There are 10 matching posts.
This year I have been thoroughly put off from blogging about things in case I alienate readers, who are interested in one subject, by talking about another. Obviously those people with an interest in competitive cycle racing and ruby love this blog, but the others are sort of left with a bitter-sweet taste on reception of my RSS feed.
So, I’ve decided to solve this by having three categories, each with an RSS feed. If you are interested in my posts, but only one subject, you can now pinpoint what you want with one of the following three categories…
Riding, racing, kit and general cycle related babble.
Ruby on Rails, and anything else coding or web related.
Personal stuff, holidays, photos, etc.
Everything
If you enjoy all of my stuff, then my entire blog is still available at it’s old URL and RSS. Service as usual as far as that is concerned.
So now you can look forward to lots more high-quality posts in a subject of your choosing, without being put off by all my other stuff.
As part of an attempt to save some bandwidth, and to prevent disruption further down the line, I am now serving my RSS feeds using Feedburner.
The new addresses for my feeds are:
I have deprecated my original feed from about 4 years ago, it now only displays a message with the new feed addresses. This will happen very soon with the current addresses from my blogging app, as I play around with the architecture.
After 3 days of torment attempting to get my new Rails app online, it is now up!
My server is now running Ubuntu Dapper with Lighttpd and FastGGI serving my blog. I ran into two problems getting this to work:
- First up Ferret does not install properly, and the file ferret_ext appears not to be compiled, and thus unavailable to the app. As soon as I started the server it would crash, citing the missing file. For this reason search is disabled, but will hopefully be reinstated as soon as I have solved my problems with this.
- The RMagick Gem wouldn’t work, and was again cited as missing by the server. I uninstalled the RMagick gem, and installed the ruby library available in the Ubuntu packages. I’ll put instructions for this in another Blog post.
Hope you like the new design, I certainly feel it was worth all the effort it took, though I’m slightly disappointed that my search isn’t working for the moment.
Yep, my blog now has a new design! It has been over a year since the last one, and I really wanted to try out a design I’ve had on my sketchpad for ages. I can’t remember where I saw this style of blog (dark, light, dark, stuff at bottom etc), but if it does come to me, or I see it again, I’ll put a nod in the right direction, or change it completely if I find it is too similar (This is why using a random sketch isn’t so great, you forget sources, and it looks like plagarism).
Design
- My navigation is now at the top, but has now been joined by the search box.
- On the bottom I now keep a short bit about me (which will soon link to the about page), some other links to RSS feeds and the like, and a sampler of gallery pictures (also to be implemented)
*Tag clouds now appear at the bottom of most pages when there is no comments or older posts to display.
Site Layout
- The front page of the site now displays the latest blog post, as well as some of the more recent ones.
- The Weblog page is now where the archives reside, and all my old posts can be found there.
- On each post’s dedicated page, there is short summary of the post up top, just so people coming straight into the site know where they are.
- There is a dedicated Tag Cloud page, making it easy to see what subjects I blog most often (bike is currently miles in the lead).
- These is now a comment feed for the whole site, individual comment feeds for each post, and feeds for each of the tags. I’ll link to the last of those soon so you can try them out.
Architecture
- Ruby on Rails! Yep, my new site is the result of about two weeks of work with RoR. No more PHP for this blog!
- Now with tagging. One of my main reasons for changing languages was that I could add new features such as tagging without having to work with old stale code, plus it let me learn something new.
Operating System
Installing mod_fastcgi on Fedora Core 5 was a bit of a nightmare, so I decided to change linux distributions. As I write this I’m currently running Ubuntu Breezy under Parallels desktop on my Macbook. Dapper doesn’t run under parallels for some strange reason, and the Edgy beta is also having problems, so I’ll stick with Breezy for these initial tests. If it doesn’t work out, I’ll also be trying out Solaris 10 and FreeBSD.
Currently the site is served using… Lighttpd under Ubuntu Dapper.
Still to Come
As with all my new releases, it’s a little rough around the edges. The backend is still pretty much a Rails scaffold affair, and I still have features to add such as…
- Gallery, might simply be Flickr, Zoomr, or Smugmug integration, I’m undecided as yet.
- About, Bike, Contact, Design, and Links pages.
Side Effects
Obviously all these changes mean that some URLs have changed.
- All the old blog post addresses will still work, but redirect to their new equivalents.
- The RSS feed for blog posts has also now moved to feed://douglasfshearer.dyndns.org/feeds/blog, so make sure you update your RSS readers.
See anything wrong?
If you see anything wrong, please don’t hesitate to get in touch: dougal.s@gmail.com
I came across an interesting dilemma of UI and back-end design. Most blogs use an address slug similar to ‘http://url.com/year/month/day/post-title-as-slug’. Part of my design was to create this automatically from the title once the creation form has been submitted. So what happens when we edit the title of this article? Do we change the slug or not?
If we change the slug, we lose the original address. If word has got round and many people are now linking to your post, you’ll lose all those visitors. It’s a bit like your zip code changing just because you painted your house a new colour. A solution to this would be to keep the slugs in a separate table from the blog posts, so that old slugs remain, and new ones are added. In this way old versions of a slug will still point to the now updated article. This does however involve a lot of unnecessary complication.
If we decide not to update the slug, it will be out of date, and perhaps not tie up accurately with the title. This is a minor flaw, as it is unlikely that a title would undergo a major change. We can show the user what the slug will look like using a quick bit of javascript in the UI, and tell them that they can’t change it in the future. This seems a nice way around an otherwise complex problem, and is going to be the approach I will take. To allow the article title and slug to be changed right up to the point it is published, I won’t generate a slug until the post leaves the draft state.
It may be nice to have the option to edit the slug manually, that’s something that could be definitely be implemented as an advanced admin feature. This is something that
Dolphinling has been talking about with regards to Wordpress.
Got an email today to say that I have done enough to graduate with a Bsc in Ordinary Sciences. While being good news, it does however mean that I have to do something for the next year, then I’ll have a go at a Masters.
Better get on with my Blog rewrite then…
I had a cup of tea at about 2300, the result of which is that I can’t get to sleep, and it’s now 0226.
I now have a good idea of where I want to go with the Ruby on Rails blogging engine and CMS for running this site. The framework that RoR provides should make it much easier to achieve what I desire than pure PHP did. My PHP code ended up being fat and bloated, and adding any functionality became a nightmare.
Using Rails should also solve the problem of templating, as the view code shouldn’t be as mixed with the business logic. We’ll see how it goes, but the current want list is looking like the following:-
- Admin – Multiple users, Permissions (definitely a later iteration)
- Posts – tagging, textile, rss feed, drafts, previewing, recycle-bin, searching.
- Nice addresses – http://douglasfshearer.dyndns.org/2006/09/04/tagline_of_post.
- Commenting – akismet and CAPTCHA integration a must
- Editing – Pane to add photos in gallery into post + uploading new images.
- Pages – just html, textile will allow me to do this where I desire.
- Gallery – tagging, resizing/thumbnailing w. cache (similar to what my php version did), sets, photostream
- Links – tagging, thumbnails
This is exactly where I realise that my current blog formatting is pretty poor when it comes to lists.
On friday morning our router in the flat decided to have a small fire to itself. It does now mean we have a shiny new wireless router with more toys, so we are all up and running again.
Only problem I’ve noticed so far is that the loading of my main blog page is painfully slow, with generation times of around 2 seconds being reported. Not so good! I’ll have a look at finding out what is causing this once the stress of exams is over.
Got the Hope floating disk rotors I mentioned in my last post, I’ll have a picture of them up soon.
I noticed the other day that there was a bug in my layout under IE, this has now been fixed. Now I just need to fix the display error with my bags Ad on the front page of the site.
This week I’ll also be implementing Google Adsense advertising. Not something I really wanted to do, but it will cover at least some of my expenses!
As you can see, along with all the info at the bottom of my posts, I have added some stars for ratings. There is no connection between these and the backend as yet, but I’ll get that working later on.
The new backend is looking good, with the blog being powered by it already. An object orientated approach does make for some extra processing overhead, but it’s worth it for the ease of organisation, and control over what different elements of the system can do.
Update
Ratings now work! My iTunes inspired ratings stars look pretty nifty, and actually work now! Now onto creating the gallery I’ve been talking about for so long…
