Disjointed Kitsch

I recently had a problem with the spam filtering software that is part of my hosting package at eapps. While it’s never “really” worked since I’ve had the account, it did a good job tagging spam and keeping it out of my mail box. Several months ago, eapps moved some of the servers and since then there have been a number issues related to the configuration of the server and server programs that have had to be fixed.

For the most part I have fixed these items on my own, despite reporting them to the “ticketing” system at eapps. The typical customer service response is to lay the blame at my feet and say “well, you must have done something”. If I were not an internet software developer by trade, that might work. I know how a Linux system works since I’ve built a number of them from the ground up and oversee a server farm of nearly 100 Solaris/Linux servers. I spend my day solving problems on them. However, I pay for the hosting at eapps and expect to be treated with a bit more respect than I’ve been getting. Even if I didn’t know what I was doing, I’d still expect my problems to be taken seriously and looked at from the point of view of “the customer is right”. Even if it ends up that I did something, which I know in this case I didn’t, I’m still paying them to investigate it and tell me it’s my fault.

Ya, I could just re-install since it’s probably a config issue anyway, but it’s the principle of the thing. If I’m going to pay to be called a moron then at least I’d like them to prove to me that I’m a moron.

I’ve spend the better part of this weekend setting up my new Intel Mac Mini. I opted for the Core Duo model with the 80GB ATA drive and the CD burner.

I managed to get my hands on a SyncMaster 710 MP flat panel monitor with TV tuner. The mini will live in the kitchen so we can use the monitor as a TV too. About $300 from Staples after mail in rebate.

I wanted the install to have as few wires as possible and since the mini comes with an AirPort card installed the networking part was taken care of. It also comes installed with Bluetooth so a Bluetooth enabled keyboard and mouse were in order. I didn’t want to jump to Apple’s keyboard right away without investigating other Mac keyboard options. Turns out, there are precious few designed specfically for the Mac. Sooo.. looks like it’ll be Apple’s.

I’m not going with Apple’s Bluetooth mouse, however. I’ve gotten used to a two-button mouse with scroll wheel and I’m lothe to give it up and go back to a single button. I would have taken Apple’s Mighty Mouse if it was wireless. I’m currently considering a MacAlly Bluetooth mouse that’s not out yet. It’s rechargable and is 2-button with a scroll wheel and it’s under $50. Yes, the recharder will bring in a wire, but I won’t be replacing 2 AAA batteries. For now I’m using a USB keyboard and mouse.

I orderd 2 GB of of memory for the unit today. Apple wanted around $300 for that. I got the same spec memory for $148. Hopefully I won’t have to call in a Apple Tech to get it installed.

When I first read about “Ajax” I though: This is what we’ve been waiting for to start the next generation of web based applications. I saw the promise of ultra-rich web content delivery using a variety of XML sources being assembled by the browser using JavaScript and async-HTTP requests.

Alas we’re still not quite at that nirvana stage yet. While “most” of the technology is in place, some of the core features that would really make this technology fly are missing (or at least I’ve not located their API yet).

The XmlHttpRequest/Response objects work fine, but if we’re going to allow the browser to parse XML and display content, the XML parser on the client side (i.e. in the browser) has got to become much better. The DOM API stinks!! Sure - it’ll work, but at the expense of making you crazy. Sure, JavaScript 1.6 fills this need by making XML a first class object, but so far the release candidate implementation lacks XPath support. What’s a “future looking” developer to do?

At this point it seems best to generate very simple XML structures on the server. To “pre-assemble” the XML for best parsing by the browser instead of having XML that represents the best “data structure”. Creating server-side processes that deliver simple XML to the browser from original, complex XML seems to be the easiest path. Even better would be to just deliver the HTML and leave all the XML processing on the server.