Barry: Thanks for letting me know about the formatting issue. For a couple minutes, the article was posted badly formatted, and Google Reader must have crawled and cached it at that time.

nice approach. Have You considered indexing the Article_View resp. User_View?
As the only thing needed is an aggregate over the votes (ArticleVote resp. Karma) I would suggest creating an index on both views. That way one places the burden of updating the aggregate column totally to the indexing engine.

Additional benefit comes from the fact, that SQL server is able to infer the index on the view, even when the view itself is not refrenced in the FROM clause. So even when querying like
SELECT
Sum(v.Vote) as Sum_Votes, v.ArticleID
FROM Vote v
GROUP BY v.ArticleID

the query analyzer is able to infer the index on the Article_View! So even existing queries benefit from the index.

Martin

I’ve used this technique before with good success. Microsofts SQL Server also has the ability to create “indexed views” which are roughly equivalent and a good option if your view logic is a simple select query.