Buffers can't explain that, because they are not counted towards individual app memory usage.

Applications are free to allocate as much memory as they want, even if they never use it. For example, iTunes might really only need 10 MB of memory to operate normally, but it allocates 300 MB just for kicks. If you look at something like ps or Activity Monitor, you'll see 300 MB for its "Virtual" size, but probably 10 MB or less in the "Real" memory column. In fact, if it never uses that memory even once, it's not even written to swap.

But there's no easy way of distinguishing this scenario from the case where it's really using 300 MB on a regular basis, and the operating system just happened to swap to disk 290 MB of it.

The amount of VM an application allocates is _totally fictitious_, and I've never found a good explanation why some apps do that.

Because it's simpler, and it doesn't do any harm.

To be precise, it's usually not the application itself that requests memory it doesn't need. It's the memory manager of the application. That way, the memory manager can be implemented much more efficiently.

Take for example the Java VM (I know you don't like Java, but in terms of memory management, you can't get much more efficient than Java). It will always request the whole heap as virtual memory from the operating system, and rely on the fact that the physical memory is allocated on first use. That's because Java uses several optimizations that make memory management in general faster, but resizing the heap extremely slow. Because of that, virtual memory is allocated optimistically in order to reduce the chance that the heap has to be resized afterwards.

Itoons – Bloatware? Who the f*** do they think they are?

Lately there has been some talk among a fringe group of itoons users, some of whom don't even own an i- anything, that the highly esteemed and universally loved itoons is bloated. What do they mean by bloated? Hard to say. They just blather on and on about functions they don't understand and then try to complicate the issue by going off on irrelevant technical tangents. Although some of them are software developers who proclaim to know more about the inner workings of a computer than a journalist such as myself who uses one every day, they fail to understand the basic fact of the matter: no matter what they say, I can argue with them until their lips are numb.

My research has shown that this stubborn and grumbling minority of 9 out of every 10, or a mere 9% of all computer users, typically own old black "laptops" that someone like you or I wouldn't be seen with in a junkyard. Goodness knows what they do with them and whether in fact they have a lifestyle at all. Probably they spend their evenings listening to "audiobooks" or playing "go." Don't they know that itoons can do that?

Apparently it is impossible to get it into some people's heads that a music player can and should do so much more than just play music, and that the good folks in Cupertino know how we like to do things better than we do ourselves. Although I have asked for people's honest opinions, some people just won't change their minds. If you are one of those sorry few who has sadly chosen a competitor's product, don't feel too bad - perfection is not the fare of the common curmudgeon. As to the rest of you, have no fear: we know what you think!

Bad news Kirk, you've been scooped. The good news however is that the author appears to share at least some of your views.

Do you have a link for that? I can't find it via Google. Or did you just make it up...

lol

Hmm, that's interesting...

First, talk about bloated - when I run CGoban for a while (not right away on startup) it can take up to 300 MB of memory.

But I just launched it now, and it's "only" using 119 MB. VM is shown as 371 MB. So if I understand correctly, this doesn't match what you've said. Did you mean that the real memory and VM figures should match?

Actually, that's easy: When the application allocates 300 MB and never touches it, the task manager will not show this in the real memory column. When the application uses 300 MB, it will be shown in the real memory column, regardless of swap.

This said, you're still complaining about an app running slowly on a machine which is way under spec: Mac OS X calls for a minimum of 1 GB RAM.

I was under the impression that RSIZE (or "Resident Memory") only showed what was using physical RAM, not memory that was "in use" but swapped out to disk. Some quick Googling seems to confirm. Is that not the case?