One of you sent me an email asking how the code I wrote was different from simply using:
import java.net.*;
InetAddress addr = InetAddress.getByName("www.yahoo.com");
System.out.println("IP Address = " + addr.getHostAddress());
I have to admit, its not all that different. The truth is, I didn't even know it could be done so simply. Aside from that, there are a few advantages from having written my own code. The two biggest are that I learned a great deal doing it about writing for tcp sockets and UDP packets in Java -- something which is surprisingly easy. I also got a deeper understanding of how the conversion from data type to data type happens in the JVM. For example, a byte converts to an integer of course, by assigning the 8 bits of the byte to the lower 8 bits of the integer -- but I didn't realize that when it do so it would treat the high order bit of the byte as the sign - meaning you get a value between -127 and +127 rather than between 0 and 255. There's probably a way around that, which I'll run into accidentally no doubt -- but for my purposes it was enough to know it, and simply test the result for being negative, and if so, add 256 to the result.
It makes sense to do it this way I suppose. You can't really go the other way. If what you're expecting isn't always positive, but is mixed, you'd never be able to figure out which was which. At least this way, if negatives don't make sense you know what to do. Its unusual to need a numeric value, but not ever need one larger than 255, but one such case is in IP addresses. The DNS protocol was written in the days when people still cared about specific bits to save space and transfer time. If it was re-implemented today, no doubt it would be DNS:XML and it couldn't be UDP because the packet would be too big. It would happen on Port 80 over a web service, require an application server to handle, and perform about a thousand time slower.
Of course, the other cool thing is that rather than relying on the operating system to resolve the value, I'm able to specify the dns server to use. That makes the code useful for blackhole lists and other such things. Its also a proven protocol so I can use this code for lots of other kind of lookups that are similar. That may help me in some of the anti-spam work I'm doing.
Knowing that the built in method exists, it would probably make more sense to try to extending the class and adding the ability to specify a name server rather than starting from scratch with my own class. Still, I'm glad I did this. It was a day's work not at all wasted.
If you're interested in the code, let me know. Obviously its not going to something I can or should sell.
Comment Entry |
Please wait while your document is saved.
complement to represent negative numbers, basically this means a negative
number and it's positive equivalent sum to 256 which overflows the 8 bits and
leaves zero. A 2's complement number is formed by inverting all the bits and
adding 1, so for example 42 in 8 bit binary is 00101010 so -42 is 11010101 +1
which is 11010110. So if treated as a 2's complement signed short int 11010110
is -42, you would then add 256 to this which gives 214, and 214 in binary is
11010110. If you add 11010110 and 00101010 you get 100000000 (9 bits - it
overflows and leaves a byte of zero)
There are 10 types of people in the world, those who understand binary and
those who don't :-)