I’d been thinking of using cvs to help manage offline file-based blogging with Blosxom, but I recently read a couple of posts by Marc Nozell and Rael (to whom the title of this post is attributed) describing how they use rsync. It’s a nicer idea than using cvs, so I’ve gone for it. I made an alias, ‘blogsync’, which does this:
rsync -tazve ssh ~/blog/*.txt gnu:blog/
and stuck the alias definition into my .bashrc file Hey presto. My trusty Linux Vaio is always with me, and I can blog there wherever and whenever I wish. The ‘net cafe at London Gatwick airport was getting a bit too pricey for terminal access, but I passed it last week to discover that it now offers simple 10baseT connections for laptop users. Five quid for 40 minutes. Still steep, but perhaps worth it for the odd time I absolutely desperately must sync up.
To complement the script that allows me to share my feed list with others, including Radio Userland users (with the coffee mug link), I’ve now written a simple pair of scripts which do the equivalent of the 5335 script mentioned earlier.
The point of the 5335 link is that the target is 127.0.0.1, that is, localhost. This time, I didn’t want to run a script of any significant size on my localhost; rather, I thought that if I could just run a simple redirector, which I could configure and get to redirect calls to 127.0.0.1:5335 to a location of my choosing, I’d be able to concentrate running the ‘complicated’ (relative term) part of adding a feed to Blagg‘s rss.dat, on the host that serves my weblog.
So 5335-redir.pl is a simple, configurable redirector, which I run on my localhost, and bladder is the script that receives a feed URL (via the redirector) and adds it to rss.dat.
Very simple. In the spirit of Blosxom and Blagg, I hope.
Here’s a step-by-step list of what happens:
- I click on a coffee mug link next to a feed reference that I want to add to my aggregator.
- The link points to http://127.0.0.1:5335/…. and 5335-redir.pl catches the request.
- The request is redirected by 5335-redir.pl to bladder, running on my weblog host.
- Bladder simply adds the feed reference (sent in the redirected link) to Blagg’s rss.dat file.
- The rss.pl script comes along on it’s regular cron way and rebuilds the My Feeds list and the newly added feed appears there.
In the same vein as the experiment with Peerkat, I’ve now exposed the feeds data for Blagg (which is the RSS aggregator mechanism I’m using with Blosxom) as a list on the right hand side of the weblog here.
It’s very simple; Blagg uses a text file, rss.dat, to keep a list of RSS feeds that I want to subscribe to. I wrote a simple script to read that information and to create a list file that I can then pull into the weblog template like the other lists here. Very simple. I’ll probably cron the script to run every so often, to keep the list up to date.
Now all I need to do is resurrect the 5335 script so that I can other feed info inserted into (appended onto, probably) Blagg’s rss.dat file when I click one of those coffee mug icons.
Well, I got everything back in order relating to the old experimental weblog, and I’m running an instance of Peerkat to expose the blog data. It’s back at the same address (http://www.pipetree.com:8080) but is just for reference. I’m pressing forward with this new Blosxom-based weblog.
There seems to be a wonderful, um, blossoming of scripts, tools, and ideas arould Rael‘s [hmm, note to self - I need to write some vim macros so I get nice <a href=... tags for various commonly used terms, like here: 'Rael'] Blosxom and the new Blagg mechanisms. How about a Blogger API interface for Blosxom, or a plug-in architecture for Blagg, to name but two, in addition to the tiny hacks such as the Google search list and the bookmark ‘to read’ list, both on the right…
I’m sure it’s because Blosxom’s scent of simplicity has people doing what they want, with code, because there’s only the slightest whiff of required compliance, to fit in with how Blosxom works, and that’s something that all platforms, languages, and minds share: files.
The essence of Blosxom, using basic tools and OS-level building blocks, inspires one to keep things simple, and celebrate the humble file as the lowest common denominator for data interchange. Now that I’ve got the possibility of including lists of things on this page (or on other blog pages, for that matter), I’ve managed to address a ‘problem’ I’ve had for a while in a really simple way. I have a bookmark folder in my Galeon web browser called ‘to read’ where I keep references to articles that I want to print out and read (on the train, mostly) – most of the time, though, I don’t have access to a printer.
With a simple script, I just parse Galeon’s bookmark file to pull out the items in the ‘to read’ folder, and make them into a simple list, which I then drop into my file area reserved for Blosxom. A simple reference to this list in the template, and ecce, I have my ‘to read’ list available on every workstation I sit down at.
The move from Peerkat to Blosxom coincided with open disk surgery on our colo box that hosts this website. I’ve got this weblog up and running but need to install Python so I can boot my Peerkat-based experiments previously at http://www.pipetree.com:8080.
I’m really taken with the simplicity of Blosxom. As the weblog entries are simply textfiles, I think I’m going to start using CVS to give me the ability to write offline too. I travel a lot, and my trust Sony Linux laptop goes everywhere with me.
My interview on jabber.org has been posted. Yikes!
I visited Matt’s blog today after finding his Google script that I’ve adopted, and followed an interesting link to an article, “Search the web via IM”, in which Matt is interviewed regarding Googlematic. It’s interesting that the ideas expressed at the start of the article:
Do you use your IM client so much that you feel like you live in it? Would you like to use it for tasks other than one-to-one communication — say, for looking up Web pages?
are in line with what I wrote about in an article about Jabber and bots called “Is Jabber’s Chatbot the Command Line of the Future?”.
I do like the idea of permanent windows opened and connected to bots providing services.
Well, after having a look at Peerkat, from the pen of that maven Rael Dornfest, I turned my attention to his Blosxom project, a beautifully simple weblog mechanism that celebrates the power of the *nix shell, tools, and simple files. It’s just up my street, and I’m a definite convert. Nice one, Rael.
I grabbed the latest version, 0+3i, and hacked in a bit of support for simple lists, such as you see on the right hand side of this page. Just as you write blog entries by editing .txt files, where the first line in the file becomes the entry’s title, so you create lists by editing .list files, where the first line in the file becomes the heading for the list. You include a list in your template by including the list’s name in square brackets, like this: [listname].
Because lists are just files, you can generate them in lots of ways. The Google search list on the right was created using a slightly modified version of Matt Webb’s GoogleSearch.pl script. I’ll probably cron the script to search Google every hour. See – powering Blosxom with standard tools like cron. Lovely.