Dear (Provider of Academic Article Databases),

Article databases are really one of the best things that has happened to academic research in the last, ten or fifteen years. It makes accessing research painless and efficient, and the more available research is, the more valuable it is. I don't think anyone will disagree with this assertion.

Here's how it works today, as best I can understand it. (Note this is a fairly psychology centric perspective.):

There are two classes of journals: those published by the APA and it's partners and those published by for-profit academic presses like Springer and Klewer. The former are the big masthead journals and there are about 30 of them (Developmental Psychology, Journal of Personality and Social Psychology, etc,) I suspect this is true of many disciplines: universities and professional organizations publish a number of journals (In Women's Studies, Signs and differences, are published by Chicago and Hopkins respectively). Anyway, and then there are smaller for profit journals which tend to be less prestigious but can cater to specific niches and sub-fields. These journals are the ones that Google Scholar picks up on and they don't tend to be archived in the same way that the NFP journals are.

Ok, background, stated. Generally what happens is that the professional organizations or universities create databases (APA and Psych Artcles; AAA, AnthroSource; Hopkins and Project Muse, for instance.) that index all the abstracts and sometimes full text are then licensee the database to a provider (CSA and Ebsco are two examples of this). Some of these databases contain FP journals, many do not. In humanities disciplines, all this content ends up in jstor after a few years anyway.

So it's incredibly fragmented and the databases, are, to my mind, sort of poorly organized. The metadata is sort of weak, and the searches are straight boolean, so it doesn't learn, it doesn't see connections between what you search for and what you find is what you find, it's very rigid, and very much not the way, we do things on the internet these days.

Now I don't need fancy ajax crap, lord knows, and frankly, I think some sort of desktop interface for the database, might be really quite effective and be really helpful for people who are actually doing research.

So I guess my suggestions to academic database providers--because this is a letter after all ;)--are as follows:

1. Allow some sort of adaptive meta-data, potentially with some sort of user generated tagging system 2. Integrate/mashup content from multiple 3. Create some sort of more adaptive search capacity that has learning/adaptive algorithms. 4. Have better offline content. 5. Have better BibTeX support (because if you're asking...). 6. Perhaps a little ajax so that it it can do some predictive caching and the web pages work a little faster. 7. Allow me to paraphrase a lolcat for this: Y it so ugggglry, can has purrrty intarfaces now pls? Because why not.

In any case, It'd be nice, but I'm not holding my breath.

Onward and Upward!