Still Not Perfect

I’ve noticed that Flickr has been having performance problems of late (mostly while I was developing a Flickr photo gallery plugin for WP). They mention them in FlickrBlog here. You can see their traffic graph for one of the problematic days there too. Flickr’s growing popularity is certainly stressing their systems. Also it appears that there was some sort of attack that contributed to the problem. So the problems appear to be due to network design decisions that couldn’t quite scale to meet Flickr’s new popularity, and malicious activity probably also due largely to their new popularity. Looks like it’s hard to be popular.

The Flickr engineers were able to respond reasonably fast and the issues seem to be mostly resolved by now. As popular as Flickr has become of late, I wonder what their bandwidth bill looks like.

Mozilla has launched the version 1.0 of their popular Firefox web browser. This release seems even faster and even more stable. Besides, Firefox is not vulnerable to all of the the exploits that plague Internet Explorer. This means browser hijack attacks and unwanted popups are a thing of the past (if you are a Firefox user [hint hint]).

The migration of users from IE to Firefox continues for the 5th straight month. How much longer until IE fires a volley back? In the meantime Mozilla seems to understand what users want much better than Microsoft.

Get your copy here, but prepare to be patient - their servers are getting hammered. Or, get the torrent.

(Update) Mirrors:
http://64.12.168.21/pub/mozilla.org/firefox/releases/1.0/
http://207.200.85.49/pub/mozilla.org/firefox/releases/1.0/
http://trillian.cc.gatech.edu/pub/mozilla.org/firefox/releases/
http://mozilla.osuosl.org/pub/mozilla.org/firefox/releases/1.0/
http://mozilla.ussg.indiana.edu/pub/mozilla.org/firefox/releases/

Some Wordpress plugin ideas I’ve been considering:

1. Server load indicator - Allow you to display a bar indicating how heavily loaded (load average, mem usage) your server is. This one is probably only easily done on a Linux server (/proc filesystem would be instrumental).
2. Visitor origins - Like this one just more customizable (can use a custom map image). The only problem with this idea is I have to get the geographic IP data from a source that probably will charge money. I’ve already investigated Maxmind’s GeoIP City and Geobytes GeoSelect.
3. Flickr photo gallery - I’ve already begun work on this one. I’m tying into Flickr’s API to get a list of a Flickr user’s photosets and photos. You can see my progress so far here. I still want to add human readable URL generation and Flickr comment integration before I feel it is ready for primetime.

Update (11/27/2004): After some thought I’ve come to the conclusion that the server load indicator really isn’t a WP plugin (it doesn’t interact with WP at all). Depending on how it was implemented the visitor origins one might not interact with WP either.

This morning I posted about some of my issues with Pictorialis II as a simple photo gallery manager. I decided to look into how PHP implements the imagecopyresized() and imagecopyresampled() functions. PHP basically just calls GD graphics library gdImageCopyResized() and gdImageCopyResampled() and passes the hard work off to GD (I guess I just didn’t remember it at 3:30AM this morning - then again I’m not sure I’ve ever looked at that particular part of the PHP source). So I downloaded the source for GD 2.0.31. It turns out my uneducated guess last night about the algorithm used for imagecopyresized() and imagecopyresampled() was basically right. The GD functions gdImageCopyResized() and gdImageCopyResampled() are in the file gd.c if anybody really wants to study them. This explains why PHP’s imagecopyresampled() is a bit slower than than imagecopyresized() but produces such prettier results. Imagecopyresampled() basically samples all of the pixels that imagecopyresized() would simply ignore in the resize and averages them together with the weighted colors of the pixels that imagecopyresized() would have directly copied into the new image.

I’ve been trying Pictorialis II out as a photo gallery. You can see the results of my experiment here (example taken offline - it was getting spammed to death and I’m not using it anyway). For the most part I am pleased with it, but I’ve got a couple gripes: Read the rest of this entry »

Firefox 1.0 RC1 was released today. It is not yet on the Mozilla homepage, but you can download it from their FTP site. It includes about 250 bug fixes. Keep in mind that this is just a testing release, so if you aren’t interested in helping test and reporting bugs I suggest you hold off for a few more weeks until the product is officially released (November 9?).

Word on the street is that it starts up faster than Firefox 1.0 PR and renders pages about 2x faster. I can confirm that it does seem to start up faster, but downloading pages seems to actually be slower - pages still render fast once you’ve downloaded them though. Also, most of my plugins seem to be broken for RC1, so if you depend on having your plugins working you might want to steer clear of this release. Afterall, it is for testing before the final release.

All in all I am still very impressed with Firefox. Thanks to all of the people at Mozilla for their hard work. And it looks like their hard work is paying off - they are gaining market share.

I love PHP. I guess that’s probably because my programming roots are in C/C++. However, you’d think that I would love Java equally then. The fact is, I don’t. Java just seems so restrictive to me (I know, some of you are going to say that’s a good thing, which I won’t totally disagree with). I love PHP because it is wonderfully flexible. It also has a wonderful toolset to go along with it. Things like the Smarty template engine and the PEAR extension repository. This flexibility and toolset makes rapid development in PHP extremely easy. Why is it then that PHP hasn’t penetrated the enterprise web application space? Here are a couple of reasons (and maybe a couple of enhancements to consider, PHP devs): Read the rest of this entry »