Tags, Taxonomies, and Categories

Most people are familiar with categories in web-based thingies. Look at Yahoo! or the Open Directory Project, or applications like Movable Type. They have hierarchical categories, where you create a category, then create categories within those categories, and so on. When you do it yourself in Movable Type, chances are you’re just guessing at how to best organize things based on your own experience and data, which is fine, it’s your view of things. In theory though, Yahoo! or the Open Directory Project have some sort of “well defined” categories created by experts, or some such thing. This is also fine… The problems come when you want to have a similar category set, so that what you call “Hardware” (Computers?) is what someone else calls “Hardware” (Hammers?) right? Syndic8 uses the Open Directory Project (DMOZ) categories, as well as some others.

Syndic8 uses multiple sets of categories, which is really more like taxonomies. At least by my understanding of taxonomies. I’ve always though of categories as sort of the “common” or “average” way to classify things, while taxonomies are more of the expert’s view, where more thought goes into it. That might not be the case, but I’ve viewed it that way. I think developing taxonomies as being more involved, since you can have multiple taxonomies that are completely separate from each other. For instance, you might have a taxonomy that works like your average category set, but them might have a taxonomy that works like a rating system, with G, PG, and NSFW categories. Taxonomies seem to allow another way to organize and classify data. (Drupal has a really nice taxonomy system.) I’ve often wished iTunes allowed for multiple taxonomies. I’d like to have a rating system for songs that had “kid-safe” and “explicit”, or perhaps “silly” and “serious”. Multiple classifications if you will… Perhaps TuneTags can provide that.

TuneTags seems to take the tags approach, like or flickr. Tags are a bit different, and somewhat like keywords, in fact they may be keywords, but in a simplified way, which makes them more powerful. Tags appear to be a single word, all lowercase, and you add multiple tags by just space separating them. This is brilliant in it’s simplicity, and can be powerful if presented properly. shows you the most active tags in the system, and while entering new data into the system, shows you all the tags you’ve previously used, with a count of usage. This seems like the fastest way of categorizing things, and with the ability to get a list of the most active tags, it’s easy to align your categorization with others, if you so desire. (John Udell is doing some interesting things using tags and flickr also provides a list of the most popular tags, which, just like, really moves the whole idea of “social software” forward, by making things easy and powerful. Those two extra things, providing a list of the most used/active/popular tags, and the display of your previously used tags are key to making things work – I’ve used systems where there is a field to enter keywords, but without those two things, it falls short of accomplishing what it really could accomplish.

This is all just off the top of my head, so if you think I got anything wrong, or missed something, let me know


Mac OS X Distro?

On LugRadio they sometimes talk about the role of a vendor/organization in creating Linux distributions. There’s Debian, Fedora, Mandrake, and others, and what sets them apart (to some degree) is what they include in their distribution.

Now, when it comes to Windows, Microsoft determines what goes into it, as it should be, but I think that while Microsoft is still suffering from the Not Invented Here syndrome, it’s nice to see Apple leave the Not Invented Here stuff in their past.

I know, there are plenty of things Apple sticks into Mac OS X that are invented at Apple, but those are often the good bits (iTunes, iPhoto, iMovie) that help define the “Mac experience.”

The most recent delightful (to me) outside thingy being added to Mac OS X is SQLite:

Another feature of Tiger is SQLite support, which incorporates the SQLite library into the system to provide an embeddable, zero-configuration SQL database engine for applications.

Developers who link their applications to this library can access SQL databases without running a separate relational database management system (RDBMS) process. The library can also be used to create local database files and manage the tables and records in that file.

Apple chose to incorporate a nice little application/framework that could potentially make some developers lives much easier. This seems to be something that keeps happening. While Apple is incrementally making small improvements, Microsoft continues to tout it’s master plan for a Brave New World of Windows codenamed Longhorn, that will completely revolutionize how you use a computer.

If you want to revolutionize how you use a computer, try some little bits of open-source here and there, and eventually, you’ll be all “revolutionized” and stuff.

Improvement can come in little steps, not just great leaps…

(I think we should start calling Longhorn “Copland ][” eh?)



I don’t have one of those fancy wireless wifi networks at home, but I do have a hub, and a router, and 4 (plugged in) computers in the basement, and 2 ethernet cables running through the house. So due to the fact that we had a small flood, and were forced to rebuild the office, we figured it was a good time to label each ethernet cable, just like the big guys do.

So now each cable has a sticker on each end, and on each side of the sticker is a number, so that we can easily trace the ethernet cable from the computer, to the hub (or router) and figure out what is or isn’t working…

Of course we really should have matched the numbers to the respective port numbers of the hub (or router) but for now some simple lookup chart taped to the wall should do. (It works for our DNS, I mean, host files…)

We also wrangled the power cords, USB cables, audio connections, and on and on…

If there’s another flood, we’re sure to be even more organized.


InDesign and Unicode

My contribution to the web today is to let people who do not have a copy of Adobe InDesign know that if someone is sending you an InDesign tagged text file, tell them to choose Unicode in the Tags Export Options dialog box when they do the export, and not ASCII, unless you can somehow guarantee that the text going back into InDesign will really be ASCII, which seems unlikely in many cases… (See every discussion on Unicode in the last year for background…)

Since I found myself trying to explain such concepts to someone who had a copy of InDesign, but no knowledge whatsoever of charatcer sets and the like, I’m just hoping this might help someone else… …And if not, at least it’s here for my own use! ;)


Uptime via Atom

I have (almost) nothing useful to contribute today…

So I present uptime via Atom, which will tell you how long my Mac has been on.

It’s like, totally experimental, and may change, or go away. Or something. I dunno. I was just looking for an excuse to use Perl, DBI, SQLite, and `uptime 2>&1`, while experimenting with Atom.

Atom is harder than RSS, and that’s alright. It kind of forces you to do things right. That is, if you consider the way Atom does things as the “right way” of doing things…

Anyway, I’ve been cranking out code, not words, so that’s all I’ve got. Later!