Inspired by one of Penn's (of Penn & Teller) articles in which he mentions he has his computer tell him what he wrote in his journal that day the previous year, I've wanted to implement a similar thing with my blog. Now that, as I mentioned previously, I've updated my blog such that its much easier to implement search and such, I've added date range filtering to my site's search. So now I can easily see what on Delicious and my blog I was doing last year.
I've also otherwise updated search on this site. You can now quote terms to match an entire string, stick 'tag:' in front of a term to only match that term against tags as opposed to the title and body of the entry as well, and you can stick '-' in front of a term to indicate that it must not be found in the entry.
I've redone my blog's layout to remind myself how terrible CSS is -- err I mean to play with the more advanced features of CSS 2.1 which are all now available in IE8. As part of the new layout I've included my Delicious links by default but at a smaller size and I've replaced the navigation list options with Technical, Personal and Everything as I've heard from folks that that would actually be useful. Besides the layout I've also updated the back-end, switching from my handmade PHP+XSLT+RSS/Atom monster to a slightly less horrible PHP+DB solution. As a result everything should be much much faster including search which, incidentally, is so much easier to implement outside of XSLT.
I now have search and an archive available for my site. I previously tried to setup crappy search by cheating using Yahoo Pipes and now instead I have a slightly less crappy search that works over all of the content that I've produced on my blog, uploaded to flickr or youtube, or added to delicious.
The search is only slightly less lame because although it searches over all my content, I still implemented it myself rather than getting a professional package. Also, the feed supports the same search and archive as my homepage so you can subscribe to a feed of Cadbury if you're so inclined and just skip all this other boring stuff. My homepage and feed implement the OpenSearch response elements and I've got an OpenSearch search provider (source) as well.
The move of my website to NearlyFreeSpeech.NET is mostly complete except for a few server side things not working yet: RandomGrammar and parts of Vizicious. I'm still very happy with the NearlyFreeSpeech.NET hosting and so far I've only spent a few cents on hosting. At this rate I'll only spend a few dollars a year.
I've moved all my pages to use the same CSS and hooked it up with cookies to my Kuler color options so now changes to the color theme will stick and apply to all my pages. I haven't figured out the caching for this yet so you may have to refresh to see changes to color applied.
I've switched from using my own home web server of which one of the harddrives died, to using NearlyFreeSpeech.NET, an actual real live web hosting service. So far I'm very happy with them and they give me almost exactly what I had on my own home server: ssh access, vim, php, java, etc. etc. The only notable things they don't do are (1) cron jobs which I use currently and (2) SSL which I don't use currently. I can replace my cron job usage and I suppose I'll have to reevaluate my web hosting if I ever need SSL. At the moment many of the server side things like Vizicious will be unavailable. I'll work on getting those working again at some point.
I've setup a minimal search page that uses a Yahoo Pipe to sort of search through my content. I say sort of search because I only get full text search over my recent item feeds and otherwise I just search over my tags.
To get real search I'm going to have to keep an archive of all my content on my own website. This is a pain but on the other hand it will let me easily backup my content or display old items on my page. Why didn't I just use a prebuilt solution?
I use my recently added books feed from LibraryThing, a site I've mentioned before where you track, review, recommend, and share your books, and I put the recently added books on my page. I thought it might be nice to include the book covers so I suggested adding book covers to RSS feeds in LibraryThings 'Recommend Site Improvements' group. The next day I had a response from the founder and lead developer Tim Spalding who had started implementing the feature. I noticed a few bugs, reported them on the same thread, and he fixed them soon after. Fantastic! It makes me want to upgrade to a paying account.
Incidentally, if you notice the Ghost in the Shell book appear multiple times in my RSS feed its due to the previously mentioned iterative bug fixes. The same item appeared multiple times slightly differently with each bug fix and your RSS aggregator may have picked them up as distinct items.
tag but when I submit that tags turned into an