pros cons controlling md5 crush uri primary pivotal database

i'm building database store information operation objects (such systematic papers, specimens, dna sequences, etc.) have coming online identified url, an identifier such . controlling guids primary pivotal vigilant seems reasonable idea, i've followed controlling md5 crush guid. you'll md5 crush your browser station bar rodent over correct mislay buttons tasty connotea book mark. example, bookmark

http://delicious.com/url/e4a42d992025b928a586b8bdc36ad38d

where e4a42d992025b928a586b8bdc36ad38d ais md5 crush .

does anybody have views pros cons approach?

for me an advantage proceed (as opposing controlling an automobile incrementing primary pivotal generated database itself) i have lot links between objects, controlling md5 hashes i store links presumably record (say, outcome information mining/scraping), following import bulk database. same way, database rebuilt scratch, urls objects won't change since md5 hash.

i'd acquire any thoughts presumably sounds sensible, presumably there (better?) ways doing this.