Entries by Max Howell

This unintentionally turned up on the Amarok blog, so apologies for that. The original post is here.

I warn you, this post is not one of my best. Nonetheless I have written it — I just wanted to document what I did this weekend.

I just finished importing the emails I sent when I went travelling, eighteen years of age, into WordPress. The trip is an interesting read. Certainly the first few are pretentious and make me cringe, but you can really see me grow up during the 72 days I was away. Especially when I get to Thailand and have to make it on my own.

I wrote the script to turn the HTML static page I had into an RSS feed to import into Wordpress using Ruby. That was a surprising pain in the arse. I kept slipping up due to not knowing the syntax quite right, or not understanding nuances of the weak typing and type conversion. Still it’s a language which feels promising. I need to do read some better reference code. The documentation on Ruby out there is rubbish. Suggestions welcome.

I was going to fix my activity-feed/life-feed thing. Especially since Steve at work is getting me all excited about DataPortability and Microformats and that. But I will have to rewrite chunks since for some reason my server no longer has the PHP SimpleXml module compiled in.

Looking through my epic eight years of blog entries I realise I must spend a whole weekend at some point just categorising and tagging it correctly. Categories are tough. I don’t like my current set. Work is easy enough to classify. But Detritus? Life? Should Life be about me or about Life-issues like philosophy? I feel the latter. I may make a category called Me. Boring crap like this can then be neatly avoided by those who wish it .

I have to say, classification and naming is a tough problem that I absolutely love to get right, but I find it as hard as the next person.

Tagging is easy but I want to be consistent with my tags. That’s the hell of web2.0 init? Tagging well. Tagging consistently. Mass tag editing is conveniently lacking too so you can’t easily change your mind later about naming decisions.

Also I find myself tagging posts eg. Amarok, but not categorising them as such. I’m using the two systems separately. The category is only Amarok if the whole post is about Amarok. But if it mentions Amarok, I tag it as such.

I checked up on my site backup solution. I have a cronjob running which does a backup everyday at 1:03am server time. On the last day of each month it makes a monthly backup. I created this system about six months ago after catastrophic data loss (MacFuse at fault grr!) But I’m not backing up my database. Which means I could lose all my blog entries. I don’t want to put my password in plaintext in the backup script though. Which made me wish we could do push notification across the Internet already. My backup script should be able to stick a dialog in my face once a week asking for the db password even though it’s on a completely different computer. Isn’t this just an extension of D-Bus?

I cleaned up my ClaimID somewhat. I also fixed the MicroID I was broadcasting on my blog to get the verified sticker at ClaimID. I cleaned up MyPlaxo even though I hardly use it. I feel Plaxo has some interesting possibilities in the future. That being cross site and application syncing. I think they actually will allow me to sync iCal with Facebook and Last.fm events within a few months. That is pretty awesome.

I also made my site OpenID 2.0 compliant, even sending the XRDS header. Although MyOpenID do all the actual authentication work. I added a Pavatar too. Which made me want to pick a new avatar image. I host my avatar at methylblue.com/avatar/, and more and more sites nowadays allow you to specify a URL rather than upload an image as your site avatar. Off the top of my head, ClaimID do and so do Blogger. Also a bunch of other sites nowdays allow you to use your Flickr avatar image, which is almost as good. At least it’s easy to change it in one place and reflect that change across your online presence.

And speaking of Blogger, they now let you login with an OpenID. Which is both fabulous and interesting. Does this mean Google will be rolling out OpenID authentication across their whole suite? I can only hope so.

I did most of this due to interest in what Steve is doing at mokele.co.uk. He’s combining a lot of new web technologies and making a system he calls OpenFriend. He was calling it OpenRelationship, but I think he wanted to be taken more seriously . He’s making an open Facebook essentially. You keep your data. Your friends keep there’s. Everyone authenticates their friends in a distributed social network. It’s fun and I want to be along for the ride.

I also edited this post enough times to realise I want inline editing for WordPress. Does that exist yet?

Comments here please.

I will prolly adapt Wocka for OSX.

Installwatch, should work on OSX. I’ve already almost compiled it. It requires itself to be loaded as a library before other libraries, and it then displaces all applicable file system functions with its own so it can write the install log. It’s clever

This is possible on OSX with some env variable, on Linux it’s LD_PRELOAD. If that doesn’t work though, I’ll compile my own version of make that has installwatch built in. Although this isn’t as handy since you are then restricted to installations that use make.

OSX packages either install via dmg and thus are self contained directories, or use the NeXT packaging format. This format writes installation logs to /Library/Receipts. This system is awesome an very unix, and further reason I love OSX. So all wocka has to do is create a receipt for installations.

No need to port pacman to OSX, or use the Fink APT db which I was considering.

I always wanted wocka for debian too. But don’t use it enough. So I’ll make wocka flexible so other people could do that if they choose.

I wrote a life-feed implementation, based on the idea by Jeremy Keith.

It’s here, and the source is here.

It needs databasing, as for instance, the recent music feed from Last.fm is just 10 tracks, which is usually about 40 minutes. And I want at least a week’s music history. Also the feeds are downloaded in background php processes, and thus the page the user gets is usually old. So some ajax magic to update the page would be good.

It’s all completely free, so take it modify it, and send me patches if you do good stuff. Subscribe to the projects category for updates.