My reason for mentioning the pointlesness was that it has become very difficult to tell whether an OS is bloated, or has a small footprint. This is mostly Microsofts fault, since it has done an admirable job of blurring the line between OS and ordinary application. However, since small footprint almost always means giving up a lot of functionality people are starting to take for granted, it makes any comparison rather unequal, not to mention unfair, and thus futile. In other words, what the discussion lacks is a precise and quantitative definition of 'bloat' and 'small footprint'---and I have a very strong feeling that such a definition will never appear, too.
I also think it doesn't matter anymore, given todays prices of CPU MHzs and solid state MBs, and the fact that the OS is no longer the main resource hog. Neither the Pegasos II nor your average Wintel clone are lowly A500's where every byte and CPU cycle is made to last. This is proven time and time again by the various wish lists which appear on this and other websites: people want to have G5 CPUs and over 1 GB of memory. You don't need that kind of power just to run the OS.
People will be quick to point out that you can still strive to save as many resources for your programs, which I agree with. It is most definitely a Good Thing. But without an objective way of measuring the increase in productivity and throughput as a function of the amount of free resources, especially in the difficult case of 'sufficient resources' and 'more than sufficient resources', the argument falls rather short. (No sane person will run the programs they want to run on a system which has a chronic lack of required resources---that will bring any system, even one with a 'small footprint', to its knees.)
If you care about such issues, you diligently search for programs which are written in hand-optimised assembly language, and sacrifice every shred of comfort for the sake of efficiency. But then 640 KB of main memory ought to be sufficient for everyone, right...?
