I recently wrote some criticism of views that RSS feeds are “only for new stuff”, and given that I had a request today for a randomized Maricopa Learning eXchange feed, it was time to put my feeds where my mouth is/was.
Before today, the feeds we generate as fixed static files (updated every hours as a cron script) on the MLX where the 10 newest items over all, the newest items and random items per Maricopa College, newest items in a set of subject areas (Biology, Humanities, …)– and available as both RSS 1.0 links and JavaScript include files (unlike FeedtoJS, these are static files that just echo the output created from the hourly updates, cheap caching).
I just updated that set of “fixed” feeds to include ones that are random selected items in the discipline areas. That was an easy update, and something that just fell off the radar. This is dependent, of course, that our MLX package “creators” provide an academic area when they create their packages. It was un-reasonable (IMHO) to create a “fixed” vocabulary, so it suffers from some of the slop of folksonomy, or user supplied free form tags. But it is not bad, e.g. the MLX items for Nursing, Biology, Humanities, etc. all have new and random feeds.
We had before the ability to generate a URL link and an RSS feed of the 10 newest items form ANY search result, dynamically generating the feed (meaning doing the database query, and returning the results as XML). What I just added today are new options to generate a randomized result of specific queries.
So for an example, a listing of all MLX items containing “critical thinking” and associated with Chandler_Gilbert Community College is linked by a specific URL:
http://www.mcli.dist.maricopa.edu/mlx/search_results.php?college=CGCC&scope=5&srt=1&words=critical+thinking
And you can get RSS feeds for the 10 newest in this same query:
http://www.mcli.dist.maricopa.edu/mlx/get_rss.php?college=CGCC&words=critical+thinking
As well as a feed for 10 random in the same query:
http://www.mcli.dist.maricopa.edu/mlx/get_rss.php?college=CGCC&words=critical+thinking&rand=10
As a bonus for CDB readers, the random number parameter on the end can be any number of desired results (well not any number, it is capped at 20 as there are pranksters out there).
Now that we have these sorts of feeds, one could, in theory, use them with FeedtoJS, but keep in mind, these results follow this path:
* query sent to MLX database
* results parsed to XML output
* XML processed by Feed2JS script
* content echoed to web page as JavaScript
It might be a lot of traffic, so I may explore, someday, a strategy for caching the dynamic generated search results.
Anyhow, it was fun to get my fingers in the MLX code again. Much more work there is coming next month, as well as a new face lift (CSS Design).
