The problem with text match search is that it misses on a step removed concept search. I'm advocating classification with hierarchy, so that if you were to click on Legal, it would give information on lawyers, on trade marks, on patenting, etc. I think the answer is to limit the hierarchy to startup related issues, and then users select from this list to classify posts if they want to. Then we can have a very helpful semantic tool for finding things that search just cant.
Yep, I know the reasons for tags. I still think they're a truly bad idea in this context.
A limited selection of tags solves one problem, but then creates new agonizing moments for contributors where they try to figure out whether they ought to select one option or twelve. And a few users, the very ones I probably least want to pay attention to, will assume that their latest blog post about their new GTD web application crosses boundaries in unique ways and would be relevant for all fifty-eight tags. It takes a strikingly small number of bad results to make tags worthless...and I've seen very few occasions where the tag features of a website are not just that.
I could be wrong...but I'd like to see an example of a site in the vein of News.YC that uses tags effectively, before I'd even think about admitting it.
Thats a good point. If you remember the good old days, Yahoo was a hierarchy of classified links administered/edited very carefully by humans after user submissions. You would need to have that layer of edit control by the system admins to prevent abuse, and I'm guessing the resources for that are not here.
(I sure wouldn't post my GTD web app in non-relevant classes though! :)
