2.5NM is probably a bit of an exaggeration for how far out I'd teach my students. I kind of borrowed 2.5NM from one of comments about a someone calling "short final" 3NM out but agreed that seemed a bit far even for the the point I was trying to make so I split the difference between what the calculator says a wings level at 500-600AGL, 3 degree glideslope "normal" final should be and that, ending up with 2.5NM. In reality 2NM is probably about the widest I'd be willing to tolerate from even a beginning student without a darn good reason (e.g. traffic ahead).
Ignoring the actual distance involved for a moment, understand that I was trying to illustrate the point that in high traffic scenarios with many planes in the pattern, a wider pattern with normal final distance isn't lazy but rather good judgement compared with flying a normal tight pattern and having to extend way long of the runway.
Again I was trying to sufficiently illustrate the point while not getting stuck in the weeds about what constituted a wide/long pattern vs a tight pattern. Clearly I failed to do that since we're still discussing exactly that... I blame any ambiguity on the OP for not defining what he considers an "engine-out B52 pattern" which is somewhat contradictory while complaining about long patterns... I kid I kid. Seriously though, read through the comments and its clear there is no "accepted" distance to fly your pattern and what constitutes wide or long or both for one person might be normal for someone else. About as close as I think we can get for guidance is the circling radius of 1.3-1.4NM for a CAT A approach.
Regarding the specifics of a "normal downwind" being 1/2 mile from the runway? I'd say it could be depending on your plane but I'd say that's probably really tight for most GA. At 30 degree bank and 70kts, a 180-degree downwind to final turn will take you 1500ft or 0.25NM... Which leaves about 10 seconds for base with no wind.
I'm usually a decent bit faster on downwind (90) and base (80) which increases the turn radius/diameter and that's still assuming your 30 degree turns are "tight" and on point going right to 30 degrees and holding it there to the 90 degree point in your turn. If you roll into it more slowly and/or roll out of it slowly with maybe a lift of the wing to check your position in the turn, you'll eat up more of that time... Which does not leave a lot of time to level, take in the site picture, reconfigure (2nd notch of flaps, trim, airspeed), descend another 200-300ft (somewhere around 40-45 seconds at 70kts and 3-degree glideslope) and make a radio call out, even as a competent pilot, let alone a student pilot.
Sure you can do some of this in the turn or more aggressively (a 6-degree glideslope at 70kts is a 700fpm descent which puts you right around that 10-15 second base mark) but its still tight.
Its been awhile so I'm having a hard time picturing the TLAR sight picture I look for and its not like I have some hard and fast rule of "I fly a pattern x miles from the runway" but looking at my track logs for my last couple of flights in which I did pattern work in PA28's and measuring the distance, my typical base leg was somewhere around 0.7 and 1 NM between downwind/final tracks depending on the wind. I did not have any significantly shorter than than 4000ft (0.66NM) with my widest pattern being around 1.1-1.2NM. Most of those were practice power-off 180-descents for my CPL too; not that having power should change my pattern but if there were ever a time you were going to cheat inward towards a tighter pattern, it'd be when you know a power failure is coming...
That being said, I'd say planes and airport particulars matter too. The pattern and landmarks I used to fly in the Citabria at my local drome when I lived out in California is permanently etched in my memory and was a good bit tighter than what I usually fly. I was coming wings level on final at 0.25NM from the runway threshold, my base leg would be 0.5-0.6NM from the runway threshold and my downwind was between 0.25 and 0.5NM from the runway centerline... There were geographical, obstacle and noise abatement reasons for the higher, tighter pattern and of course efficiency was the name of that game as a tailwheel, somewhat remarkably (sarcasm), flies almost exactly as a trike in the air so it was takeoff and landings we needed. In that plane there were no flaps, no transition to manage, a super simple though coarse trim control and we were pretty much chopping power close to mid-field and then adding power back in to manage descent rate once we got the airspeed dialed in.