Peter Sefton has posted an e-print of his forthcoming Serials Review article "Towards Scholarly HTML" (dx.doi.org/10.1016/j.serrev.2009.05.001) on ptsefton.
In an attempt to comply with Elsevier's author agreement, he states:
At the moment that link seems to resolve to an open version of the article, whether or not you have a subscription to the journal but I guess that will change; when it is "published" you will only see the article if you are clicking from inside a network that's on their list of subscribers. If not, you will need money to see it. But I can post the article here with the copyright statement you see below and remind you that you need to use the DOI to cite the paper should you wish to. No naughty linking back here (unless it is to reference these comments I'm adding). And no linking to the version I'm about to put in ePrints. OK? Even though you know that if you do link to the DOI some people may not be able to see the article in the future, don't do it, use the DOI link. There, I think I told you.
Note: the Elsevier version is no longer freely available.