Both you guys are right about the awkward navigation through the TOC.

One of my early tasks (still ahead) in this latest project is to add searching to the TOC, and #1 (and easiest) is by a range of dates. The search would pull just the threads that have activity (or were started) within that range, and form a virtual TOC (a filtered TOC subset) through which you can navigate back and forth. There may still be a performance hit if older HTML versions of thread TOC trees (what you see appended to each message on the tree, and what's concatenated to form the main TOC) have been purged from the cache and have to be recreated.

Hmmm, got an idea: I can outsource the CPU cycles to you guys, by sending the thread data in XML format out to your browsers where it'll get transformed to TOC HTML instead of doing the work on my server. You wanna see an old TOC, you can render it on your nickel! Yeah, that's the ticket!

The term "TOC search" is a private mental artifact, and I should explain that I'm thinking that the best way to return search results on a threaded BBS is through a "virtual TOC" composed of all the threads containing messages that satisfy the search criteria. That way, you'd get the messages in context, represented as part of a conversation rather than an isolated fragment.

Aw, what the heck, here's the URL to the prototype. (Of the two test boards, "Site" has is more interesting because it has a bunch of old Zone GeekSpeak messages to exercize the TOC). This talk of changes to the TOC will make more sense when you can see where I'm going with the TOC, which is where I put all my efforts in this first prototype. It uses AJAX, which is a trendy buzzword acronym for Asynchronous JavaScript And XML, which boils down to using script to hold a private conversation with the server in the background, while the page stays in front of you peacefully minding its own business.

In this case, script on the TOC page contacts the server in the bckground requesting a particular TOC page (the numbers you guys have hacked out of the URLs). The server gets it or generates it and sends it back to the script, which then displays it for you. When you click Earlier the script requests the N-1 TOC page, and when it receives it, erases the current TOC and replaces it with the new one. The script caches TOC pages (until you click Reload) so that once you've downloaded a page, you can go back to it instantly; there's even an index of pages you've downloaded so you can click on a number and go to it directly.

Like I said, this is a first crude prototype, and the upgraded TOC is about all of interest to professional BBSers like Zoners (I think in terms of designing something that people will be using several times a day, every day). There's a second prototype in the works right behind this one that should be more impressive, with the...nah, no spoilers, I want unprejudiced reactions.