“Being the Product Manager at ScienceDirect responsible for linking, I felt compelled by your post to try to explain the background to the College Edition linking problems, and to tell you about what we’re doing to solve them.

The issue with article-level links to College Edition is an unforeseen consequence of a major infrastructure release that we deployed on March 3rd. Basically, two factors contributed to the problem:

- The first one was that of the article locator page popping up. The article locator isn’t new – it’s been around for a very long time, but it only used to present itself for about 200 Elsevier medical and life science journals. A logic consequence of our architectural changes, however, was that the page was presented for pretty much all of our 2,000 journals on ScienceDirect. This created much more confusion than we anticipated – clearly a miscalculation on our part. We have reverted this course of action, and the article locator page is now again appearing only for the original 200 or so journals.

- We also have a software code bug that has been breaking article-level linking from PubMed and CrossRef for College Edition customers for some time now. We have a solution for this, but we need to prepare it, test it to make sure it works, and then deploy it. I’d love to be able to make a promise about when we will able to implement this fix, but fact is that it might take a few weeks. As soon as we have a fixed date, we’ll make sure that our College Edition customers get informed about it. After the fix, seamless linking to the full-text of subscribed journals in your College Edition collections will work without a hitch.

Due to the complexity and scale of ScienceDirect, it took us some time to identify all issues surrounding this problem and come up with a good game plan – but that doesn’t justify the mixed – and at times clearly wrong – messages you and others have been getting from our support organization and from our sales reps. Even as we were still investigating the issues, we could at least have made sure that you got a consistent and correct answer from us. We didn’t do that, and I apologize for it.

Please continue to blog and vent your opinion about what we’re doing. Although a bit painful at times, it is very valuable to us.”

Instead, like Heather, we’d use the displayed URL:
http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6WNY-45SJBXF-1R&_user=582442&_coverDate=07%2F31%2F2000&_rdoc=1&_fmt=&_orig=search&_sort=d&view=c&_acct=C000029699&_version=1&_urlVersion=0&_userid=582442&md5=41d3387119e7ce37fc81122cc53a510f
They’re not actually PERMANENT but they seem sufficiently stable for ereserve.

(It’s interesting that if you use their system to “email” the article, they don’t use the DOI, but a citation token.)

I actually just sort of skimmed the thread and forget the details , but anyway, you aren’t the first to notice problems with their external linking mechanisms.

We ended up using the url included with the citation export, which seemed to match the browser window url. That was 6 weeks ago, and I haven’t checked back to see if the links still work. (nobody has told me that they don’t, though.)

When I was in London a few weeks back there were protests on the street about Elsevier outside their offices. You’ve got to be concerned about a publisher that attracts that sort of attention!