Just realized something: FFXIV is a terrible DPS player's DREAM come true...

[Shurrikhan]

Basically, those kinds of people are already seeking help.

Not always. I've many times had a friend inform someone who honestly thought that, say, Tornado Kicking at every possible occasion was good play, or didn't realize multi-DoTing was helpful for their Repertoire, and I've either been too engaged with healing, tanking, or otherwise making up for that lost DPS to notice. As soon as the friend has made any note of it, I've played enough of each to advise correction by typing between GCDs, but until then I wouldn't have figured to bother.

Or take someone who has a parser but is looking for advice. There it helps to see when their DPS isn't rising at expected rates (e.g. blowing Mage's too soon, not making use of multi-Flare, etc.). Those who know the class can estimate pretty clearly what is or is not being done by per-2s to per-3s rolling summaries, without having to watch their casts. That's one more useful tool, that also happens to be free (over sufficient samples) of misunderstanding.

The only thing the parser in-game allows for is shoving your number into the face of people that don't care for them, good or bad.

We're working under the assumption that the UI can't even be hidden? Otherwise, this is again not a change. For every person who'd be less afraid to reveal a number, there's another who no longer feels the need because everyone else has access to a parser.

Do I trust every player to be capable of following basic rules for interpretation (when you just screwed over the Monk's stacks, yes, his DPS will be lower relative to the duration of the fight; when you decide to chicken out of a whole pull after one of the DPS already popped CDs in preparation, yes, his DPS on the next will be reduced, and yes, gear makes a difference)? No. And I think there are plenty of opportunities to mitigate that, should SE put in a parser of its own. My personal preference as a performance index is a relative potency / second meter (preferably complete with battle log integration and sort and coverage analysis as you would find via FFlogs), because it nullifies internal DHit and Crit variance while taking all but a marginal contribution from secondary stats (unless thus adjusting the base) out of gear difference.