Sose, we are arguing on the semantics. Like Skepticals noticed, we are talking about the same things in a different context. If we are only interested in demonstrating how things work in a limited context, indeed we usually sacrifice some accuracy to make things easier to understand.
The problem with those simplifications is that sometimes they lead people to make faulty assumptions (i.e. that the MAC address of a network host can not be changed using software) or to consider that what they just heard/read is something absolute and has the same meaning universally, in any context. In the computer networking context, for example, indeed an IP address is described using Software (that runs on the computer), while the MAC address is suggested by the Network Interface Card (which is a hardware device installed on the computer). However, like Smurf said, it is also true that the hard-coded MAC is provided by software (that is run not on the computer's CPU, but on the NIC). That information could be considered irrelevant enough to ignore when we examine a computer, since for all practical purposes, a NIC can be treated as a hardware device and nothing more, but it does not mean the information is not correct.
That's why it is important to be completely accurate on the context in which we speak.
In your original post, you didn't clarify that the center of your attention was a computer and it's components, so I assumed that you wanted to give an accurate definition for IP addressing, as a concept, beyond any implementations.
I don't know the arguments of each side (what did your fellow networker claim?), but it is possible that you were both right, you just couldn't agree because you examined the same thing under a different point of view, and so were using different criteria.
NOTE: SOSE IS AN IDEALIST, AND CRITIQUE INCLINED, BUT ALWAYS STAND TO BE CORRECTED.
Likewise.
