RSS intrigues me. I know it involves XML and it's a way of syndicating your website (though I'm not quite clear what that means), but I don't really know what it's all about. So I've been reading up on it this evening. The two RSS readers I've come across so far are: NewsDesk 1.1 and SharpReader. They both seem very similar and have both been developed under the Microsoft .NET framework. This means downloading the .NET framework, which I've just started, but it tells me I have about an hour's worth of downloading left, so I think I'll hit the Cancel button once I've posted this.

Do I really need an RSS reader. Probably not. But, like I say, I'm intrigued to find out what it's all about, so I'll probably give it a run out some time soon.

I was doing a little long-overdue weeding of my IE Favorites this morning. I came across one called "Thoughts of a Technical Author", which was the name of a temporary weblog I kept for a while, while I was testing Radio Userland. When the free trial period expired I stopped blogging.

I clicked the link to see if anything was still there. The link opened a directory at radio.weblogs.com containing the contents of my blog, as individual HTML files. I was quite pleased to see it all still there. I've copied all of the files and will, at some point, save their contents as entries in this MovableType blog.

It was July and August last year. I'd used Radio Userland, thought it was okay, but not as configurable as I wanted, and not independent enough. For a quick, simple blog it's great. Setting it up on another server seemed like it wasn't what it had been designed for and I couldn't find much documentation. So I decided to use MovableType - and then did nothing, except think it over, for more than a year, until this week.

Problem
I work from home one or two days a week. I use Outlook as my mail client both at work and at home. The trouble is that, because my machine at work is always on (so that I can access it remotely, if required) I have to remember to close down Outlook every evening before I go home. If I leave Outlook running at work it downloads my mail off the server, so that I can only access it by using VNC. But I don't like leaving my work machine VNC-able, so this isn't a good solution. Anyway, VNC is slow. On the other hand, if I access my work mail from home, my emails get downloaded to my home computer, then when I get back to work I can't access them.

Solution
Use IMAP mail. I've just switched to using IMAP mail in the past few days and it's the perfect solution. My friendly system administrator advised me on this, and, oblidgingly, set up an IMAP mail server on the machine that handles our mail at work. He also told me to take a look at Thunderbird, the new IMAP mail client that comprises the mail client bit of Mozilla, as a standalone application.

Thunderbird
It installs with a download and a few mouse clicks. Set up is fairly straightforward. I already access my work network gateway machine from home using PuTTY to establish an SSH login (using public/private key authentication). The PuTTY session has several defined tunnels, one of them being from a localhost port to port 993 (for secure authenticaton) on the server at work that handles mail. All I then have to do is tell Thunderbird to use the specified port on localhost and it tunnels through to the mail server, allowing Thunderbird to read the contents of my inbox, etc. The good thing about IMAP is that it doesn't download anything unless you tell it to, it just marks mails as read and leaves them where they are.

The Thunderbird install at work was even easier. Just plug in my username and the name of the mail server.

The only slight hiccoughs are with the home setup.

a) I can't use my full email address as my Thunderbird email address on my home installation, because my mail server at work won't allow mail from outside to bear its domain name, but I want people at work who receive my mail to be able to click Reply and send the reply to my work inbox (not the inbox of some external mailbox). The way round this is just to use my network username as my Thunderbird email address. I only mail internal work people using Thunderbird at home. I still use Outlook for all other non-work mail. So I can use my network username because this is aliased, on the work network, to my full work email address, so recipients can reply to my username (no @, no domain name) and the mail gets through.

b) The IMAP server passes to the client a security certificate that specifies the name of the server. If you specify the server's name as localhost (or 127.0.0.1) you get a security certificate warning message, because there's a name mismatch. The way round this is to map the name of the remote server to 127.0.0.1 in your Windows hosts file. Windows checks this file first, before trying to resolve host names to URLs in any other way (e.g. via a domain name server). So you can therefore use the name of the remote mail server in Thunderbird. Because the name resolves to 127.0.0.1 the port it goes to is the port on the local machine, which then (thanks to PuTTY) gets tunnelled to the appropriate port on the remote machine, at which point you can log in.

Sounds complicated, but I already had the PuTTY tunnelling stuff set up (thanks to another friend at work), so the switch to Thunderbird was pretty simple. And once it's set up, it works like a dream. When I want to check my work email, I fire up my internet connection (which is an ISDN line and therefore not always on), start the PuTTY session to my work network's gateway machine and log in. I then start Thunderbird, log in and read my mail.

Let's see if I can upload an image, and insert it here:

OK, that seemed to work!

The problems I expected to have getting MT working were with the server-side stuff, particularly the Perl modules. My ISP isn't helpful when it comes to installing Perl modules. They have the take-what-your-given-or-go-elsewhere approach to customer service. I should probably go elsewhere, but ISP's are like banks - you don't change unless you are seriously, seriously pissed off with them. I'm not quite there yet.

However, the only real trouble of any sort I had with the MT install was not really the install at all. I downloaded and printed out the full installation and configuration documentation from the MT website and that guided me through nicely. I went for the standard install, using a Berkeley DB database. I'd rather have used MySQL - but my ISP ... No, the only problem was when everything appeared to be installed correctly. What then?

The MT documentation falls short when it comes to actually telling you what you do to see your new weblog, or how you add entries to it. You have to work this one out for yourself. The answer to how you add an entry is that you have to point your browser to the mt.cgi script in the cgi-bin and click "New Entry" in the left column. Seems obvious once you know, but I'd thought you would add entries from one of the weblog pages, rather than by going directly to the Perl script.

To see anything you need to rebuild the site by clicking "Rebuild Site". This is not obvious. If you've never built the site in the first place it's not intuitive to have to click "Rebuild Site".

Maybe I missed a document somewhere, but I don't think so. It seems like the MT people guessed installing and configuring would be the hard bits, so they documented all that pretty thoroughly. They seem to have assumed that once you'd installed MT successfully, you can just bash on. What they need is a little "Getting started" document: the stuff you do after you've installed, before you start messing about getting the configuration just how you want it - e.g. view your site, post your first weblog entry.