There's a lot of confusion regarding penetrations, confusion made worse by "code wonks" trying to 'improve' things.
Let me be plain here ... I was present during, and took part in, the basic ANSI full-scale fire testing done at a "major testing lab" that developed the data upon which the codes were written.
The purpose of the testing was to see what effect various penetrations had on known, tested, and proven wall assemblies. Penetrations evaluated included both 'through' penetrations (pipes) and partial penetrations (boxes).
From this data were derived the 'rules' regarding the use of putty pads, the placement of penetrations, etc.
Putting it simply ... the rating of a wall is directly dependent on the simple thickness of the drywall. Since walls are tested as complete assemblies, with temperatures monitored on the "safe" side, the thickness measured is the combination of the coverings on BOTH faces.
Let me state that differently: a wall with 1/2" drywall on both sides tests out the same as a wall with one side open and 1" of drywall on the other.
In that manner, it really does matter whether the panel is surface-mounted or set into the wall. The effectiveness of "boxing" the panel in drywall depends entirely on how well it is done, as well as how many cable / pipe penetrations there are of that "box."
In the testing, 4" boxes did not cause the wall to fail. 4-11 boxes did make the walls fail. Thus, is was proposed that 16 square inches of penetration be allowed in each stud bay.
FWIW, there was little difference seen whether boxes on opposite faces were 'back to back' or separated, or whether the cavity was stuffed with insulation.
It's the stud 'bay' that matters .... so, conceivably, if the studs were 12" apart you could have a penetration of one face every foot and the wall would still pass the fire test.
When it got into the hands of the code writers, they chose to word things so that a 4x4 box, regardless of the mud ring used, needed a putty pad ... and these had to be at least 24" apart, without regard to the stud spacing. Strictly speaking, these "safer" requirements are not supported by the test results.
XXXXXXXXXXXXXXXXXXXXXXXXXXXX
Naturally, cutting a long slot, one bridging several stud bays, will affect the fire rating of a wall. With almost all household walls using 1/2" drywall on each face, plus the lack of 'rated' doors or windows, it's safe to say that we're not talking about a rated wall in the kitchen or dining room. It's not an issue.
It's pretty hard to install multiple receptacles close to each other and still have everything come out in a straight line. Nor will the tile guy thank you for it! Plus, you invariably have a (technical) code violation when the thick tile forces you to use long mounting screws on your devices.
Folks often say "I HATE plugmold!" I'm not sure where this comes from, though I can see how a plugmold / Wiremold strip can look bad after a few years' catching paint and grime. (This is especially true when the Wiremold has to zig and zag all over the place before it reaches the plugmold or box.
Yet ... what if you could avoid all that? What if you fed the plugmold directly from the box, and set it into the wall? There would be very little projection, and the long, straight line would be a joy to the tile guy.
The big question is: How do you connect the plugmold to the house wiring? Those connections have to be accessible, and the components really ought to be connected to each other. Where does the GFCI go?
The GFCI placement is an issue, as the breaker is (now) supposed to be AFCI.
