Darwin & others,
Adding an entire archive of PDFs to EndNote items manually must be extremely tiresome. Given the popularity of EndNote I bet many researchers are doing just that anyway (they probably curse a lot in the process  ;D).

So if there's no tool to programmatically add the PDFs then maybe smart folks at DonationCoder could look into how hard it would be to make one? If such a tool could be completed it could draw a lot of interest to DonationCoder. And hey, it might due to saved time speed up scientific progress a (very, very, very small) bit.

The main problem to solve would be how to match the pdf files and the bibliographic items. If journal articles contain their DOI id number as extractable metadata then that might be a solution. I opened a few journal articles in UltraEdit and browsed for doi numbers or something similar in the ascii text but couldn't find anything. Also, I exported metadata from some sample PDF files with the free tool PDFTK. Here's one example of what I got:

InfoKey: Author
InfoValue: x
InfoKey: GTS_PDFXVersion
InfoValue: PDF/X-1:2001
InfoKey: Producer
InfoValue: Acrobat Distiller 6.0.1 for Macintosh
InfoKey: Creator
InfoValue: InDesign: pictwpstops filter 1.0
InfoKey: ModDate
InfoValue: D:20070613112031+01'00'
InfoKey: GTS_PDFXConformance
InfoValue: PDF/X-1a:2001
InfoKey: Title
InfoValue: 14.6 News 758-759.indd NS new.indd
InfoKey: CreationDate
InfoValue: D:20070612095515+01'00'
PdfID0: cee3243181236aabefbac383315fda1f
PdfID1: 3ab8e10f7074409da8a4fc9d6eb3a2
NumberOfPages: 2
PageLabelNewIndex: 1
PageLabelStart: 758
PageLabelNumStyle: DecimalArabicNumerals

this row "InfoValue: 14.6 News 758-759.indd NS new.indd" is sort of an identification since the PDF is from Nature Volume 447 Number 7146 section News page 758. But it's not as good as a DOI. And other articles i tried did not even have that InfoValue. PdfID0 & PdfID1 are made by Acrobat Distiller I think, and does not relate to the article contents/DOI.

Embedding DOI numbers according to some standard seems like an obviously smart thing. If available it would not only allow adding a pdf to the corrent EndNote items but also the reverse. Start with a pdf, autoresolve its DOI online, then import all the resolved metadata into an EndNote item, rename the pdf according to some format and import/link it to the EndNote item.

But it wouldn't really surprise me if DOI is not embedded at all. The electronic journal systems still seems rather ineffective and user-unfriendly.  As this overview states (http://hublog.hubmed.org/archives/001306.html ):

While most of the larger publishers provided an acceptable method of authentication, the PDF files they produce are obviously not optimised for ease of use by the reader. It's almost impossible to build a tool to automatically fetch PDFs for papers [---] The implementation of all of these features could be automated with little change to the publishers' systems, but would be a major benefit to researchers struggling to deal with large amounts of literature.

(BTW, a comment on that page links to http://quosa.com/solutions.html , a tool for easy, massive article downloading and EndNote importing. No price is listed on the site (only this note on discounts http://www.quosa.org/support/helpdocs/mac/pricing.htm ). So I suspect it is very expensive. )

Without DOI in the PDFs any programmatic matching of PDF and EndNote items would probably have to be more fuzzy.

Armando,
"If I could find another reliable strategy which is sufficiently portable (I don't want to loose years of tags and annotation by just changing software or OS), I'd give it a try..."

Well you could do what I do with journal articles but for any other types of files also. That is, create a .txt file with the same name as the file you want tagged (but perhaps with "_tag" at the end) and put the tags and keywords in that .txt. I think any indexing search tool will find those tags and due to the same filename you then find the file you're searching for. If you have a consistent and unique format for your current tags in filenames that a script can isolate (like using [tag1 tag2 tag3] only for tags) then you could probably make a batch script that migrates to the ".txt system".