Jyte should autosuggest cred based on the cred that person already has.

Cred clustering? I dunno, convince me.

Say that I'm going to give you cred. I see that you already have the cred of "code monkey". I consider it, and agree that that is a valid cred for you. As I begin typing c-o-d, it should suggest 'code monkey' so that I can quickly select it and not take the chance that I accidentally give you the cred of 'code monky'.

For a good example of auto-suggest tagging done well, look at ma.gnolia's implementation.

Great idea! Currently you may manually switch to the cred tab at the bottom and see what cred they have, but auto suggest would be much nicer.

Ah, okay. I thought you meant like, if I have "code monkey" tag I should also get "coder" tag, which I'm not entirely sold on.

This could easily lead to cred convergence: people building up cred on stuff they've already got, but not getting new cred tags as often. On Friendster, where you can leave testimonials, I noticed that subsequent testimonials often tended to echo points in previous ones.

Autocompletion does sound nifty, but perhaps based on all tags instead of just the ones a person already has?

It kind of makes me sad to see someone with cred in the same thing spelled two slightly different ways. And for math, maths, mathematics it's kind of pointless to either have different people using different ones or to have to put all 3 for one idea. So, something that helps to make it more uniform is good.

Some cred guidelines or conventions might be in order...do we want them to match the same tense of tags and interests? Should they all be skills (nouns) or descriptions (adjectives)?

Can't jyte work out synonyms in tags

Something like del.icio.us tag suggestions when posting. That is a beautiful implementation.

Maybe cred could be suggested based on your own rating as well.

i believe that is a different but existing claim.