It never fails... I sent out a few hundred resumes this week. There's a very prominent link to check out my recent community project (on one of the websites where this message is coming from).

And of course Murphy strikes. The site was down for a couple of days last week and it's currently running slow as molasses. Seems like everything breaks whenever one of my employment contacts is looking at it. I no longer have this running on my own server, so most of these things are out of my control. I'm sharing resources with maybe a hundred other folks. All of the optimizations I performed to make the data fly onto the page don't matter if my shared database is choked or one of the ISP's DNS servers takes a dive. 

Sigh...     

More Microsoft madness... 

So I downloaded the 'Release Candidate' of Microsoft Internet Explorer version 7 (that's what 'RC' stands for in the software world) to see if they fixed any of the long standing CSS quirks and anomolies.

The first thing to report is that the installation failed the first time and left my machine in an unworkable state. I had to re-install a couple of times to make things work somewhat normally again. This isn't a pleasant user experience, but of course I'm used to such things.

Second thing to report is that even though this isn't production code yet (meaning it hasn't officially been released), the installation offered no choices for continuing to run IE6 as an alternative. The existing browser was over-written. This is a bit of a problem for web developers who need to check their code on multiple browsers. In order to be able to check both IE6 and IE7 I would need to use two machines.

Now for the good news (and bad news). It seems that they have indeed fixed a lot of CSS quirks (and presumably Javascript/ECMA quirks and DOM quirks and ...). You'd think this was good, no?

No. Because until IE7 is widely deployed, the developer trying to write cross-browser apps needs to add yet another 'if' statement into every section of their code where cross-browser issues creep in. It is no longer enough to detect 'Gecko' or 'MSIE' and providing behaviour based on these possibilities (this is making a lot of simplifying assumptions about Opera, and IE/Mac, and the real old browsers). Now you've got to add yet another behaviour for the differences between MSIE7 and any other versions of MSIE which you've already got hacks for. 

You're thinking that this shouldn't be much of a change, and you're probably right - at least for one particular page. But consider that there are millions, if not hundreds of millions of pages out there written with browser specific hacks in place. Somebody spent a lot of time to make sure that they looked the same on IE and Firefox (and Opera and Lynx and Netscape, etc.), and now they're going to have to spend a lot of time fixing them all once again because they will all be broken.

The good news is that once IE7 does get widely deployed, you might eventually be able to write a web page that looks the same on Firefox and IE without a lot of special case code. The bad news is that it may take several years for IE7 to get widely deployed.   

My daughter starts fourth grade next week. Back from our family re-union which came on the heels of a beach trip which was scheduled (i.e. the spot reserved) in better times. Now perhaps I can get back to the important task of finding employment.

I'll probably give Norm a hand next week with band instrument rentals. He'll need the help - and perhaps give me a chance to say hi to some of my Sonica friends.

I've done some interesting stuff recently with the website software. A lot of consolidation of things which are similiar. These leads to better code, since similiar things aren't duplicated. Unfortunately it doesn't seem to help get anybody to take notice of what either the software or myself are capable of doing. Sounds like I need a concerted marketing campaign...

But let's take an example. Categories. Forums. Subscriptions, photo albums. These are all very different - n'est ce pas? Categories are collections of articles with the same tags. Forums are collections of articles. Subscriptions are collections of forums. Photo albums are collections of photos.

Whoa, hang on. They're all collections.

So I created a software class called 'containers' with an associated 'contents'. Contents are the things which go in containers. Together they create a collection. Of anything. I no longer have different database tables for all these different collections of things. I just have a bunch of containers that contain things.

Remember not too long ago when I mentioned that forums and weblogs and newsfeeds were all just collections of 'articles'?

Similiarly, a while back I created permission lists. Which are - containers. But I digress. Permission lists can be associated with anything. Photos, weblogs, forums, whatever. Until last week I just had access permissions. But what about write or post permission? It's the same thing. Just a list of permissions. 

So now you can not only create a private forum, you can also specify who can post to it. You can create a soapbox that only you can write to and everybody can read. But wait a minute... That's what a weblog is. So I've just created a new way to create weblogs - as forums. So why would anybody need a separate thing called a weblog? A: They don't. So eventually I will get rid of them as well. No I don't mean that I'll remove weblogs from the software. I'll just remove the management of things called weblogs from the software. It isn't needed.

Where is this all going? There's a lot of unneccessary code. At some point there will be another re-write. When that happens, a lot of the extra crap will be tossed into the bit bucket. What will be left? Containers. Things that go in containers. And perhaps 'people'. That's all. And it will still do everything the software does now. It will do everything you could ever imagine doing with a content management system. Want buddylists? Want MP3 collections? Private groups? Public groups? Taxonomies? Events? In every other content management system, these are all new database tables, new software classes. More code and data to manage in some unique way. But it doesn't have to be that way.

They're just boxes of stuff.   

I never wrote the story about my DSL saga back in 2001 because frankly it would've taken a book to describe the whole story. It took a court order to get my twisted pair back from Earthlink after Flashcom went bankrupt and gave them my account. Even though I signed up with Earthlink even though I could not use their service but did so just so I could pay a $100 cancellation fee so they would disconnect my line and give it back. It was the only way to get my circuit disconnected. Even after all the gyrations, they never disconnected it and the circuit ended up in legal limbo and couldn't be used by anybody.

Fast forward to 2006. The latest chapter is no exception. Sometime last week my DSL line went down. The connection just vanished. I've spent a week and numerous service calls to get it back. But I don't have it back. I only have something that works for now. I don't believe I can ever get my service back again the way it was. 

In the meantime, I've broken into my DSL modem, and changed just about every setting on it (hundreds) in a few thousand different combinations.  I've got a new DSL modem. This is the fourth modem I've been through this week. I've got a new DSL circuit from AT&T.

What don't I have? One stinkin' static IP. You'd think this would be easy, but in 2006, static IP's are a thing of the past. The AT&T call center in India can't give me one. Even though I pay about $50 a month extra above the base DSL rate so I can get five of them. They wouldn't give me just one. But that was Pacific Bell. Now it's AT&T. AT&T doesn't like giving out static IP's for any amount of money. The local installer can't get one for me. Neither can his service office and neither can his boss. All they can get me is another dynamic IP, which I can't use (but I'll use right now - it beats no IP at all). And they're still charging me $50 extra each month for the five static IP's which they refuse to sell  me.

The worst part of it is that they are telling me that I've got a static IP. I don't, and I've wasted a lot of time trying to prove it to them - but there's nobody to listen; and the two or three people left in the company who even know what a static IP is can't get past internal politics to make it right for me and give me what I paid for. Nobody cares.

Welcome to the new AT&T. Same as the old AT&T. But be warned - if you're an AT&T residential customer with static IP's, you're gonna' lose 'em. It is no longer a supported configuration. They will turn them off whenever they feel like it, and you'll never get 'em back, and there's jack you can do about it.