Here's several replies, rolled into 1 message....
I'd just wonder if it's worth the time and effort put into it, as it's only cost effective (especially with the added cost of the pre-programmed bootloader chips) if you make a LOT of them at a pass - and it seems like the regular Teensy 3.1 is enough for the majority of people.
This is exactly my thinking, except we program the mini54 chip.
It easy to say "just redesign ... and be done with it".
The actual engineering doesn't take too long, maybe a few days to lay out the PCB and build/test it. But when you add on building a new bed-of-nails test fixture, ordering PCBs and scheduling production of product, increasing the ongoing complexity of scheduling parts far in advance so no Teensy goes out of stock, and lots of other little behind-the-scenes things that all have to happen correctly... it's quite a bit more involved then "and be done with it". It's all do-able, but it adds up to quite a lot of work.
I think ongoing product manufacturing, from forecasting sales to purchasing to inventory management to actual production to testing & packaging to actual shipping as sort of an art form. Much like olympic figure skating, done flawlessly, it looks easy. What you don't see is an incredible around of unglamorous hard work. It's very tricky to get right. If you look around most of the hobbyist hardware market, lots of products go out of stock regularly. Robin and I are actually pretty good at it... but I can assure you, that's not some effortless natural talent. It's a lot of work to do well.
The 2 biggest concerns are the financial investment to make large enough batches for Teensy's normal low pricing (on a model that's unlikely to sell much), and spending time working on this, rather than things many people want much, much more... like a debug version and a ++ board with that new 180 MHz part and new USB types and about 100 feature requests on the audio library and adding write support to the improved SD library and a biopotential sensor library and setting up a wiki and someday even putting serious work into really improving the website. Time is the really precious, limited resource. Releasing any new product, even something seemingly simple, uses up quite a lot of time that necessarily delays other stuff.
After all, PJRC is a commercial enterprise and most companies listen to the wishes of their users/customers.
I sure do try to listen. The hard part is fitting the massive number of very different requests into a fixed number of hours per day. Only a small fraction of things everyone wants can get done. I have to (try to) choose wisely.
But, to be honest: I needed these bottom-pads only one time (two of them).
This is also a pretty common response I hear. This, and that it's less convenient but still not too difficult to get to those bottom side pads when necessary.
From PJRC's perspective (I'm assuming a lot of things here) it probably makes more sense to focus on the ++3.1, as opposed to devoting development time to the 3.1.
So far, this has been my plan. Well, and a debug version. At this point, real debug capability is by far the most commonly requested feature.