learnandteachonline.com

Doc Searls is listing all the cars he's ever owned.

Thanks, that reminds me that my OPML car roll would be a good file to use to understand the search feature of GrazrScript, whenever I finally get around to studying it.

I made it to demonstrate the inclusion properties of an OPML directory, and referred to it on this draft OPML Editor support site page, and in my ebook about OPML. (Also mentioned docnography in the book.)

When I link to the car roll I usually point to it on Dave's Worldoutline but it seems to be down at the moment. Here it is in a Grazr:

I was so immersed in OPML for a while that I burned out on it a little, but lately it's been popping up unbidden. Something made me think when I was driving to work the other day that it might be a good format for resumes, especially as they could be used after landing a job. A company could make a nice browsable qualifications database with OPML.

P.S. I had an Austin America, too. Never knew anybody else who owned one. They were pretty cute, a little like a Mini.

Jon Udell's latest interview on IT Conversations gets into Syndicated Oriented Architecture (SynOA, to distinguish it from plain ol' SOA).

Here's the .mp3 file of the talk with Rohit Khare, founder of KnowNow and software architect.

The first step, both evangelists say, is you gotta "RSSify" everything. Then tag your information to help other people find what you've found. Consider special problems of syndication in the enterprise.

My note: think beyond blogs when you think about RSS.

I registered amyloo.com. Should have done that a long time ago, but I probably wouldn't have done it if I hadn't been prompted to sign up for a cheap Windows hosting service. I wanted a sandbox for playing around with some Cold Fusion/SQL Server stuff. I'm experimenting with FDF. I'll probably narrate some of it, especially if I hear that anybody is interested. I'm figuring out how to output data entered into a web form in a PDF template.

I've had my sites on shared Linux accounts for a number of years recently, but have gone back and forth. I learned on Unix starting in '94, then reluctantly moved my stuff to Windows when I thought I should, went back to Linux when so much exciting LAMP stuff was coming out and I wasn't afraid of databases anymore, and now it looks like I'll have both for a while anyway.

We also have both at work, but my heart is with LAMP. I asked the guy at work who replaced me in the web department (after I moved to the publishing arm) where his heart was and he emphatically said "Windows." But he's alright.

So the datasbases show up publically on a page that looks like a Google search results page.

But now I've moved on to thinking what kind of data I'd want to put there. The contents of my Drupal aggregator and my blog posts don't need to be there to be seen online because they are right here in a more appropriate format. For people who care about a lot of traffic, I suppose that's one more place to have links in.

What would be more useful? Putting up all my podcast posts would be nice, because my RSS feed is set to show only the last 30 and there have been 81. But I could link to my page on some of the podcast directories that have somehow retained info on all episodes.

I haven't been able to determine if you can only put tables on Google Base, or if it's relational, and if it can be relational, how would a user look at crosstabs. That would be useful.

Good for stuff you don't want to make into a big deal

Even if you can only show flat tables, I can think of one category of data I might put on Google Base: informal reports. Here's what I mean. I think about the kinds of data I handle every day, and a lot of it has a reason to be online, like a shopping cart. The database exists to make products available for purchase online.

For data that's used online a lot, an organization will naturally want to invest in building or adapting an off-the-shelf app to display it.

Example. But let's say there's a need to share with a distributed group some data that isn't often output online, it's just a one-off need. An example might be a part of a database whose main mission is to deliver a buyers guide directory.

The bulk of the data is the vendor contact info and categories, but there's also one little table that manages banner ads. It contains rotation ratios, advertiser names, URLs of the advertiser's sites and URLs of the banners' image files.

An ad rep is doing an presentation and wants to show the prospect a little slide show of all the banners ever run in the buyers guide. As persuasive as the rep probably is, she's not going to get anywhere if she wants an online app built for one presentation. But! If that one little table were uploaded to Google, she'd have a page to go to in her pitch, a nice list of links to the banner images, where she can click and show.

Bottom line, I can see gBase used in situations where you don't want to invest a lot of time or money in putting data online. You think?

Or maybe its market is individuals, not organizations. Could be it ends up getting used more for party invitation lists or soccer team rosters than for banner ad URLs.

Tell me what I'm not seeing. I'm sure it's a lot. I haven't read any commentary about it because I wanted to look at it on my own first in a real life context. Only now I'm tired of thinking about it and probably won't go read what others are saying. Not right now anyway.

The file is processing even though there were four errors: three "Attribute was incorrectly specified;" and one "There were invalid fields in the item."

So now I'm waiting. While I'm waiting I uploaded the RSS file for this blog. See how that goes. It's in "Processing - will publish soon" mode now too. Google didn't find any errors in that file.

I'm just not getting anywhere at all with the XML that MySQL spits out. Google Base doesn't like it, it doesn't validate. Trying tab-separated text now. MySQL only does CSV, but if you use the CSV for Excel type, Excel will save it as TSV and leave its hands off the commas you want to keep.

But it does give debugging info

Line # 2
Error Error parsing/validating XML feed - help
Bad data Content is not allowed in prolog.

Some folks say Google new online database service is lame and others think it may be intriguing. Let's see.

I signed up for bulk upload permission and sent Google an XML file that I can't view for an hour. That part is kind of a bummer. The waiting part.

Here's what I uploaded. It's a dump from my MySQL Drupal table that stores all the information output by my blog's built-in aggregator, and includes headlines, a text excerpt and URL for my old blog, my OPML blog, my friend Hil's Spirits Dancing blog and Scripting News.

I have appointments all morning starting almost now, so you may see the result before I do. I think it's supposed to appear here.

Microsoft announced a supercharged SQL Server 2005 today. But there will also be a stripped down free version. David Wilkinson, a fellow OPML blogger, was just saying the other day they ought to do that to compete with LAMP.

I don't work on the website every day anymore for my job, but when I did I pushed LAMP stuff whenever I could. I could install it myself without the help of real programmers, which meant I could fix it when there were problems. We also use SQL Server with Cold Fusion on the IIS server and I can't be nearly so self-sufficient with that setup, plus it just isn't reliable. I can't think of the last time we had to restart the Linux server because something was stuffed up. It might be years.

Later... No, wait a second. I didn't read that Microsoft story carefully enough. It's only a free trial. That's altogether different. I guess MS is a lot more interested in competing with Oracle than with the grassroots crowd. If B2B is the way they want to go, it seems like they ought to drop the live.com "For the People" rap. I don't really think you can do both and have your heart in either.