Note that I've just cribbed these messages by Jon MacLaren from one of our internal newsgroups for my reference.
For those of you who are using the ACM digital library - at http://www.acm.org/dl - you may have had difficulty printing older papers. It seems that the ACM have simply scanned in the pages of each paper, and creating PDF files from the bitmaps. Twelve pages of text and equations can result in a 1Mb acrobat file. When printing from the acrobat reader on the Suns or Solaris (don't know if linux is better), the reader turns each page into approximately 17Mb of postscript!! Clearly, there are big problems for printing; your machine will run out of space on /var after about 4 pages, etc.. There is a utility pdf2ps on the Suns - I can't get it to work on Solaris, but that's probably just me - that only triples the size of the document from PDF to postscript, i.e. 3Mb of ps for the 1Mb acrobat file mentioned above. The print quality appears to be the same - for around 2 orders of magnitude less postscript! There's no manual page for pdf2ps, but entering the command without any parameters will give you usage instructions. One word of warning: you can't substitute "-" for a filename to mean stdin or stdout, i.e. if you specify an output file of "-", it will create a file called "-", which is tricky to remove... [ For further information, the ACM digital library provides access to most ACM papers - journals and conference proceedings - and often goes back to the mid-1980s. Very useful. ]
Recently I've been using a citation database at: http://citeseer.nj.nec.com/cs It's fairly up-to-date (although it doesn't have older articles in it), and covers technical reports and conference publications as well as journals. Perhaps the nicest thing is that it can give you context on where an article was referenced, i.e. you can read the referencing paragraph. Even better, you don't need any username/password to look at it.
|Donal K. Fellows, Computer Science, University of Manchester, U.K.|