am = Array[Subscript[a,##] &, {6,6}];
am = am /.
  Flatten[Table[
    Subscript[a,i,j] -> Subscript[a,j,i],
    {i,2,6}, {j,1,i-1}]]

javascript:window.location=location.protocol+'//'+location.hostname+'.ezproxy1.library.arizona.edu'+location.pathname+location.search+location.hash;

To make it work, copy the above code, and paste it as the URL in a bookmark. In all modern browsers, alternatively, you can just drag this link to the bookmarks bar:

• ez

In either case, make sure you edit the javascript code and replace ezproxy1.library.arizona.edu by your EZproxy server.

Once "ez" is in your bookmarks, clicking on the bookmark will have the same effect as clicking the ezproxy Google Toolbar button.

$lastmod = (stat($filename))[9];

use strict;
use XML::XSLT::Wrapper;

my $xml_input_filename = "xml/data.xml";
my $xml_output_filename = "xml/out.xml";
my $xsl_filename = "xml/transform.xslt";

my $xslt = XML::XSLT::Wrapper->new(
		ProcessorList => [
			'libxslt',
			'sablotron',
			'xslt'
		],
	);
my $result = $xslt->transform(
		OutFile => $xml_output_filename,
		xml => $xml_input_filename,
		xsl => $xsl_filename,
	);
$xslt->dispose();

Values given to [code]]czo2OlwieG1sID0+XCI7e1smKiZdfQ==[[/code] and [code]]czo2OlwieHNsID0+XCI7e1smKiZdfQ==[[/code] can be either file names, or xml/xslt code passed as a string.

If [code]]czo3OlwiT3V0RmlsZVwiO3tbJiomXX0=[[/code] is ommited, then the result of the transformation is returned to the variable [code]]czo3OlwiJHJlc3VsdFwiO3tbJiomXX0=[[/code].

b2evolution

A google search for blog engine (I wanted something I controlled completely on my own server) gave me b2evolution as a first result. After a glimpse of its capabilities, and reading a few positive reviews, I decided to give it a try. So I downloaded the Summer Beta release, i.e. version 1.8 beta, which is supposedly stable. (BTW this is the only 1.x stable version, although a beta!)

Pros

The setup was pretty straightforward. I was really impressed with the configurability: the backoffice has an amazing spectrum of options. I also liked the multilingual features, i.e. the fact that you can specify the locale for each post separately.

Cons

The basic setup leaves you with ugly URLs, which in the worst case look like [code]]czo3NDpcIltiYXNldXJsXS9pbmRleC5waHA/dGl0bGU9cG9zdF90aXRsZSZhbXA7Yz0xJmFtcDt0Yj0xJmFtcDtwYj0xJmFtcDttb3J7WyYqJl19ZT0xXCdcIjt7WyYqJl19[[/code], and are a little better with what they call extra info: [code]]czo0MTpcIltiYXNldXJsXS9pbmRleC5waHAvMjAwMy8wNS8yMC9wb3N0X3RpdGxlXCI7e1smKiZdfQ==[[/code] (note [code]]czo5OlwiaW5kZXgucGhwXCI7e1smKiZdfQ==[[/code] in the middle!). There are some situations (not many though) where I could live with URLs like that, but a blog is not one of them.

So I tried to fix it with [code]]czoxMTpcIm1vZF9yZXdyaXRlXCI7e1smKiZdfQ==[[/code]. According to the few forum threads dealing with the issue, this should be not hard to fix, but it turned out to be a nightmare that took me hours, although I'm used to dealing with [code]]czoxMTpcIm1vZF9yZXdyaXRlXCI7e1smKiZdfQ==[[/code]. It seems that the latest version of b2evo engine interacts in some way that it took me quite some time to hack it. I will not go into details here, maybe in another post, since the solution is far from obvious. In any case, even after solving the problem, the URLs were not uniform: some links to the root of a blog were [code]]czoxOTpcIltiYXNldXJsXS9ibG9nMS5waHBcIjt7WyYqJl19[[/code], some [code]]czoxNTpcIltiYXNldXJsXS9ibG9nMVwiO3tbJiomXX0=[[/code], and some as the latter but with a backslash in the end. The best I could do to fix this is to do a 301 redirect of both cited URL types to [code]]czoxNjpcIltiYXNldXJsXS9ibG9nMS9cIjt7WyYqJl19[[/code]. But links with .php are still in the navigation system, and this bothers me. I would need to edit the template to fix this.

Another thing that I didn't like about b2evo are the skins. They are not terrible, but compared with Blogger and WordPress, they just look like amateur work; they don't come even close!

Overall, I do enjoy having the power of its backoffice, but the version that I had felt a little buggy. And the documentation is pretty poor. Some day, this could be a great blog engine, but right now there's a little something missing.

WordPress

This was a completely different story from b2evo. Everything about it is very neat, very professional looking. The skins are great. The documentation is much, much better than b2evo's. This is really a great blog engine... or would be if it weren't for one small thing: it's Rich Text Editor with all the javascript around it that just won't let you be in control of the code of your posts! There's no way of turning off these functions, and there is no simple HTML editor. For instance, I could not insert a piece of code with [code]]czoxMTpcIiZsdDtwcmUmZ3Q7XCI7e1smKiZdfQ==[[/code] tags, since the editor would always strip off all line breaks, leaving all the code in one line! How ridiculous!

Conclusion: if all you are going to be posting is text (no code), and you don't care about HTML editing of your posts, WordPress looks like a great choice. If only it had a simple HTML editor in addition to the Rich Text editor, like Blogger does, it would be my first choice.

Edit: The visual rich editor in WordPress can be turned off! See my coment

Blogger

So for now, I'll stick with Blogger. The only thing I'm missing with Blogger are categories, and I can't understand why in the world won't they add such a simple functionality that would draw so many more people to Blogger! I see many people choosing WordPress or other engines over Blogger for this and this reason only.

What I like about Blogger is the fact that all files are static HTML, that they are physically on my own server (which is a big psychological thing for me), there is no php burdening the server, and all URLs are nice as they are, with no [code]]czoxMTpcIm1vZF9yZXdyaXRlXCI7e1smKiZdfQ==[[/code].

So, Blogger, I'll stick with you for now, but if you want me to stick with you in the long run, do something about those categories, please!

<filter action="drop" type="wildcard"
    pattern="*.py" />
<filter action="drop" type="wildcard"
    pattern="*.txt" />
<filter action="drop" type="regexp"
    pattern="/google\w+\.html" />
<filter action="drop" type="regexp"
    pattern="/\d+/(\d+/)?$" />

<filter action="drop" type="regexp"
    pattern="/archive/$" />
<filter action="drop" type="regexp"
    pattern="/uploaded_files/$" />
<filter action="drop" type="regexp"
    pattern="/uploaded_images/$" />

In this way, void URLs will not be included in the sitemap.xml output.

If you're wondering "Why would I be running the sitemap generator on my blog?", it's because Googlebot does not handle well the posts where the time & date has been altered.

If you don't know what the sitemap generator is, see Google Sitemaps Help.