Essentially it's about usability:
The process should be one-search-one-click.
Authentication should be seamless.
If there is an article that I want, I should be able to enter the title, and get the PDF. Done.
I shouldn't have to think about which database will have the article.
I shouldn't have to find the journal, then navigate through years and issues, and enter multiple passwords, to find an article.
If I do a Google Search or follow a DOI or URL and end up at the publisher's site, I should be able to access the article directly.
Password requirements should be minimised.
Ask the user whether they want to stay permanently logged in, as is done with Gmail, Facebook, and so on.
A user shouldn’t have to see passwords a second time if they are accessing material from their own computer.
The usability benefits of this far outweigh the potential for unauthorised people gaining access to full-text material.
The web of full-text journal articles should be more integrated into the standard web.
I shouldn't have to go through some custom library search interface.
It should be integrated with Google and standard publisher pages.
It's about speed. See the Google philosophy on usability.
Aim to get the time for full-text article access to less than 5 seconds in most cases.
1. Go to search Page:
2. Enter article title
3. Click PDF
4. Click Export Reference
The university pays for these subscriptions, publishers and the university should do their best to streamline access.
Google Scholar is good
I worked out an approach that works fairly well with Google Scholar, but it's still not perfect.
I posted a description of this on my blog.
The integration of the University system with Google Scholar is a very good start.
The main issues with Google Scholar are:
1. The connection between the university system and Google Scholar is good but not perfect.
I'm not sure if there is anything that the university can do to fix up connections to some journals.
2. Google scholar sometimes lacks the very latest articles.
IP Filtering is another great system:
For example, if I go to the home page for the journal "Social Networks":
Due to IP filtering I can download the articles directly.
It would be great if more publishers supported this, although it does create issues with off-site access.
Athens and other systems which allow direct access to publisher site
When I say publisher site, I mean pages like these which are directly accessible on the Internet:
It is very easy to get to the publisher's site of an article using a standard search on Google or Google Scholar.
Many other systems also link to the publisher's official page for a journal article.
Thus, enabling access to the publisher's official site would increase access speed dramatically.
IP filtering and the Athens system are two ways of allowing the use of publisher's official site.
Any other systems that allow more direct access to the publisher's official site would be welcome.
Publishers that do not adopt a standardised authentication system should be lobbied to adopt such a system.