Ahh, that sucks... I, too, have cm7 (not a nightly, just the most recent stable), no overclocking, but have very decent performance.

Yes - I did also originally hate the older Firefox for Android, but the newer incarnations work very well, as long as you've got a semi-decent phone. Some of the older Android phones may be bogged down a little, but anything recent should work well. I have an HTC G2, and am pleased with the speed of the application, and haven't used the Android browser for months. Same goes for tablets - any of the newer/faster tablets run Firefox very well.

I know some aging queers that, when asked, were not too supportive of being sucked off by former patent trolls.

There is another means: https://www.facebook.com/settings Click "Download a copy of your Facebook data." and follow the instructions.

OTOH, you posted your comment from a (relatively) reliable and affordable internet connection that probably is able to do unrestricted web searches. Could be worse.


..................not that we should ever stop trying to make it better.

Well, I never close the bathroom door to take a leak, but that's an entirely different issue...

my windows 3 box is ridiculously secure; no one's been able to turn it on for years :(

It's actually called reverse tunneling, and it works exceedingly well. In a typical reverse SSH tunnel, an outbound TCP connection is made to a server, which brings up a tunnel that the server is able to send traffic over back to the client. This removes the need for the client to allow any inbound-initiated traffic or any possible NATing.

Anyway, my idea probably would have greater backward compatibility with existing networks and software

Because older applications would still be able to generate routable traffic? Maybe... but this introduces additional problems, because you are still using ports for routing purposes *and* application traffic.

Also, how do 'legacy' applications access services on destination hosts that are on non-standard ports? A old web browser, for example, is going to make a connection to "http://www.example.com" using tcp/80. How is that browser able to connect to the webserver running on mail.example.com, if it's running on tcp/81? Manually entering "http://mail.example.com:81" may be an option with a browser, but what about an application that is hard coded to make a connection to a specific hostname and port?

And again... are you suggesting that all server applications should require their own unique DNS names?

While your idea is unique, I think it would too dramatically change the way modern networking works and would create too many other problems to be a viable solution.

IPv6 also requires a lot of changes, if it didn't, it would be easy to implement.

It does, yes, but IPv6 does not change the nature of how server/client applications function.

You realize the "http" portion of your example URL is what is determining the protocol and port at the application level, right? You do not ask a DNS server to resolve "http://example.com", you ask DNS to resolve "example.com" and then the *application* generates 'http' traffic to the resolved host. Essentially, you are proposing that port numbers be used for routing. How is that ultimately better than instituting a larger numbering system for IPs (IPv6) that is already widely supported and does not dramatically change how all modern applications work?