Context. OR, A word on useless comparisons

Submitted by illogic-al on Sat, 04/04/2009 - 10:38

On my running system, from a fresh log in, KDE reveals 1268876 K memory usage in GNOME vs. 1279348 K memory usage in KDE 4. That is a change of 10472 K. Granted, this is not a huge difference.

This was taken from an article which I won't bother linking to. It doesn't need anymore traffic. I just thought I point something out as it's Saturday, I'm bored and I had a long day in the lab yesterday and need to vent and for some reason the previous quote from an article which shall remain nameless just annoyed me.
Perhaps I woke up cranky. *ahem*

You see boys and girls, whenever we make comparisons it is important to take into consideration, or provide for your audience so they they may take into consideration, extra information which may be pertinent. For argument's sake, let us call this extra information context.
For example, when making comparison between the size of an Amarok installer, and an iTunes installer on the OS X operating system one might mention the size of both to compare. Now let's say that the Amarok installer is 70 MB and the iTunes installer is 130 MB. Clear win for Amarok, yes? But what if the iTunes installer has 60 MB worth of translations, while yours has 0 MB. Not looking so good now are you? For 70 MB each you both do the same thing: play music. Context. But wait, Amarok might provide more features within that 70 MB. Context.

And speaking of MB (megabytes), some quick math. K's as a measure of memory, is something which might be unfamiliar to the lay audience. We can instead do our audience a service by doing some simple math conversion for them. Let's say, for argument's sake, that *1 MB = 1024 KB. Let us further assume that **1 K = 1 KB. Given these hypothetical conversion factors, and the Gnome and KDE 4 memory usage given above, we arrive at the following values:

  • Gnome: 1249.3 MB
  • KDE 4: 1239.1 MB
  • Overall Difference: 10.2 MB

The overall difference in memory usage is 10 Mb. Certainly less impressive sounding than 10472 K. Presenting our data in this fashion has some very obvious drawbacks. It makes it easier to understand, and so, if the data doesn't explicitly support our drawn conclusion, we may be made to look like fools. Alas, me must strive for journalistic integrity in such matters and leave the audience to make such conclusions. We may even abandon statements biased by opinion entirely, leaving these to the audience, should we fear ridicule resulting from ridiculous statements.

Case in point:

But it does show that GNOME requires less hardware to run.

Hmmm. According to the data above, and the subsequent, albeit hypothetical, inferred calculations, Gnome and KDE both use over a gigabyte of memory. This indicates that both Gnome and KDE need a system with > 1 GB of memory to run. How then, does it show Gnome requires less hardware, when it is physically impossible to get the hardware which would meet this requirement. We go from 1 GB of memory to 2 GB or, if you're being cheap 1.5 GB.

Perhaps we only have 1 GB of memory and the swap space on the hard drive would be our limiting factor. As the only hardware being compared, or mentioned, is memory however, how would we know? Ah, context. A beautiful thing really.
I won't even touch on how the > 1 GB of memory usage value was obtained. I mean, really? So it isn't possible to run KDE4 on 768 MB of memory? Did I achieved some mystical Black Magic when I did this, not 7 days ago, with KDE 4.2 and ArchLinux? Which begs the further question, what exactly is being measured here and counted towards KDE 4 and Gnome. Ah, context. Perhaps, I will touch on this point.

After our re-representation of the data above we see that there is a change of 10 MB. So what? I hear you in the audience saying; but it gets worse. We still have no context! We don't know what is running in Gnome or KDE. Do we have any file managers running? Do we have no programs running but have simply logged in to the desktop? I certainly hope so, but we can never be certain. Do we have an indexer indexing in both, one, or none? Without this information and the potential to normalize for discrepancies, who is to say this minuscule, in the grand scheme of 1.5-2 GB memory systems, isn't adding some, perhaps optional functionality which Gnome doesn't provide. Who is to say the whole desktop isn't providing far more default functionality than the Gnome desktop? If this were true, then not only would the 10 MB difference be insignificant, but commendable. Something to be congratulated and praised. But alas, we'll never know.

We have no context.

So remember boys and girls, should you ever decide to disparage someone/something publicly, at least do so properly. Provide some context. Make your attempt at humiliation/ridicule complete, because if you don't provide context you could wind up on the recieving end. In the worst case, you'll find you don't have much to go on and save yourself some embarrassment. Good day, Good night.

* 1 MB is actually = 1000 KB.
** 1 MiB = 1024 KiB = 1024 K.

P.S. I do realize this point has already been deconstructed, with more brevity earlier. But that doesn't help me vent, does it? (I wonder, does that count as providing context? I think so.)