I’ve been toying slightly with the design recently, and I just noticed today that it’s pretty obvious that almost any browsing I do happens in Firefox. The site (well top menu anyway) looks awful in Internet Explorer – auch. Better fix that some time in 2005.

It’s been awhile since Stuart Langridge released some cool javascript which allows you t do client-side sorting of tables in an unobtrusive manner. Soon after Andy Edmonds released a merge with a function to made alternating row coloring he had made. Now Caspar has done a little magic and added (cookie) memory to the script – so it remember you most recent sort from visit to visit. The code is in my pre-alpha markup area.

So I thought I did a decent job of cleaning up and recovering the site after the evil crash a few months ago. Browsing through the 404 (file-not-found) logs sadly tells a different story. Damn. I was hoping to be fully recovered by New Years Eve. That’s probably a little too ambitious.

An odd discovery today – if you don’t specify a protocol in your links, the browser apparently suppose it’s the same (be that http or https) as the current page. You can se it in action on Slashdot by viewing their source – none of the regular links has protocol specification – they just start with two slashes and the hostname (ie. //netfactory.dk/).

In the case of Slashdot I suppose they don’t use it to save bandwidth, but in most cases it’s probably useless.

Mysql is usually pretty fast by default, but to keep performance to the max sometimes requires knowledge on how mysql works and how to write queries that does their job most efficiently. Today we ran across a simple example which illustrates that mysql’s ability to use an index depends on the way you write the query. Let’s make a simple table and add an index on the date column:

create table members (
	m_id integer auto_increment primary key,
	m_name char(20),
	m_since date not null
);

create index ix_m_since on members (m_since);