Given a list of researchers, each with a number of publications (for conferences or journals), with appropriate meta-data (such as citation counts), how may one determine whether a particular paper has achieved a "test of time" benchmark, relative to the other researchers and papers?
My initial idea was to compare a journal, published for example in 2005, and compare its citation counts to all publications by that journal in year range (2005, 2005+10). If the number of citations for the input paper are in the 80%+ percentile for the journal over the period, it may be argued it has stood the "test of time" with regards to it's impact.
However this seems quite crude, how else could I quantify this benchmark, and ideally provide a means to visualise it as well?
I lack a formal background in statistics (aside from the minimum), so any ideas or views are welcome.