I see it's already hit the front page, but here is another page of advice directly from a Googler: https://plus.google.com/u/0/115984868678744352358/posts/Gas8...

The author (pierrefar) also commented in the HN thread of that post about JavaScript suggestions he'd come across.

http://news.ycombinator.com/item?id=3470396

SeoMoz now has a more technical rundown on how to do a 503 status code: http://www.seomoz.org/blog/how-to-handle-downtime-during-sit...

If other 50x error codes can result in an SEO, why bother to even use them? is it frowned upon to use 503s where an actual server error response would usually be returned?

I can't see anything in the linked article suggesting that other 50x responses will result in poorer search engine performance - I would expect that any of the transient-fault 50x responses would have the exact same search engine impact. The important part is not returning 200 OK and having googlebot index your blackout page.

It just so happens that 503 is the semantically correct response for this situation, while other 50x responses are not.

I'd really rather everyone use the correct status codes (503 indeed being the correct one here) than try figure out what their SEO impact may or may not be. I'm quite sure Google understands HTTP.

I'd be cautious of that method. Google is not a fan of cloaking, and you might wind up with a penalty for much longer than the blackout.

Or just use a CSS black curtain.

That's stupid. It's not about taking damage, it's about raising awareness.

Try this: http://maddox.xmission.com/

For a startup's site, avoiding a search engine ranking drop is more like "Bring a sleeping bag" than "Bring the caviar", it seems to me.