9340bf1e   
   XPost: comp.arch.embedded   
   From: philip+usenet@paeps.cx   
      
   Didi  wrote:   
   > On Feb 6, 12:06 am, "robertwess...@yahoo.com"   
   > wrote:   
   >> On Feb 4, 5:06 am, Didi  wrote:   
   >> I'm pretty sure it's not an issue of how smart the implementation is,   
   >> but that routing based on the ARP tables is simply wrong.  If there   
   >> are two subnets on the wire, and you want to talk to both, you should   
   >> have two IPs (one in each subnet) defined for the NIC.   
   >   
   > This is the only way on windows and linux systems, yes. Why crossing   
   > subnets on a wire is "wrong" is something you will have to explain.   
      
   That is not what Robert wrote.  Robert wrote that routing based on ARP tables   
   is wrong (and probably more generally that assuming that implementations will   
   route based on ARP tables is wrong).   
      
   ARP is an address resolution mechanism, _not_ a routing mechanism.  The fact   
   that is happens to work like one on a single (or some) implementations does   
   not mean that it is.   
      
   Most implementations consult ARP only after routing.  This makes sense for   
   reasons that have been explained at length.  Adding static ARP entries to all   
   machines on a network segment instead of just assigning them addresses in the   
   same network is entirely the wrong way to go about things.  Why are you using   
   IP at all in that case?  You may as well just have your own protocol inside   
   Ethernet.   
      
   At this point, I think you're just trolling and flailing madly.   
      
    - Philip   
      
      
   --   
   Philip Paeps                                    Please don't email any replies   
   philip@paeps.cx                                     I follow the newsgroup.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)