I used to get worked up about the short lifespan of most bioinformatics tools and techniques. A three year old software is very likely to be completely unusable.
Then one day it struck me: It needs to be counted in Dog Years (where one human year corresponds to seven years in a dog's life).
Now it all makes sense! That three year old paper is really a 21 year old technology as far as bioinformatics goes. Of course it won't work.