I’ve been having some really weird problems with my mime types lately, which has put a hamper on some of the RSS and standards compliance stuff I want to do here. Every document generated by PHP was being sent with a Content-Type: text/html; encoding: iso-8859-1; regardless of whatever was being specified with header in the code. I’m still not entirely clear on what fixed it, but I believe it had something to do with no default content type being specified in the php.ini file, but a default encoding specified. This confused things somewhere and it’s all resolved now. There is the default content type of text/html which works for 99% of stuff, and the overrides now work for the rest. Now I can rest easy.

Most of the scripts I offer on this site are things I’ve written myself, for myself. As part of this living site they are constantly updated and tweaked, yet sometimes I don’t update the scripts section with what’s running live here, usually because I tell myself that the code is so ugly that no one would ever want to use it. Well . . . it works. I’ve decided to put the updated autop code up for all to use. Impetus for the change came almost two months ago from Phil Ringnalda, though my implementation is pretty different from the one he suggested in his post.

Basically what the update does is make the paragraph code ‘smarter’ to deal with lines that end with block-level tags, therefore not creating extra spaces and invalidating your HTML. Try it out! I’ve left the simpler version there for people (like me in some applications) that don’t need the extra code at all.

Sometimes when reading what should be a very dry article you see an unintentional typo or some similar snafu that makes you spurt Dr. Pepper. Sometimes I wish my keyboard had the equivalent of a sneeze guard. From Tuning MySQL Server 4.0 Query Cache:

You may increase or degrease query_cache_size to find the value, which provides best performance for you.

So am I going to need one of these if I want my database to be tweaked to perfection? All joking aside, the query cache in the 4.0x series is, for practical web applications, one of the best improvements thus far. Though it annoyed me back in the 4.02 days when they changed the syntax for the configuration options. Why? No good reason. They’re on 4.09 now and things couldn’t be smoother though. I know of several major sites that run MySQL 4 as their main backend.

I’ve been noticing now with the fancy caps watermark that most of my posts begin with the letter I, and so I wondered how many of my posts were like this, so I put together a little query like so:

mysql> SELECT DISTINCT LEFT(post_content, 1) AS letter, COUNT(*) as count FROM b2posts GROUP BY letter ORDER BY count DESC;
+--------+-------+
| letter | count |
+--------+-------+
| < | 373 |
| I | 10 |
| T | 2 |
| A | 2 |
| W | 2 |
| C | 1 |
| S | 1 |
| M | 1 |
| O | 1 |
| Y | 1 |
+--------+-------+
10 rows in set (0.00 sec)

What does this tell me? Pretty much nothing except that before I came up with the autop function I started each post with a tag instead of text. Maybe it’s time to go back and clean up some old content.

I’m not sure how else to describe it, but if you’re browsing this with a modern browser you should be seeing the first letter of the paragraph of a post in a large serif behind the text. This is something I’ve been thinking about doing for a while, but never got around to it. CSS is a beautiful thing. It can think of a few tweaks for it, but basically the hard part is over. I’m still debating over what font to use behind the text, but I have all of the letter images (which are PNG by the way) generated through a neat PHP script so changing the font or color is a trivial issue.

Let me know what you think! Font suggestions? Too dark? Too big? Oh, and thanks to Rebecca for helping me out with the initial letters. She’s a life-saver.

I believe that in the weblog medium, nothing should come between the world and your words. Ideally you should be able to type a block of text, and it will be presented in a typographically sound, semantically meaningful manner without any intervention from you, the blogger. Unfortunately, everyone is caught up in presentation. Personally I’m very familiar with this because my aesthetic compels me to do things like curl my quotes, define every single acronym, and use paragraphs instead of multiple break tags, but this is a barrier to publishing. Before I had the curly quotes script, I would manually type in the HTML entities every time I used an apostrophe or quote, which as you can imagine can be very tedious. That’s been taken care of, and the code has even been integrated into Cafelog, the classiest weblogging software out there. I’ve addressed the paragraph problem in a rather superficial way that needs some looking at, and today I decided to take a look at the acronym problem. I’ve got it working just fine, I just need to hash out a little code that sorts an array based on the length of its key, and then I’ll put it up on the scripts part of the site. After this I want to clean up the paragraph code to deal with block-level tags, and then maybe port this BlogTimes thing I’ve been seeing around to PHP.

Update: Darn recursive acronyms! How ironic that the acronym PHP is messing up my script!

Update: It’s now online.