I’ve noticed that Jason Kottke has found Joggle (remaindered links).

Since Blogger doesn’t support comments, I don’t have the time to set up MovableType, Textpattern or anything similar and Joggle’s not live yet, the easiest way to find out more/ask questions is via the Joggle Forum on sourceforge (yes we sold our souls, no we don’t care).

It’s possibly also worth noting that despite the claims of the sourceforge CVS summary page, code does actually exist, which is browsable here.

Blogger breaks my HTML.

My template is well-formed, and valid XHTML.

The source of my blog entries is well-formed and valid XHTML.

The output of my entry inserted into the template is broken XHTML.

Somewhere along the line, BloggerTripod converts my opening and closing body tags into BODY tags, </b> tags into </B> tags and my <blockquote> tags into <Blockquote> tags, which means my page is no longer valid (since it leaves the case of the opening tags alone, giving me a million and one “invalid nesting” errors), which, pretty obviously, I’d really like it to be.

I was thinking about this ‘speed of re-rendering’ problem last night, whilst deprived of computer by my significantly better half, and came to the conclusion that it’s really not that much of a problem for Joggle anyway.

Joggle only renders pages when they’re requested, and deletes ones that haven’t been viewed for x number of days in the background. Posting a new story to your blog, or a forum site will trigger a re-render on that page, so it will pick up the template change almost immediately (in fact, it’ll pick it up when the template first gets changed, but that’s by-the-by), so all we need to do is delete the other pages, and the next time they get viewed they’ll be re-rendered using the new template.

The only problem I can really see with this is if someone has an archive page that gets a lot of hits (interesting article, review, whatever), and a lot of people simultaneously request it whilst it’s in the process of being re-rendered.

What to do?

I really want to give Joggle baked files, but having read Jason Kottke’s comments about how quickly Movable Type re-rendered serveral thousand files in about thirty seconds, I’m now a little intimidated about this.

Movable Type rebuilt the entire site – 2403 individual archive pages (including 1432 comments), 59 category archive pages, and almost 5 years-worth of monthly archive pages – in about 30 seconds. Fast.

Fast like a monkey going for a banana, but you host your MT site on your own server, whereas Joggle follows the Blogger model of a centralised server, and is likely to suffer from it in much the same way i.e. Blogger takes about five minutes to recognise a publish command and make the change visible.

Of course, Blogger serves hundreds of thousands of active bloggers, and so possibly has some good queuing system to handle thousands of simutaneous publish requests, but that doesn’t stop the delay being annoying. For the time being, a publish request will probably spawn a thread which publishes the page, but should the number start spiralling out of control, a queuing system it’ll have to be.

Just like everyone else in the world is probably now saying, ESPN have redesigned to use an almost all CSS layout.

This is good news.

Firstly because it’s another big site turning to use CSS-only layout (see: Wired, W3C, Netscape devedge), but secondly, and perhaps more importantly, because it’s a Microsoft site, and that means they must be confident that the majority of the visitors to the site are going to have proper CSS-compliant browsers. Come in Netscape 4, your time is up.

With luck this could mean that tools like Frontpage might start producing ValidHTML and making people use CSS instead of tables, font tags, and all other such horrors.

As Zeldman points out, the ESPN site still has it’s faults, but it’s a massive boost.

James Snell writes about WhoAmI Self Hosted Identities, and proposes a protocol to query a person’s FOAF information through Jabber, and explains how it can be used for by web applications and sites to discover your preferences and interests, without you having to enter them again and again, or have the sites working them out over time (much as sites like Amazon currently do).

The problem appears early on (as they so often do), that this emergent data would need to be stored in some kind of standard fashion so that it’s easily accessible (the information could be stored on your local machine for instance, or a webserver you have access to; you have control over who can see what information and how much, etc.). FOAF is easy, there are fixed fields you can specify, but who can determine what kinds of data systems that use emergent properties will need, or even generate? Some ‘catch-all’ isn’t good enough, and it would take companies working together to surmount this. Something I’m sure that hasn’t escaped the attention of the MS Passport and the Liberty Alliance.

UPDATE: Obviously the answer lies in the semantic web. RDF is the answer to world hunger.

This might be neat.

Joggle auto-annotation:

  • maintain a list of a) words and b) links that are automatically replaced when they feature in a blog entry (e.g. ‘Google’ gets replaced with ‘<a href=”http://www.google.com” title=”Google” alt=”Link to Google”>Google</a>’ )
  • provide an “annotate” link which allows you to review the words of the entry you’re making and define new annotation definitions.
  • a public and private annotation list would be nice, with the public list being a subset of the group of all users’ annotation lists. This should be done automatically, so that when a word reaches a critical level, it gets transferred into the public list

Bookmarks are useless – what’s the point of having a series of bookmarks that only gets stored on one machine? At work this is no problem for me, a fast connection and Google and I can find anything anyway, but at home I’m on dial-up, this means bookmarks. Of course, all of my bookmarks are made when at work.

So what I want to see is a central resource for storing your bookmarks on the web (not a new idea), integrated and accessible from within the browser. For the sake of argument call them “Netmarks”. For the sake of argument (and ease), make your browser Gecko-based. And now add some items to the “bookmarks” menu which makes XML-RPC calls to a server storing your current URL, as well as storing it as a bookmark (Netmark this page, Manage Netmarks, etc.).

These should all be retrievable from any Gecko-based Mozilla-based application (where’s the XRE when you need it?) based on a user-id and probably a password.

This is all possible with what we’ve got right now. Instant global bookmarks with no extra effort from the end user.

Now someone just has to write it.