[
Lists Home |
Date Index |
Thread Index
]
On Monday 09 December 2002 11:30, Miles Sabin wrote:
> Alaric B. Snell wrote,
>
> > On Monday 09 December 2002 04:27, Paul Prescod wrote:
> > > Actually, you can't do significant development with TCP/IP without
> > > understanding how IP addresses and NAT work.
> >
> > Yeah you can! Particularly with the java.net APIs; you wrap your
> > hostname in an InetAddress object and then pass it to a socket
> > without really needing to know that it's been resolved to an IP
> > address underneath.
>
> Respectfully, I have to disagree. For _significant_ development with
> TCP/IP (ie. anything moderately robust and non-trivial), the java.net
> APIs are woefully inadequate ... hence the (now not so) new java.nio.*
> APIs.
Ok, we're disagreeing on 'significant' - I meant 'useful', in that you can
write clients for most standard TCP protocols with java.net having being told
what TCP provides without knowing how it works. But I wouldn't write a
production-grade Web server with just that, no :-)
>
> Miles
>
ABS
--
A city is like a large, complex, rabbit
- ARP
|
