Tom Morris

Back in March 2006, I wrote a blog post called The Conversation Garden, asking whether or not anyone is going to do anything interesting with the humble message board. Hil over at Spirits Dancing points out that Twitter is the conversational marketplace that I described. Despite Facebook's looming presence, I still think the Web has a lot going for it.

Despite Twitter and despite blogging reaching an apex (perhaps?), I still participate in a fair few message boards, and a curse-inducing number of mailing lists. There's still a significant signal:noise problem, and it's not going away. People still aren't thinking boldly enough about this kind of thing. While in the nineties, it was "just install a message board", now it's become "just install a blog". Big deal.

POWDER Grouping has been released as a W3C Working Draft. What does it do then? It specifies a framework for describing patterns of IRIs (that's URI+UTF-8 for those of you who can't keep track of what the current verbiage is for them funny things you put in the bar in yer browser). Why would you want to do that? Well, you might want to say in a machine readable way things to the effect of:

All the pages on www.example.org/porn/ are pornographic.

All the pages on www.porn.com are pornographic, except those in the /notporn directory.

All the pages on www.twitter.com/tommorris/ are written by a person named Tom Morris who also has a homepage at tommorris.org.

POWDER Grouping describes the formal semantics of a group of resources. Membership in that group is done through a variety of means. It can be through providing an explicit listing of URIs or by providing a list of patterns which they all meet - for example, a regular expression or a simple string match (of the sort "/notporn").

Once you describe a set of resources, you can then make declarations about that set - for instance, that they all contain a certain class of content, or that they are all made by one person - or whatever else it is you want to say about them.

Like a lot of the current work going on in the W3C's Semantic Web workgroups, it seems at first glance well specified. Keep it up!

I tried to post the following comment on the BBC Internet blog attached to this entry by Ashley Highfield about Internet speeds:

Broadband 'speed' is not the only issue. The actual usable monthly download limit is also important. It matters very little to me if I can get 24 Mbps, if I can still only download 20Gb a month over it. If we are talking television, the difference between it taking fifteen minutes to download EastEnders and five is not important. But if you can only download (or - sorry - stream, since this is bbc.co.uk/blogs and we have to be politically correct for the rights holders) half the programmes you want to watch because you hit a monthly cap, that's the limiting factor, not the speed. Everything else that I might do - checking e-mail, browsing the web, SSHing into my servers, a few online games - are quite fine on 2Mb.
We need to demand from our ISPs that they provide not only high-speed connections but high-quality connections, which includes the ability to use them without download limits.

The Perl script (it's the BBC - what else?) that Movable Type uses to accept comments seems to not respond or something.

A few years ago, when I started blogging, I would often explain to people that a lot of my blog posts don't have titles. This is something I inherited from Dave Winer, whose software the current incarnation of this blog is roughly based upon. On my blog, there are two types of posts - short and long posts. Short posts fit inside one paragraph element, and are styled with a blue background. They all appear at the top of each daily archive page. Long posts have a title and one or more paragraphs underneath them - these appear at the bottom of the daily archive page, and have a white background. Both have permalinks. Both are types of blog entry.

My reasons for this are simple. Sometimes blog posts are too short and fleeting to require a title. Mostly these are links, short off-the-cuff remarks or single paragraphs. Many bloggers used to call these "asides", although I don't like that word. But still, people would tell me that this is very strange, and that really everything should have a title, and that I was just being picky or a nuisance, and to please shut the fuck up and let the normal people carry on the conversation.

The idea that everything in the world has a title is silly. A huge swathe of artists call their great works things like "Untitled No. 7" for a reason - because they can't think of a title, but have to in order to exhibit the damn thing. Instant messages don't have titles attached. There are a lot of UNTITLED.DOC files out there. Blog comments don't have titles. Not giving something a title is a normal part of human experience.

I still post things without titles. And so do most of the people telling me what a weirdo I was a few years ago. Twitter, Jaiku, Pownce, Facebook's status updates, FriendFeed and so on all prove that.

Can Dave Winer and I now get an apology from all the trendy people telling us that we were stupid and insane? If you don't want to lose face in public, you can always send a Twitter direct message. They don't have subject lines, after all...

It looks like the religious nutcases are trying to prevent the passage of the new Human Fertilisation and Embryology Bill. Sorry, but if you think that this is a bad idea, I think you are not only wrong, but you deserve nothing but contempt and scorn.

If you think that we are going to anger some almighty unseen actor if we conduct scientific research that could lead to medical advances that might be able to treat painful, life-threatening diseases such as cancer and Alzheimer's, please go and get fucked.

Here is why medical advance is important. About nine years ago, I watched as my paternal grandfather slowly died in the hospital. He suffered from Motor Neurone Disease, a neurological disease which is currently incurable. He was diagnosed at quite a late stage, and very little could be done except try and make him comfortable for the few remaining months that he had to live. John was a very practical sort of person - he was an architect by training, who spent his life working for the government designing public buildings - mostly prisons. When he retired, he spent his time painting - and got to the point where he could sell pictures.

When the disease took it's toll, it was horrific to watch. Imagine a person you love being unable to do any of the things which make them happy in life, and eventually unable to communicate - not being able to speak, write, draw or move. The quality of life for John was decimated in only a month or so. A week or two after this happened, he died in his sleep in the hospital.

Preventing others from having to go through this kind of thing is what the religious are opposing. Of course, they do so for other reasons. But their illogical and stupid reasons do not change the fact that they are pro-disease, pro-cancer, pro-motor-neurone-disease, pro-Alzheimer's, pro-AIDS, pro-heart-disease and pro-suffering. Their argument rests on a biological exceptionalism that sees human beings as uniquely different. Sure, we are - we are the only species to build skyscrapers, write novels and build global communications networks. But that does not make our genes special.

There's a reason that the Medical Research Council, the Wellcome Trust, the Royal Society, the Parkinson's Disease Society, the British Heart Foundation and Cancer Research support this bill. It is the reason I support it also. This type of research has the potential to find causes of diseases that kill and harm thousands of people every year. There is no good argument against this. It is the religious who object to the prevention of the sort of pain and suffering that my grandfather went through in the heartbreaking final months of his life. Anyone who attempts to halt efforts to prevent others going through similar pain deserves nothing but derision and scorn.

I've just joined FriendFeed. It emits FOAF, but it's filled with Too Many Fucking Blank Nodes! They are using nodeIDs for all the knows though. But here's the thing - the people actually have URIs. Use 'em!

Oh, and my FriendFeed is here.

(What? You don't understand this FOAF thing? Well, I do! Hire me and I'll help you make your FOAF not suck!)

Apparently, there's been a huge trainwreck keynote interview over in Austin at SXSWi. Full details here, here and here. Twitter and IRC were going wild during that period, with people saying they were walking out, booing and all but throwing bottles at the stage. LOL!

Seems like typical softball, self-serving conference bullshit. It's the reason I go to BarCamps instead. You should too.

I wonder if anyone can help me. I'm trying to connect to the Internet over GPRS with my Sony Ericsson W810i mobile on my MacBook Pro. I generally prefer to use the USB connection cable that comes with my phone, rather than using Bluetooth - both for data security and because it means less power usage (in my experience, it is also marginally more reliable in terms of disconnect speed - when I tell it to disconnect via USB, it generally does it quicker than over Bluetooth). I can't seem to connect via USB though. Here are the settings as I've got them:

System Preferences 'Network' panel - under Bluetooth, I've got "Orange" as my username and "Multimedia" as the password (as is listed here). There is no Telephone number listed, but if you click 'Advanced' it's got Vendor set to Sony Ericsson, Model set to GPRS (GSM/3G), APN set to "orangeinternet" and the CID is set to 1.

The panel for Sony Ericsson W810 is set for exactly the same settings. I've tried all ten of the possible CID values, and I've tried putting the APN setting into the phone number field.

This worked on my Mac the last time I used it (before the hard drive failed), but isn't working now. Suggestions would be greatly appreciated. Reward for successful instructions: ten million ISK (EVE Online) or an invite to a "Web 2.0" site of your choice (Github or Jaiku or anything else I've got invites for).

I've been using a variety of wiki systems for some years, and I have to say I strongly dislike most of them. Starting with WikiBase, and through Wakka/Wikka, PBwiki MediaWiki, Confluence (a so-called "enterprise wiki" - which is to say a wiki made so complicated and hacker-unfriendly that it's possible to sell to those running corporate IT departments), SocialText (also for 'enterprise') and a variety of different personal wiki systems.

I can't stand them. The simpler ones tend to be a lot more useful than the complex ones - and MediaWiki is a tolerable middle point. But I really don't like most wiki systems as they currently stand. There are a few simple reasons for this.

First off, I don't like writing text in a web browser. I have TextMate and It's All Text! which allow me to edit any text area in a proper editor like TextMate. Can't I do that with wikis? Yes, sure. And I do often. But I want it to go further than that. I'm offline a lot. I want to be able to edit a wiki while offline and then push my changes up to the server when I'm back online. A wiki is - at it's core - a version control system for web pages. But compared to version control systems that programmers use - even now-redundant systems like Subversion and CVS (hint: Git) are a lot more useful than most wiki systems.

Another thing I hate about wikis is the syntax. Almost every wiki system does it differently. What was so wrong with HTML that we had to come up with a thousand reinventions of it? Now instead of learning HTML, everyone has to learn a hundred crappy, less-expressive versions and keep track of which particular dialect of this mess we should use on each particular wiki.

Third, every wiki should have proper diffs. I track changes on wikis via RSS, and what I'm interested in is the diffs. Okay, most normal people don't read diffs. Do I care? Not really. So, give me diffs of every change.

Fourth, most wikis I use are slow. Compared with typing "mate foo.txt" into my command line and working on a text file, I have to go to the site, find the page using rather primitive search tools (compared with using grep or most editors 'Find in Project' commands), then log in, wait for another page to load up, remember the stupid wiki syntax and then edit the page. All that is discouraging. Okay, it's fine for ordinary users - but it actively discourages me from contributing to most wikis. These problems are often exarcerbated in 'enterprise wikis' which tend towards featuritis at every opportunity.

Here's my ideal wiki system. It'd consist of a folder full of textfiles, tracked with Git. Then when one types "git push", it adds all of your commits to the wiki, which are then published on the website. And that's it. Recent changes? "git pull" then "git log". Maybe if you want it to have some extra functionality, you have a simple script called "wiki" which you could just pass commands to like "wiki orphanpages", and it'd give you a list of orphan pages. All the non-geek users could use a web UI much like they do already with wikis - go to a website and make commits on there.

Oh, it looks like Git Wiki may be the starting point for what I need...

There's a podcast up over on Nodalities where Danny tolerates me waffling for about an hour on stuff I've been working on related to the Semantic Web. Enjoy!

...and he's making otherwise intelligent people swoon rather embarrasingly.

I'm calling you guys out - yes, you Ian Davis and you Paul Miller. I get it. You meet Nova, have a few drinks and then your readers are subjected to late night 'Web 3.0' blog posts.

If we want a phrase that's better than 'Semantic Web' then 'Web of Data' is far more self-explanatory than the cryptic 'Web 3.0'. Buzzword hotel sure is full - what with Graphing Social Patterns in mid-flow over in San Diego. All the Valley geeks have got their knickers in a twist about Facebook and OpenSocial/MySpace. They are banging on about how important platforms are. Last 'platform' that I used was AOL. I prefer the Web though. Lots more social graph nonsense (repeat after me: 'social network data' has no information loss, is understood by normal people and doesn't make you sound like a preening Valley type - sure use 'social graph' when you are talking to your VCs, but, please cut it out when you take the shirt-and-tie off).

If we are going to build a Semantic Web and still be sane and have a head full of hair at the end of it, we need to stop coming up with jargon. Jargon is fine if it describes a well-defined technical thing. An example might be 'reification'. It has a clear technical definition. I can go and read something like the RDF Semantics or Concepts and Abstract Syntax document and figure out what the person is talking about. But 'Web 3G' or 'semantic graph' or - shudder - 'social graph'. What on earth is everyone talking about?

Some friends and acquaintances of mine were at the WebCamp Social Network Portability event in Cork today - Dan, Stephanie, Uldis, Aral, Ben and others. I wanted to go along, but didn't. Instead, I had a nice lie in and watched a live stream on Ustream and asked questions in IRC and got mentioned numerous times during the talks. The Internet lets me be a ghost and inflict myself on an event in Ireland without even leaving my desk.

There are some slides up - danbri's keynote, Uldis' on three-line FOAF files. Steph has put up a video of her talk about social network portability from the perspective of the user. There'll be more up on the wiki, I expect.

In the closing session, I remotely asked this question in the backchannel to the panel: So, do you think that these problems are going to be solved by large groups of people in huge mailing lists and working groups or by smoke-filled rooms filled with hackers more focused on solving problems than coming up with philosophical thought experiments?

I think it's an important question. Personally, I think that any of the approaches to this need to be driven by rough consensus and, more importantly, running, well-tested code working to well-defined and well-tested specifications. Convoluted mailing list discussions or blog posts or conferences should be a sideline, and the more important part is writing the damn code. This is where I tend to part ways with things like the Data Portability Working Group, where the discussions seem to go on endlessly about things like identity, privacy and truth. Now, those topics are interesting and worth talking about, but it's a bit like if your plumbing is leaking and you hire a philosopher, who comes around to your bathroom and tells you that since there is no such thing as truth or the external world and that we should be satisfied with our own subjective realities. Your toilet is blocked up, the bath is leaking and the shower burns your skin off. Perhaps you need a plumber, not a philosopher.

We've got FOAF, SIOC, SPARQL, hCard, XFN, OAuth and OpenID. Let's sit down and implement the damn things rather than pontificate on a mailing list about the nature of identity. If it doesn't work, or there are legal or privacy concerns, we fix them as they come up. That is what I was getting at in my question. The Data Portability group will be successful in my mind when it's stopped talking to itself and started writing user stories, test cases and code. If it doesn't start doing that, then the criticism that it's just a group for vendors to join to make themselves seem like they are doing something vaguely open is justified.