Gonzaga--heh, the irony is that I agree with everything you just said AND I also actually do the same thing, in a sense, on the System Shootouts website (ie, I run down pretty much every "raw number" spec I can find and line 'em up against each other).
The difference, of course, is that unlike most gearhead "comparisons" which only use pure benchmark & horsepower numbers, I *also* include software, ease of use, customer satisfaction, security issues, and other less tangible fields in my comparisons.
I used to "score" every category (like the xvsxp.com guy does), but quickly realized that this was pointless since different features have different levels of importance to different people. Now, while I do still declare "winners" in most fields, I don't draw any actual conclusions about an overall "winner".
The exception to this is, of course, the processor itself, where I still use a *very* unscientific "rough equivalent" score simply because too many people *still* use clockspeed as their *only* comparison.
News 
Market Data 
Discover 
AI
