Danoff's right. Some motherboards can do "cable select", where they will choose which is the master and which is the slave by which connector you use. Usually the connector that is furthest from the motherboard end will be the master. You then set the jumpers on the drives so that one is master and one is slave.
Make sure that you match the speeds of IDE devices, because they will only run at the speed of the slowest device. Therefore, if you connect a CD writer and a hard drive to the same channel, the hard drive performance would be seriously decreased. If you have an Ultra-ATA100 drive, and an ATA33 drive your system will run faster if you have them on separate channels, even if the ATA33 device shares a channel with a CD drive. Mind you, you will find that if your BIOS supports ATA100, and you connect a '33 drive to it, the POST will complain bitterly (although it will start, so you can pull data off a legacy drive).
You will also find that a computer will attempt to mount all the primary partitions sequentially, and then any extended partitions after all the primaries, unless you use Disk Administrator to force the drive letters. This will mean that if you have two drives, each with a primary and an extended partition, they will be mounted thus:
master-primary
slave-primary
master-extended
slave-extended.
Some manufacturers manually assign d: to the CD drive in their build image (on the Product Recovery CD), so you would get:
c: master-primary
d: CD/DVD-ROM
e: slave-primary
f: master-extended
g: slave-extended.
Whether you can tolerate that is up to you, obviously. However, on multi-drive systems, I tend to set my CD/DVD drive to R: (for Reader) and my CD-Writer to W:
Finally, with NTFS partitions, there really isn't any need to partition them unless you wish to do so for organisational reasons. In the bad old days, larger partitions were inefficient, but this has been pretty much eradicated. Well, until you get into terabyte territory at least...
