Through Queen's Library webpage I can access PDFs of many journals going for years back. There are probably around 20 or 30 journals that I cite from most often. I just realized how easy it is to look at the titles of every paper published in a journal for the past 50 years or so.
So my plan is to take the time to work through these journals, downloading papers and making notes regarding which project ideas the papers are important for. I'll go back as far as I feel is important. After this is done once for a journal, then all I have to do is keep up.
I also have a plan to review the most recent year of articles in all of these journals every year. I put a repeating date (Feb 12) to review the journals into my palm pilot. This way I will feel comfortable that important journal papers will not slip past my notice. Attached to the date is the list of journals I will review. I can keep this list updated. I suppose I can do this for conferences too, if they have online archives.