Keith Devens .com

If you want to avoid polluting your namespace, a good bet is to use subdirectories for years. In my experience a year's worth of random stuff fits nicely in a single directory.

Thanks for the suggestion. That's a good idea. It's what the W3C does for many things. Unfortunately it doesn't work for pages I want to keep current. That is, I wouldn't want to have a URL like '/2001/about' for my about page (for example), since that implies the page is out of date.

I'd say, for things as site-general as an 'about' page, go ahead and put it at the root level.

For anything else, such as one-off experiments and random things, I use a /year/month/ scheme. And then, whenever I update those things, I tend to just upload the new version at a new date-based directory for some crude versioning.

For random experiments or temporary things, that's a good idea. But the reason I mentioned my wiki in the beginning of the post is because a lot of the pages there are the type of pages I'm considering. For examples, my wiki pages on jEdit, ProgrammerFonts, Calvinism, Presuppositionalism, Regular expressions and even Acno's Energizer.

Also, I mentioned Chris Langreiter's site because I admire it. Earl's site too. Both use Vanilla, Chris' software. Maybe I should call my namespace 'space' (but that's always irked me).

What about the last thing that we talked about?

The last thing we talked about was doing it like Wordpress does it, but as I pointed out before you seem to have misunderstood my question. My question wasn't about implementation, it was about site structure (i.e. what should my URLs be?).

Oh. I see. I have no idea of how the site is really implemented, but from you have told me it is all driven by a database.

If it were me, I would choose to put nothing at the top level. I would choose to do it like Google Directory and build a taxonomy.

Top/Computers/Python
Top/Computers/CSharp

etc.

I'd just create a subdirectory called 'stuff' or 'sundry' and throw all the randomness in there. A little bit more organisation within that for things like experiments might be useful.

Build a taxonomy! Nooooo!

In any case, don't you know, taxonomies are out, tags are in

I'd just create a subdirectory called 'stuff' or 'sundry' and throw all the randomness in there.

That's probably the best idea. I've actually already ripped out my wiki software and replaced it with the software that'll serve my "pages". If I ever want to promote a page to the root or vice versa I've already got a rename feature that'll serve a 301 (I've also got delete that'll serve a 410). Since all of my individual pages that exist are already under /wiki, I'm just going to leave it as my home for random pages. It's basically a private wiki then, which I'm ok with.

Also, my pages software will let me create URLs based on the year or year/month, so I still may use that for some things. Thanks for the input, all.

In any case, don't you know, taxonomies are out, tags are in.

That is like the "labels vs. folders" debate has cropped up in many discussions on "Outlook vs. Gmail". I believe that programmers like labels, but many end users still like folders.

Silly end users. When will they ever learn?

Well, that was said tongue-in-cheek, but in any case I'm not developing end-user software. However, on the web, tags do seem to win (Flickr and Gmail are certainly well-received end-user products).