Review of leading group , YU HU & Mindy ]

Last class meeting we discussed some interesting issues, such as in 
the first paper, issues covered :
How to make clear the meaning of "end point"  ? 
Using which machenism to define correct and suitable "end point" ?
The layer of fuctions lying in the internet should be dynamic ? if so, can
we try to find a machenism to decide it ?
Trade-off issue between performance and cost is open to more research, it
is also open to find a measurement for it ?
And etc.

In the second paper, we focused on some of these issues:
New requirement comes out, modify "end point" better or insert some
function into the network ?
The new function should be invisible to "end point"?
Does it reasonable to imagine that we can add some new specific
applications athigh lever instead to add functions into network? Which
kind of standard and which possible measurement are suitable to it? In
class , we imagined it using a 
children browser instance. 
In caching problem, we discussed much on "up-to-date" issue , "user
numbers counting" issue and etc. 
Is the second paper sufficient to argue against end-to-end
principle? There is also a trade-off on it.
How can you trust the third parties ? ( It is really open to think .)
We also covered some problems about : certificate, notary public , some
"end point" argument in systems such as Napster and Gnutella?


Next class meeting, we want to discuss some technical and philosophical
topics when you are appointed to design a national direct demecracy voting
system ?
( How to implement these functions and in which level and its merits and
drawbacks )