[pgrey]

I was sort of the mindset of trying this, and just letting it run initially, as some have mentioned.

Then, I had a day (also in a suburb of Seattle, mostly), where it was cold, and traffic was pretty plugged-up, and I swear it did an auto-S/S about 75 times, in maybe a 25 minute drive, and I sort of snapped. I've replaced a few starters in the past, and even when they were pretty exposed, they were almost never easy. Simple, yes, but something was always wedged or similar.
Nowadays there are probably 20-50 other miscellaneous pieces that have to come off, first, and the starter isn't probably 55+core, or something like that, I bet it's 755, or some such. There's got to be some sort of wear/tear on the battery, too, it's hard to imagine it not being the case, say if you do 900 more 2.5A drains/month, it's going to add up, slowly, but surely.

I think if I'd seen more of a fuel-savings indication, I'd have been more inclined to persist.
Personally (given the spirit of the idea anyway), I think it should (also) tell you an approximation of lbs/CO2 you're not outputting, it's got to be a much more impressive (looking) number, and more motivating, for some (it could simply be another item-selection, doesn't have to be either-or).
In the meantime, the person who's spent too many hours swearing at an older, stuck starter has me turning this feature off, whenever I remember (or at that first stop).
The cruise-control option someone mentioned is intriguing, given the issues with running rear-defrost constantly...

 

[pgrey]

What worries me is that I think it's restarting after a few seconds because the battery voltage is drooping. The alternator starts charging again, voltage seems good, the engine shuts down at the next opportunity, then it notices the voltage is low so it starts up. The system should be smart enough that if the battery voltage is low just once during an ignition-on session, it should continue to charge for the duration of the trip, instead of constantly shutting off to just start again a few seconds later because the battery is weak.

This part is pretty electrical-physics-101, at least in theory. If it sees a significant voltage-drop, it sure seems like it's either a lot of draw (lights, heater, etc, needs to keep running regardless), or the battery needs a bump. Seems like the system could do a decent job of calculating some reasonable of charge/time, and use that as a minimum, even if it's across multiple (manual) restarts.
How long it actually waits, before continuing to cycle again, is an interesting question.

I wonder how many people who have this (or a similar system, in any car) and get a jump-start, only to realize that they drove two blocks, stopped at a stop-sign or whatever, and oops ;-]

I'm a huge EV proponent (hybrids, too for the right profiles), but this sort of attempt to be a semi-hybrid-sort-of-not-really, I'm not so sure...

 

[pgrey]

Agreed. Having to take time out of your day to go to the dealer to have this done is ridiculous. I just updated the Sync system firmware in my 2015 F-150 from a USB drive, why can't Subaru let us do the same?

Yep, this has been my chief complaint with these darn TSB's, from the start (see some of my earlier comments).
They should be able to do a WiFi update, or a USB update, just confirm an MD5 hash, before the actual update starts (avoiding any sort of mis-matched update), and if something fails, revert to the previous (I'm still fairly certain the h/u has this capability, built into most system-BIOS for 20+ years now...).

Or, we could ALL have to schedule a drop-off appointment (my dealer/state hasn't had while-you-wait appointments since Mar...), for all of these infotainment updates. I did some math, in one of the previous posts, and it would pretty much entirely fill the service schedules, of every dealer, if everyone has to bring in their 2020+ cars, for updates, at 1-2/year, so yeah...
Or, say if you live in some far-flung corner of a state, maybe 150mi from the nearest dealer, that's 600mi round-trip, yearly, for a couple of updates, hmm.

I get where people are thinking about the abundance of caution, given that the controller ties into things like RAB, but I bet it does on a LOT of cars that do OTA updates, most of them, my bet anyway.

 

[pgrey]

Haven't read the whole thing but this cought my eye:

View attachment 486871

This is how Subaru should've done it, but they went conventional, good old ring/pinion.
If you read reviews of cars that use the belt-system, they're much smoother, given the inertia of this type of system.
I just don't see how the parts aren't going to wear faster, it's a system that's been around since at least the 50's, without all that much change, except for some better timing, on the engagement/spin-up, of the starter (this certainly brings wear down, but by by how much I've never seen facts/figures, on this side).
How much faster will it fail, overall, well, that's going to take a bit, to see...

Either way, I still dislike my auto-S/S system, particularly when I have the brake applied and it restarts, it still jolts-to-a-start. If I'm going brake-accelerator, it's MUCH smoother (I would bet this is a CVT engagement-timing issue, but that's just a guess).
A mild electric-hybrid design would solve the whole deal, without a complete redesign, one can only hope...

 

[pgrey]

Subaru programers set it up this way for a reason - in order to achieve higher MPG ratings (and subsequent lower CO2 emissions) from the EPA. Sure, Subaru can change it, but they won’t. Hopefully they reduce the hoop jumping it takes to temporarily disable it, but they won’t make it defeat-able permanently.

True, but they could allow for a shortcut, on the main screen (one that doesn't disable when going > 5mph), that'd be enough to satisfy me.
I also wouldn't be surprised if a 3rd-party OBD or similar device comes out, that would disable it automatically, at start-time, in the longer-run...

 

[pgrey]

That makes sense because otherwise the car would have worse emissions and fuel economy in certain test cycles, but I do wonder if it takes into account the extra battery energy that would eventually need to be re-captured by the alternator. The battery before and after the test cycle would need to be at the same state of charge for it to be a fair test.

It seems like a test, say on a track, where you stop/start 20 times, say 5 laps or something, it'd be as easy as measuring the fuel-use (using an interim-feed gas tank or similar, for accuracy there)?
I wonder the same, lots of factors here, perhaps the savings/emissions-results are negligible, on days where you're using the AC or heater, say 90% of the drive.
Testing w/o climate-control seems like cheating, to me, but I haven't gone and read the CAFE test-spec or something ;-]