it has tabbed windows.

Adium

A great multi-protocol instant messaging client.

After all this, I finally have a Mac I can use comfortably. And who knows, I may even get one for myself at some point. Five months later: No. I'd rather shoot myself in the face.

It is well known that when you do anything, unless you understand its actual circumstances, its nature and its relations to other things, you will not know the laws governing it, or know how to do it, or be able to do it well.

My first attempt at the process was with TekArtist's instructions, which suggested using the Livna repositories. However, trying to install the Livna RC6 package resulted in a complaint about needing a newer version of the fedora-release package. Confusingly, when I attempted to update the fedora-release package, yum told me there were no newer versions available. Looking in /etc/yum.repos.d/ (warning: contains strong language), I couldn't see how to bump the Fedora Core version number. (This was weird for someone more used to the Debian way of things, where a distribution upgrade is as easy as changing one line in your sources.list and doing an apt-get upgrade.)

A bit more searching gave me these instructions from Alexandre Gomes, which suggest downloading the fedora-release package and using rpm to install it. However, that package in turn relied on fedora-release-notes, which relied on fedora-release (I think; my notes are incomplete). The solution was to get both simultaneously:

This successfully updated fedora-release. The next step was to do the actual upgrade of the whole distribution! The command for this is yum -y upgrade (the -y option means "answer yes to any questions asked"). yum went off and got a lot of packages and calculated their dependencies. The process then failed, due to a conflict between the fedora-logos and redhat-artwork packages, of all things. A yum update redhat-artwork resolved that. Running the update command again, I got:

Error: autofs conflicts with kernel < 2.6.17
Error: hal conflicts with kernel < 2.6.17

Predictably enough, the next step was yum update kernel, which brought it up to 2.6.18. However, running the upgrade again resulted in the same error. So evidently some old kernel was hanging around. Investigating, I found:

[root@localhost Desktop]# yum list installed | grep kernel
kernel.i686                              2.6.18-1.2200.fc5      installed       
kernel.i686                              2.6.18-1.2798.fc6      installed       
kernel-smp.i686                          2.6.18-1.2200.fc5      installed       
kernel-smp.i686                          2.6.15-1.2054_FC5      installed

There were indeed some old kernels that needed removing. Now, for the life of me I couldn't work out how precisely to issue the package name and version to yum in order to remove it, despite the section of the yum manual about doing precisely that, so in the end I used rpm instead, which just does what you tell it instead of trying to be friendly.

[root@localhost Desktop]# rpm -qa | grep kernel
kernel-2.6.18-1.2200.fc5
kernel-2.6.18-1.2798.fc6
kernel-smp-2.6.18-1.2200.fc5
kernel-smp-2.6.15-1.2054_FC5

[root@localhost Desktop]# rpm -e kernel-smp-2.6.15-1.2054_FC5 kernel-2.6.18-1.2200.fc5

Pow. Gone. I ran the upgrade command again, and... success! yum went through and upgraded about 1,400 packages. One reboot later and I was running the shiny new Fedora Core 6.

Minor niggles: One of the features of FC6 is the compiz visual-effects thing (because we all need wobbly windows and desktops that spin around on a big 3D cube). However, I couldn't get it working, apparently because of something to do with my video card. If I ever manage to, I'll document it here as well. Also, strangely, you can't get Firefox 2 from the default package list in FC6, which strikes me as bizarre for a major new OS distribution. You have to get it from the development repository, like so:

[root@localhost ~]# yum --enablerepo=development update firefox
[output snipped...]

Updated: firefox.i386 0:2.0-2.fc7

Complete!

Yes, it looks like I've made it to FC7 already. :)

The first problem that I encountered was that the installer would stop working at the first point where I needed to switch CDs, with the message An error occurred unmounting the CD. Please make sure you're not accessing /mnt/source from the shell on tty2 and then click OK to retry. Hitting Ctrl-Alt-F2 to switch to the terminal, I tried unmounting the disc manually:

sh-3.1# umount /mnt/source
umount: /mount/source: Device or resource busy

Switching to the log terminal with Ctrl-Alt-F3, the following message was visible:

14:50:36 ERROR: exception in unmountCD: (16, 'Device or resource busy')

The install process would also freeze completely sometimes (generally a few screens in, most often immediately after choosing what packages to install) with the following error in the log:

14:50:36 WARNING: Try 1/5 for file:///mnt/source/repodata/filelists.xml.gz failed

Doing a bit of searching indicated that the FC5 installer seems to have difficulties reading from the disc if it's in a high-speed drive. So, I swapped out the DVD-ROM drive for an old CD-ROM job. Plus various combinations of other people suggested disabling APM/ACPI, APIC and even DMA. To do that, it's necessary to edit the kernel arguments when the system is starting up. I issued the following:

apm=off ide=nodma pci=noacpi acpi=off noapic

This, combined with the lower-speed drive, allowed me to install FC5 successfully. Although annoyingly it stored the kernel arguments during the install process, which meant that during the first normal bootup it wouldn't start, probably because DMA was disabled. (The boot messages showed that udev couldn't be started.) If this happens, you need to edit /etc/grub.conf and remove them.

If you've got here from searching for the same problem, I hope this helped.

It's common practice to find out when someone was last around by asking one of the bots "seen Somebody?" - the bot will then respond something like "Somebody was last seen on #perl 3 days, 7 hours, 52 minutes ago saying 'You what?'" This, however, is limited to a per-channel basis for the two bots. If Somebody was in #perl three days ago but hasn't been in #london.pm for six months, then dipsy and purl are going to give conflicting answers to that question when asked.

#london.pm's main bot is dipsy, while #perl's is purl. Both were originally based on infobot code (which now appears to be often referred to as "the venerable infobot", which should tell you a lot), but have seen a lot of modification and custom hacks. (dipsy eventually got a total rewrite, and it's been proposed for purl for a long time although nothing's ever happened.)

Ideally speaking, there should be some sort of infobot data format which bots could send to each other, perhaps with a safety measure in the form of a secret key which would never actually be transmitted - in pseudocode key := salt( secret + msg ).

Today I tried to log in and got this message.

Your login was deactivated for one or more of the following reasons:

• Repeated failure to comply with the Open Directory community's editorial guidelines and policies.
• Continuous poor and/or abusive editing.
• Self-promotion and biased editing, including, but not limited to, cooling your own site, title or description manipulation, unfairly editing your own sites or those with which you are affiliated.
• Unfairly tampering with competitors' listings and submissions.
• Inability to function well within the Open Directory community.
• Uncivil and intentionally disruptive behavor.[sic]
• Violation of Open Directory forum and email privacy.
• Spamming the directory.

We do not disclose the specific details of login removals. However, the decision to deactivate your login was made by consensus of the meta community, and thoroughly reviewed by DMOZ staff to ensure that our decision was appropriate and warranted. Our decision to remove your login is final. Removed logins will not be reinstated, and you will not be granted a new login. We wish you the best of luck in your future endeavors.

Needless to say, I'm very upset about it. I would be merely pissed off, but it's the insulting final paragraph that really did it for me. I didn't do ''any'' of the things in that list of bad behavior. Go look at the directory category. See all those subclassifications? The careful organization? That was all me. All of it. Before I became editor the category was a shambles.

And what did I do to deserve this? God knows, because they "do not disclose the specific details of login removals". I know that once I was unable to edit the site for a while, and my login was temporarily suspended; however, after an email to them some anonymous ODP drone was happy to reinstate me.

They didn't even have the common courtesy to send me an email about this. I only found when I tried to log in. I have no idea when this happened, and they won't tell me. But they're only too happy to tell me that their "decision was appropriate and warranted".

So, since the "Open" Directory Project has decided to treat me this way, then forget the "Open" Directory Project. I will never use it again, nor any site that derives data from it. Their loss.

I guess it doesn't really matter, anyway, because in the end the wiki model is going to hammer them into the ground. For fun, have a look at this thread on their BB. Can you say, "in denial"? Browse around that site a little more and you'll get an illustrative glimpse of the incredibly arrogant way the members of the ODP hierarchy treat non-members.

In fact, here's an interesting snippet:

Only in very special situations where the integrity of ODP is in great danger editors will be removed without warning.

What does that make me then, Genghis Fucking Khan? I demand an explanation.

Well, I just got a snotty response (duh, it's the ODP forum). I emailed exactly the same message to staff@dmoz.org, as suggested. Of course, nothing happened.

If you are asking what licence this software is released under, you are asking the wrong question.

Time to start a damn movement. And I even thought of a slogan: Just Say Mu to Legalese.

The availability of streaming audio across the Net has impressed me, deeply. I've spent the last few years feeling very dispirited about the state of music in general. However, broadband and audio streams have lifted my spirits immeasurably. Now, there's no restriction on what I can listen to. The corporate-owned and -controlled radio stations and their meagre output pales into insignificance compared to the incredible choice of music I can find online at sites like Shoutcast. In the last few weeks, I've been able to listen to Japanese and Korean and Romanian and Brazilian and Armenian and Arabic and Hawaiian and Vietnamese pop in addition to the Russian stuff I mentioned; genres I would never normally encounter, like darkwave, roots, ragtime, gabber, "80s alternative college", soca.... All completely inaccessible from the standard radio spectrum. And that's before you include the police and fire and railroad scanners, ham radio feeds, talk stations, recorded live comedy broadcasts, stations that play only music from video games, radio dramas from 50 years ago....

In short, you can find every kind of music under the sun, entirely uninterrupted by the kind of crass, annoying advertisements that you have to suffer through on a "real" radio station. I've discovered a plethora bands whose music I want to buy, and never once felt the urge to smash the radio to make the idiot disk jockey SHUT THE HELL UP - because for the most part there are no disk jockeys. Most importantly, the independent nature of the stations means that they're not forced to play whatever their parent corporation - like Clear Channel - forces them to, and stuff it down your throat.

Choose what you want to listen to. Choose streaming audio.

"Towards the Semantic Web" - Ora Lassila, Nokia

• Driver: Automation
• Short term goal: Interoperability
• Long term goal: "Departure from the Tool Paradigm"

Automation

Remove humans from the loop by making inforamation understandable by machines - accessible formal semantics. Hypertexts become semantic networks. As it stands, the WWW is an architecture for linkages ("pointing") - linkages are data, but data has no meaning without interpretation, and human interpretation doesn't scale.

The Semantic Web is a team sport - the problem is to achieve critical mass.

Accessible formal semantics: Ontologies. Features:

• controlled vocabulary
• concept taxonomy
• other relationships between concepts
• they are "a specification of conceptualization" (Thomas Gruber, Towards Principles for the Design of Ontology Used for Knowledge Sharing, 1995)
• they enable reasoning
• robust ontologies prevent the problem of "semantic drift".

The Resource Description Framework (RDF) allows the extension of ontologies. RDF:

• a data model of Directed Labelled Graphs
• an XML-based syntax for serialisation of DLGs
• nodes and arcs in an RDF DLG are named by URIs
• graphs decompose into object/attribute/value "triples". A statement: "subject/predicate/object".

The Web Ontology Language (OWL) is layered on RDF and offers more expressive power. The Semantic Web is built in a layered manner; not everybody needs all the layers. (Unicode → XML → RDF → RDF Schema → OWL...) XML is not a magic bullet; it's not enough just to use it. (Not only is XML not a lingua franca, it's not even a lingua.)

Ora prefers the "Darwinian" approach to the formation of ontologies rather than the committee approach. He feels that in the end a "mix and match" of individually-created and formally designed ontologies will prevail.

Reasoning and Inference

Reasoning allows one to draw inferences based on generalised "rules" - enabled by ontologies. It eases interoperability by allowing inferences between data in compatible formats. The Semantic Web, via ontologies and reasoning, will improve interoperability of information services.

Tools and Beyond

Before: human-operated software. After: software assistants for a range of tasks.

Formal Semantics

"The manifest destiny of AI". Machine learning could bootstrap semantic annotation for existing content. Ontologies → Reasoning → Agents

"Migrating Thesauri to the Semantic Web" - Brian Matthews, Council for the Central Laboratory of the Research Councils

Searching the Web with current technology has difficulty with ambiguity (chocolate chips versus silicon chips) and complex queries ("find me the paper on X that Y wrote on date T"). The Semantic Web will allow searching on descriptions of documents as opposed to by content.

There's a mismatch between terms of search and catalogue terms: translation is necessary. Restricted vocabularies [ontologies] - Knowledge Organisation Systems - can catalogue all items.

Levels of semantic structure may vary on different systems:

1. simplest KOS: lists, maybe with cross-referencing (dictionaries).
2. next level: classification schemes and taxonomies
3. thesauri introduce more relationships: equivalence (exact, inexact, partial, one to many) - for multilingual thesauri; hierarchical (broader term, narrower term); associative (use, used for, related term).

Existing thesauri can be used to provide descriptions to use as a starting point to develop ontologies for the Semantic Web. The legacy of traditional thesaurus development is that rigorous schema exist, but they were developed for human consumption. Traditional thesauri often provide additional relationships not covered by broader/narrower terms, etc. (Subtype of, instance...)

Thesauri on the Web: provide data in RDF and allow people to mark up their data using these terms. Migrating to the Semantic Web requires:

• increased precision in thesauri
• human intervention.

One possibility is term-to-term modelling (e.g. modelling the relationships in DAML+OIL). Alternative approach: use RDF Schema to model concepts used in thesauri.

The task of migration is a complex one, requiring new tools and heuristics (a revised RDF thesaurus format? Alternative mappings?), but once it has been done, the process of searching for information on the Web will be improved dramatically.