Discussing the OpenID proposal, Cienna Rand had a great suggestion:
[W]ould it not be possible to create a resident-run openID server? One group could handle the verification of who you are in-world and then provide authentication to others. Adoption could be driven by other third party sites referencing the community server.
This is exactly something I intended to discuss, only I hadn't quite thought of it in this context. Yes, it would be possible to build a passable OpenID server without help from Linden Lab, and that's a flaw in my proposal. (So now you know, for the record, I'm not perfect.)
✗ Don't ask for things you can do yourself.
Many of the suggestions I give are designed to prevent work on the part of Linden Lab to manage the proposal system. The biggest thing you can do to make less work is not propose features you don't need. The biggest group of these are plain old bad ideas, but many of them are simply things that you don't need Linden Lab to do.
When I thought of this rule, I had seen a proposal similar to #469 Listening to your favorite music regardless of land, which is short enough I'll quote it here:
Give players another audio control that gives them the ability to locally listen to their MP3 collection or enter a URL to listen to their favorite streaming audio.
This is a useful feature that everyone should agree should be possible. The reason this is a bad proposal is because it already is possible: when you play Second Life, you have at your disposal a whole computer, capable of doing wonderful amazing things such as this. If you want to listen to your music or a stream of your choice, you are fully capable of running a separate music player program designed specifically to do exactly that.
I hadn't connected that thought with the OpenID suggestion, but it still applies. You could very well build your own OpenID server for Second Life accounts. It would still have to verify you in-world instead of through secondlife.com, but residents would only have to do that once, not for every site. You would still have to remember an additional password, but it would only be one, not one for every site. That would improve the current account situation, and still encourage third-party web sites, though perhaps not as much as if Linden Lab built it.
However, many things are possible through hacks and workarounds. How does this rule not prevent almost every suggestion? Those "but"s are the main reason. While a resident-run OpenID server would be both possible and useful, it would really work better as an automatic process without the in-world validation step. In this case, there's also something to be said about trusting secondlife.com's authority with Second Life residents more than, say, openid.slworld.com. If your own idea requires too big a hack or too messy a workaround, you might go straight from idea to proposal.
Building your own OpenID server would best serve as a proof of concept for Linden Lab's implementation. Any idea that's reasonably implementable by a resident should be implemented first, to prove it works and that it would make a good feature. Science tells us your proposal will be one thousand times more applicable with real proof and experience behind it. And who knows? You may find it wasn't such a good idea after all. (Not that that would happen with your idea, of course.)
In both the client and the world, it's in Linden Lab's best interests to leave out features. A lot of making good software is making a good interface, and a lot of designing a good interface is leaving out as much as possible. Apple is historically the king of leaving things out, but Linden Lab prides themselves on it too, in that the Second Life client is a relatively slender download in the neighborhood of 20 MB. (They then depend on it by not having patches when a new version comes out, instead redownloading the entire client.) Linden Lab is also a resource-strapped startup, unable to build even the features they want to do as fast as they'd like.
Leaving out features is absolutely critical to Second Life's development, so it's remarkable and wonderful that Linden Lab is still so open to accepting community suggestions. You shouldn't squander their limited resources by asking for things you could already do or make yourself.