Linking to Articles in Course Materials

Here is some sample text showing how it is possible to link straight through to an article using an inline, persistent link. The article I have chosen is Combined Searching of Web and OAI Digital Library Resources, for no good reason, really. The link is persistent because it makes use of a DOI. If you look at the link, you'll see it looks like this:

http://dx.doi.org/10.1109/JCDL.2004.1336145

The first part (http://dx.doi.org/) is a DOI resolver - it takes the DOI (10.1109/JCDL.2004.1336145) that describes a digital object, looks up where that object resides, and sends you there.

If you don't believe it can send you to different places, try these:

• 10.1045/april2005-hammond
• 10.1145/985692.985762
• 10.1109/JCDL.2004.1336145

Sometimes it is useful to be able to pull links into a page dynamically. By this I mean that the links are pulled into the page from a database elsewhere, using a bit of web browser magic. For example, the following links all come from my delicious account:

In principle, anything that produces a list of links e.g. as an RSS feed, or even better, a JSON object, can be pulled into a page at the point it is needed. Updating the database that provdes the feed will update the links that are fed to the page.

As the above shows, it's not too hard - even with my limited design skills - to find a way of presenting the link information that acknowledges the source of the links as e.g. a list of links provided by the library. (It looks better in Firefox than it does in IE...;-)

There is an issue here of course relating to persistence and uniformity in the provision of links over time. To this extent, it may be that a CT maintain a link database of links they feel they would like to use in a particular section on of the course, and this live/maintained list is used to generate a static listing of links once per presentation.

As another quick example, how about something from ROUTES on T396? The following will grab the RSS feed of a ROUTES search on the course code 'T396' and squirt the links into the page.

Note that this doesn't have to be done live (although in this case it is). It's easy enough to hide/reveal links that are statically hardcoded into a page.

This is something new I'm trying out. As before, I can pull in a list of links from somewhere using JSON or RSS or whatever, and then - as long as they are links to web pages rather than websites - I can perfrom a search over just those pages.

For example, reusing the live links pulled in from Delicious (as described above) as the target URLs for my search:

If you were searching the same set of journals over the same period using the same search terms in several different search environments, you might expect to get the same answers each time you make the search. (Could we demonstrate this?)

However, the same is not necessarily true of commercial search engines.

For example, the Dogpile metasearch engine has a very nice visualisation of the the degree of overlap in results from the same search on the Google, Yahoo! and MSN search engines (the 'GYM').

(If you're reeeeeeeeeaaaaaaaalllllllllllly interested, here's a report they wrote comparing results from the GYM search engines.)

Even within the same search engine, it is possible to receive different results to the same query depending on the territory the search engine is viewed in, as this comparison of the results from google.com in the US and google.cn in China shows.

(Here's something similar comparing the top results from the same search on Yahoo! and Google.)

As far as Wikipedia goes, you may be intersted in something like this screencast of a Wikipedia animation that shows how a particular entry might change over time. Unfortunately, the script that lets you do this for real on a Wikipedia page you choose in your own browser (requires Firefox) appears to be broken :-( (although it's probably fixable over a day or two).