The shadow meta-journal

I've been noticing that Google Scholar searches are getting better at coming up with PDF versions of articles. Sometimes these are hosted from journal sites -- and pop right up if the university has a subscription. But more often they are hosted from the authors' websites -- as in, "Here is my lab website, here are PDF's of my papers."

Now, this is against most of those copyright agreements that authors have to sign to put their work in journals. Those boilerplate agreements almost always allow distribution of preprints (i.e. ugly double-spaced manuscripts) but not formatted proofs or journal PDF's. But most of the papers you will find in a search are original journal PDF's.

I just want to thank everyone who puts their papers online. It's incredibly useful. Personally, I'm moving more toward submitting to open access journals, at least where the papers are a good fit.

Can I also say that Google Scholar is incredibly easier for finding something at a journal site instead of using the search methods from the publishers themselves?

By far my biggest complaint is the whole implementation of the digital object identifiers (DOI) by publishers. Remember how those were supposed to make everything easily searchable with a single number? Now, some journals require you to include DOI's in bibliographic entries, presumably so that they can do online hotlinks in their online fulltext. But have you every tried to search for something by DOI? Most journals give you no results at all. And there's no way to search for things across many different journals with the DOI. It's all a mess. And here most of us are busily entering DOI's in our bibliographic databases. What a waste of time!

Of course, the Google Scholar citation search is less complete than ISI. Heck, ISI is even tracking my weblog. But I am really beginning to lean on having the first line of text in the search results. It really cuts my time sifting out stuff I don't need or want.