Hypersmurf said:
The Attack (Melee) action is a standard action; that is not a melee attack, but rather a standard action that includes a melee attack.
Which is exactly the point of the message you were replying to - the table doesn't differentiate between attack actions and melee attacks either!
That's splitting hairs. It's also inconsistent; the PHB never "definitively" defines when something's and action and when it's an attack. Many actions are listed in the section "Special attacks" and some "actions" (such as disarm) per table 8-2 are actually attacks. The terms are used almost interchangeably - a simple attack in the PHB's sense is thus both a standard action, and a melee attack (which might be exchanged for certain special attacks, or might be gained in certain circumstances). Unless it's explicitly stated how to interpret it, differentiation is misleading - you'll have to interpret is as one or the other based on the context, instead.
Further, the differentiation is meaningless. It's meaningless since there's clearly no difference between a situation/ruleset in which you have both an action and an attack with identical consequences and an alternative situation/ruleset in which you have only one action which can also be used as an attack.
Finally, distinguishing between a melee attack (notably absent in table 8-2) and the action "Attack (melee)" simply begs the question whether there's no similar distinction between a sunder action and a sunder attack.
Of course, there's an obvious answer why the PHB doesn't distinguish between these to things: it's written for humans, not machines, and it's extremely confusing to consider two identical actions (actions in the english, not rules, sense) to be two completely different things without some obvious distinguishing factor.
In other words: the accuracy of the ruleset does not extend to the level detail required for such distinctions to be reliable. Inferring the absense of a sunder attack based on the interaction of an underspecifed table and a text blob separated by a number of pages certainly isn't reasonable, especially since the text suggests it is possible, and similar attacks definitely are possible.
I'm sure you can find details of this post to disagree with, but that's kind of the whole point - these details are simply underspecified and it's inherently arguable both ways, therefore.