Four months ago I blogged about the frustrations of using HTML5 video. I said at the time that I have to transcode the video into 4.1 different formats (mp4, ogv, webm, flv, and jpg). However, I was reminded today (at NC Dev Con) that Flash supports h.264, which is to say, MP4. So I don’t need to transcode to FLV for the Flash player fallback; I can just send the MP4 file to the flash player.  So I’ve updated the player on my site to drop the need for FLV. I’ve also switched my flash player from OSFLV to the nicer Flowplayer. Not that you would notice any of this, unless you’re using IE 6/7/8.

If you are viewing this blog post on my website (as opposed to an RSS aggregator), you will notice that the title of the post is written in a hand-written font. This is all through the magic of CSS3 web fonts. And in fact, the font that is being used is “Kip the Great”, which represents my handwriting as of early 1999. I’m limiting use to the title because the font is pretty hard to read. That’s only partly due to my bad handwriting—the font is also not well balanced (lowercase e and o, in particular, are too large), the characters don’t all align on the same baseline, and it is weighted so that it always looks bold (except for lower-case t, which I think I shrunk because it was originally too large).

Looking back at the font, I can see some subtle ways my handwriting has changed over the last 1.2 decades:

• I use an actual ampersand instead of that stupid backwards 3 thing

• I write capital D and F with two motions instead of one

• I write capital H with three motions instead of one

• I loop the tail of my lower-case G, like in cursive writing

• I put a line through my upper- and lower-case Z. (A habit I picked up in Calculus 2, where it became a real problem that my Zs were indistinguishable from my 2s. I also started adding a tail to my lower-case t around the same time, to distinguish it from a plus sign, but I only do that when I’m writing equations.)

That being said, there’s only about a 1% chance I’ll ever update it. Designing a font was one of the most tedious things I’ve ever done.

Longtime readers of this website probably remember1 that, when I created the current layout (nearly four years ago), I commented that the Homer icon beside the title of my posts was just a placeholder.  Well, I finally decided to do something about it.  I opted to just use my gravatar, as I started doing in the comments last year.  In case you’re curious, my current gravatar is from a doodle of one of The NiNjAS.  But if you were one of my true fans you would know that already.

I finally got around to implementing a search feature on this site.  If you search for something in the bar over on the right side of the page, you’ll get a search over all the blog posts, photo, and photo albums on this site.

I did all this with the Zend Framework’s PHP implementation of Lucene.  It seems to give good results very quickly, although there are some issues I’m having that are either misunderstandings or outright bugs.  Or maybe it’s because I’ve been up till 4am fiddling with this thing.  Tomorrow I’ll try to simplify my scenario and see if I still get the same results... if not, then I guess I need to figure out where I’ve missed something.  I’ve found that the way you design a search index is quite a bit different from the way you work with a relational database, mainly because you intentionally denormalize data for the sake of faster searching, when you wouldn’t do that for a relational database (at least not until you found that some particular JOIN or something was a huge bottleneck).

Anyway, I need to get some sleep.  If you try out my search box and notice something particularly odd, let me know.  (Well, something other than the styling, which isn’t fully presentable yet.)  Most of the problems I’m having are with more advanced queries that aren’t working the way the documentation claims they should, but it shouldn’t be anything most users would ever see.

Update:  I figured out what was wrong, it was due to using the default query parser, when I needed to construct the query from APIs.  After writing my own query processor, all is well.  More info about the specific problem I was having can be found on this Stack Overflow question, which pointed me in the right direction.

This post is to let you know about several small tweaks to this site that I’ve been working on lately, even though you probably don’t care at all. :)

One of the cooler things is that I’ve written some PHP code to programmatically generate gradient images.  If you’ve looked around the web you know that gradients are essential to modern web design, and I figure there’s no need to fire up Photoshop everytime I need one.  (Now if I can just write a glossy floor generator I’ll be totally web 2.0 compliant.)  You can view the gradient generator source code, if you’d like.  Of course, this kind of thing is so easy to do with PHP and a bit of 7th grade math that it’s almost not worth posting.  But I figured I’d share anyway.

Thus far I have put these gradients into action in two places on this site: as sexy new comment headers (as seen here, for example); and in the background of any picture in our photo album.

I can’t remember if I ever posted about this, but I wrote some Javascript a while back which is currently in use on the photos page, which scales the photo to fill your browser.  jQuery is awesome.  (And every modern browser1 can scale images without making them look grainy.)

I’ve also fixed the bug with the stored name/e-mail from adding a comment.  I have to apologize for the accidental breach of privacy, which would have exposed your e-mail address to other visitors to the site for up to an hour after your visit.  Now the name/e-mail fields are filled in by Javascript, so they are not cached server-side.

Another small change is that timestamps on posts and comments are now converted to your local timezone.  You can still hover over the timestamp to see an ISO-8601 timestamp, part of the datetime microformat I adopted when I added hAtom support.  Speaking of which, I finally found a way to validate hAtom: there is a site, transformr.co.uk, which will take a URL to a page supporting hAtom, and it will generate a true atom feed for it.  Here is mine.  I’m still not sure who would benefit from that though.  If you know enough to use the hAtom feed, then you probably know enough to click the little feed icon in the address bar too.  Oh well, it’s there if you want it.

I experimented with using Feedburner to manage the feeds for this site, but I haven’t been pleased with the statistics they provide.  They seem to be inaccurate, and I think the problem is that Feedburner is geared towards sites much larger than mine.  So I’ve decided to revert to hosting my own feeds.

So.  If you subscribed to a feed from this site in the last three months, you may need to resubscribe sometime next week in order to keep getting updates.  If you’re not sure which feed you’re subscribed to, you should start getting notifications sometime next week that the feed is dead (if you’re subscribed to the Feedburner feed).  If you don’t get any notifications, and you keep getting my blog posts, then you don’t need to do anything.

If you have no idea what I’m talking about then none of this applies to you.

It’s been a while since I implemented a spammer’s honeypot here on Vacant Nebula.  It has been extremely effective, so much so that I disabled the captcha.  All I do is put a hidden form before any blog posts are displayed.  Humans never see it, but spambots all see it, and apparently they are configured to submit spam to the first form on the page.  In fact the only spam that has gotten through in the last year has been spam that submitted to all forms on the page, not just the first one.  (I think this just happened once though.)

Fast-forward to a few days ago, I noticed that the excerpt of a page that Google shows displays the hidden comment submission form.  This doesn’t particularly matter, but I’d prefer it not be there.  So I added a check on useragent, and if it appears to be a search engine bot the honeypot is not displayed.  Well apparently spammers use a two-step process.  First they scan for blogs with forms while pretending to be googlebot.  Then they submit to those forms pretending to be a normal user’s browser (usually IE 5.5).

I know this because I got about fifty spam comments in the last two days.  If they were scanning the page with user agent reported as IE 5.5, they would have still seen the honeypot.  But the comments were submitted with user agent of IE 5.5.  Anyway, I’ve gone back to printing the honeypot for everyone, but only for the homepage.  Any permalink pages will not have the honeypot.  I’m pretty sure spammers don’t bother to go to the permalink pages, and search bots should only be indexing the permalinks.  Hopefully, both problems are solved.  If not, I’ll have to go back to a more fragile solution (something requiring Javascript, something requiring cookies, or even reinstating captchas).

Or maybe the spammers were just trying to wish me a happy twenty-seventh birthday by flooding my site with links to porn.

Last night I marked another item off my mental list of things to do with this website, by supporting Gravatars for the comments.  Since I haven’t been requesting e-mail addresses in the past, none of the old comments will show Gravatars.  If you’re a frequent commenter on this site and you want me to retroactively apply your e-mail address to any comments you have left on this page send me an e-mail telling me what e-mail address you use, or just leave a comment for this post with the e-mail field filled in and I’ll use that address.

I’ve come up with a very simple solution to the problem of browser-cached CSS files.  What I mean by this is: when you update the CSS which manages your website’s presentation, it will take a while before some visitors actually see those changes.  The reason, of course, is that browsers (this is at least true of IE and Firefox) will cache CSS files pretty aggressively, without checking very often to see if they have been updated.  Usually refreshing the page will solve this, but most visitors aren’t going to care enough to do this.  Meanwhile, your site will look pretty broken to them (especially if you’ve done something like styled a list so that it looks like a horizontal toolbar instead of a bulleted list, for example).

So here’s the very simple solution.  Add the following rule to your root .htaccess file:

      RewriteRule ^(.*)\.[\d]\.css$ $1.css [L]

I’m assuming that you have a common include file or template or something which prints things like the page header.  If so, whenever you update your CSS file (say, style.css), you update the link tag in your header to use style.0.css.  This will look to the browser like it is a different file from style.css, so it will download it again.  But Apache is really just loading the same CSS file through the magic of URL rewriting—you’re just ensuring that the user picks up your recent changes.  You can repeat the process the next time you tweak your CSS, just change the header to style.1.css and so on.

This is a quick update to say two things-

1. Two new sets of baby pictures are up.  They are both under the “Emma Leigh arrives” photo set.  I have to say that photos from Wal-Mart turned out much better than I had anticipated.  We even got full-resolution unaltered digital images on CD, along with full rights to reproduce and redistribute, something I really wish we had with our wedding photos.

2. You’ve probably noticed the site looks a little different.  If it looks broken, try to refresh (browsers tend to cache CSS files very liberally).  I finally implemented a tag cloud, and I’ve made one pass through all my blog posts to add tags to them.  Let me know what you think about it.