A court recently ruled [1] that LinkedIn had to allow scraping, but it has to be public data (i.e. you can't log into your Facebook account and then scrape all your friends' pages).
It's not the scraping itself that I would be worried about, it's using this data to create what could be seen as a competitor (especially if you make it commercial). For a lot of these websites, traffic is essentially money: diverting users to another website using their own content would be very badly received
I was more thinking about a local web-app that consumes the site and then outputs it in a nice mobile layout, without actually archiving or saving anything.
