how should I properly implement non-SE-enabled links

no-follow? javascript? some redirect script?

[muszek]

I have a big site about a browser-based online game. That online game has a stinking load balancing system - they use www1.example.com, www2.example.com... www20.example.com, which makes it hard to create a link to any of that games document.

Some time ago I have written a set of functions that creates link pages to particular documents (players, teams, etc.) on example.com. But every link instance contains of twenty-something actual links. Not only I'm leaking PR this way, but also all target pages are all identical if visitor is not logged (so google sees same content for all those links).

I want to rewrite this tool a bit, to make those links irrelevant for search engines.

How should I do it?
-- no-follow attribute? - does it work as it should?
-- some kind of javascript? - is it bad for whatever reason?
-- some redirect page + robots.txt?

Please tell me how would you solve this problem.

[jchampliaud]

If I'm not mistaken Matt Cutts suggested a no-follow attribute in cases like yours. I'd have a look around his blog.

[sugarrae]

>>>Not only I'm leaking PR this way, but also all target pages are all identical if visitor is not logged (so google sees same content for all those links)

Well, it you kill the links for the search engines, they aren't going to follow them and you won't have PR for the internal pages anyway, if I'm understanding what you want to do correctly.

>>no-follow

I wouldn't use that personally. Yahoo *does* index no-follow links (they also show in linkdomain searches for the target domain) - though whether or not they ignore them, etc, is something you'd have to get from them. But the fact that they index them at all makes me not want to use a no follow tag on anything I *absolutely* do not want indexed. I'd stick the destination content in a folder and 403 forbid it and go about my way.