Beefy Boxes and Bandwidth Generously Provided by pair Networks
go ahead... be a heretic

Re^4: Sorting "SuperSearch" results by votes?

by LanX (Sage)
on Jun 05, 2009 at 08:26 UTC ( #768710=note: print w/replies, xml ) Need Help??

in reply to Re^3: Sorting "SuperSearch" results by votes?
in thread Sorting "SuperSearch" results by votes?

Moin Almut 8 )

But I'm pretty confident that sorting by reputation can be "more useful" than sorting by time.

I wouldn't be so confident :)  Actually, I think using other means to narrow down the search (additional keywords, etc.) such that sorting of the remaining hits becomes more or less irrelevant, is the better strategy

It's just ranking, you equally have the possibility to narrow down by keywords with either sorting method.

Honestly, if your interested about the feedback of a CPAN Modul X::Y and you get 100 hits, wouldn't you preferre to start reading the 10 posts with rep>20 before getting to rep=0 ???

Of course there are much better and sophisticated ranking methods possible with the metadata we have┬╣ ... but sorting by rep is much simpler to implement just by extending an SQL-Query. It's easier to get the break even of investment and payback.

Cheers Rolf

(┬╣) like weighting the votes by XP of the voter, or by "voters who voted like you".

UPDATE: Or weighting by voters who regularly vote for "too-late-posts" (posts which slipped out of focus of average voters).

  • Comment on Re^4: Sorting "SuperSearch" results by votes?

Replies are listed 'Best First'.
Re^5: Sorting "SuperSearch" results by votes? (implementation)
by tye (Sage) on Jun 05, 2009 at 13:53 UTC
    but sorting by rep is much simpler to implement just by extending an SQL-Query

    Um, if you sorted on rep using SQL with Super Search, you'd turn something specificly designed to prevent undue load into something performantly exceedingly awful. And adding an index on reputation would rather suck also, since reputation is the one thing about a node that changes an order of magnitude more often than anything else about the node.

    - tye        

      OK, no simple SQL-Query, but surely you could cache the normal (max 100?) results of one search-run and sort them within perl???

      Anyway if it's too complicated to realize, never mind ...

      Cheers Rolf

Log In?

What's my password?
Create A New User
Domain Nodelet?
Node Status?
node history
Node Type: note [id://768710]
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others studying the Monastery: (3)
As of 2022-08-08 01:19 GMT
Find Nodes?
    Voting Booth?

    No recent polls found