Notes on the History Block

• When it was written (completed). This line is marked with a large bullet that's really an at sign. The idea is that the line tells at what time the page was written.
• When it was extended. These lines are marked with small bullets. Most pages have none; a few have more than one.
• When it was released as incomplete. This line, if present, is also marked with a small bullet. It can be distinguished from any other lines with small bullets by the fact that it comes before the large bullet.
• When it was released as a structural node. This line, if present, is marked with a small x-bullet.
• When it was mentioned on a history page. On these lines the month and year text links back to the mentioning page.

Of course, most of those events apply only to essays, not referends, but referends do still have history blocks. The blocks won't have any lines marked with bullets, and may well be completely empty, but they're still there in principle.

Now let's look at a few examples and see what we can see.

• Number Maze was written in April 2004, and was also mentioned on the history page since it was one of the better essays for that month. It was mentioned again with no change in July 2004, then was extended twice, first in September 2004 and then again in November 2005.
• Physical Awareness was released as incomplete in June 2004, and wasn't completed until May 2006 (when it was also mentioned on the history page).
• The Rule of Correspondence was written in April 2001, and hasn't changed since. That's not a very exciting example, but it is a typical one—most pages have only a single line in the history block.

You've probably noticed that the bullets are also links. I want to say a few words about that, but first I need to explain how the history block evolved.

In the beginning, the index was just a list of backlinks. When I'd designed it, I had decided that links from history pages wouldn't produce backlinks (see Exceptions to the Backlink Rule), but even then I wasn't completely happy with the decision, and in July 2004 I changed my mind and put backlinks from history pages in a separate section underneath. That section was a primitive form of history block! The current form is different only in that (a) it has bullets, sometimes on lines that weren't there before, and (b) the lines are in chronological order, not alphabetical. In particular, the lines where the month and year text links back are really just the same old backlinks in disguise.

What about the other lines, the ones that weren't there before? How did I decide on a design for them? That was a highly constrained problem, actually. It would have been perverse to name the batch month and year and not provide some kind of link to the history page, but at the same time it would have been bad to make the month and year text into a link, because then there would have been no way to recognize the lines with real backlinks. The only solution was to make the bullets into links. I resisted it for a while because I didn't like the way the underlined bullets looked, but there really was no other choice. (And, by the way, the underlined bullets don't bother me as much now.)

I also didn't like that the bullets linked to the history pages, since that made the lines with real backlinks have two links to the same place, but in that case there was one other choice. For every history page there's also a details page, a.k.a. the “index for each batch” that I mentioned on the essays search page. I figured that a link to the details could count as some kind of link to the history page—and then, wow, I realized that that method had a beautiful side effect. A bullet indicates an essay status transition, but a details page is just a big list of links to essays with status transitions. So, for every bullet there's a link, and for every link there's a bullet that links back to it—in other words the bullets, too, are real backlinks!

As long as we're talking about backlinks, let's take a minute to review the backlink rule and its exceptions. In the strictest sense, everything I said in Exceptions to the Backlink Rule is still true. In particular, history pages do not produce backlinks. The index contains a list of backlinks, links on history pages don't produce entries in that list, Q.E.D. If, on the other hand, we let backlinks in the history block count as backlinks, then the exception for history pages goes away and a new exception that cuts in the other direction arises: details pages do produce backlinks.

Finally, I'd like to tell you about one small regret I have. Part of my original concept for urticator.net was that it would be timeless. I'd add some new essays occasionally, of course, but the new essays would fit together seamlessly with the old, and it wouldn't matter when anything was written. (I don't think I've explained that part of the concept before, though the start of Anonymity is close.) The history block doesn't really change any of that, but it does make time a more prominent feature of the site, and I regret that. On the other hand, it's pretty subdued, not too prominent. It's also useful. I know it shouldn't matter when anything was written, but sometimes I've still wanted to find out. And then there's the best part, which is that the history block makes it easy to associate from one essay to other essays that were written at the same time, and so adds a whole new dimension to the associative network! So, all in all, I'm quite pleased with it.