Squashing bugs and historical context lost

A few days ago I was going through the error log for the webserver and noticing the rather large number of Internal Server Errors I was getting for this site [1].

That was easy enough to solve: my software wasn't handing not-found pages correctly; it was aware that the files weren't there, it just wasn't bothering to send back the proper response to the webserver. I hadn't bothered to actually write that part of the code.

Mark [2] had noticed. So had about a dozen robots [3] (who so far had made about 700 failed requests this month).

I fixed the problem (although the error pages themselves are rather spartan) and then spent some time going through the error log seeing what problems I could fix. One typo on the template page probably accounted for half the failed requests; the rest were typos in individual entries and missing files from way back. For instance, an entry [4] over two years ago when I was still playing around with formating entries. It pointed to some mockups I had done that I had neglected to actually copy over when I went live. It's interesting to look at those [5] mockups [6] and compare the format I'm using today [7] (which just revealved another display bug that I'll need to get around fixing. Sigh). I actually liked the rounded corners I used originally but I used a hack to get them and (as the sidebar on those pages show) they don't always work like I intended them to, which is why I dropped the concept.

I eventually solved the problem I was having with Lynx by using CSS (Cascading Style Sheets) and because of the loss of historical context [8] for that entry [9], the opening sentance now makes absolutely no sense whatsoever. Of course this is the layout you are looking at right now.

[1] https://boston.conman.org/

[2] http://www.conman.org/people/myg/

[3] http://www.robotstxt.org/wc/robots.html

[4] /boston/2000/02/12.1

[5] /boston/2000/02/12/bd2.html

[6] /boston/2000/02/12/bd3.html

[7] /boston/2000/03/01-02/21

[8] /boston/2002/07/23.1

[9] /boston/2000/02/12.1

Gemini Mention this post

Contact the author