I'm a little bit leery about certain implications here. Specifically, of PCs using swarmanoid morphs routinely having multiple stacks.

Maybe. I am not opposing the idea as such. But, it does seem to me that the basic ... uncrackability of such a set-up is being taken for granted. What with the extensive hacking rules and AIs and rapidly advancing / evolving technologies and so forth, there would be very few things that with sufficient skill and motivation, would be unhackable. One also has to ask, if the swarmanoids' multiple stacks are all coded to prevent unauthorized access (so to speak), then why aren't the single stacks on other morphs enjoying similar protection? For that matter, if a swarmanoid has multiple interconnected stacks, I don't see any reason why a variant of this could not be done with other morphs. One could have a stack in the usual place, and several "back-up" stacks. The latter would not necessarily have to be implants, either - could be set in jewellery or clothing or other personal items. OK, this last would be very handy fpr most PCs, and may well be where the technology goes next anyhow but, in pure game terms, could also be very unbalancing if introduced right now - rules rewrites, rethinking the entire basic set-up, etc..

Ah well - this is no different to remote scanning any other morph with a stored cortical stack (everyone). Right now an enterprising thief can remotely scan my travel card and access my account and balance. Why not any other morph with a stack. What I proposed was to share the stack in order to address the "where is the stack stored" question. We assume that a stack is encrypted and firewalled to prevent access - if access is compromised then the stack can be copied or read. Obtaining elements of the stack (full or partial) does not guarantee it can be read - moreover, there is less chance because there is more chance of partiality. If one lops of another's head, then the full stack is available to be hacked and the stack withdrawn intact. If one takes part of a swarm, then the chance of a full stack recovery reduces with partiality - perhaps linearly. Returning to the "many redundant stacks within a swarm" proposal, each stack level (or redundant layer) is encrypted (and perhaps differently encrypted) so that an inspected element is incomplete, and a significant swarm-portion would only render an incomplete stack at each level - a different incomplete stack. This "remote viewing" is no less different than scanning a non-swarm morph for its stack, or more easily, scanning a stream during stack recall, or perhaps maliciously, interfering with a recall and slicing a recalled stack with malicious code. Perhaps every stack recall out there is not the "true version" - every recall process is infected with extraneous code. Oooh how's that for conspiracy? What you thought happened on the last mission is not what actually happened - your "soul" was hacked on the way back. But you will never know. Who protects the backups? Are they true versions? True, if we vacuum up a swarm we can take the stack. But we still have to crack it/them, just like any other.

hewhocutsdown: I like it. I am not a fan of "don't know - make a roll" but if you were to introduce a mechanic it seems like a penalty to swarminoid morph players, and not others. Rather than a general % based on swarm loss, I would entertain a specific mod to a separate action: eg. stack recall :- penalty = ratio of current mbs to total mbs (as percentage)... you've got 300, you started with 500 so a neg 40% to the recall attempt ... all without going into specifics of how many microbots contain how much of a stack, how many bots are there in a swarm, how many layers can be stored, how many, how many... I would like to keep it simple. I like your percentage idea regarding stack related actions.

So only Swarmanoids lose their stacks when they die?

I've always just pictured that the stack was kind of like the queen of the swarm...

I figure morphs can have more than one cortical stack, but how it's linked would vary; it's probably not a standard implant. As for Swarmanoids, I figure there's no single model type out there, and it could be up to players or the GM as to which it is. A "queen" drone could be there that's more armoured than the rest and carries the stack itself, or there could be mini-stacks in every drone, each recording a certain part (and just assuming that any the PCs can recover would be sufficient, unless story demands otherwise), or there could be multiple full stacks that all contain back-ups. No reason why you can't have whichever as the situation demands it.

I've ruled in my campaign that the swarmanoid PC has holographic ego storage evenly distributed across the swarm, with 20% of the swarm being the minimum needed for a full-resolution copy of the ego. That is to say, the entire ego is stored on each bot, but with a resolution proportional to the % of bots recovered, under a certain %. So if at least 20% of the swarmanoid fragments can be recovered (a few large handfuls) he'll be able to be restored fully with the production of more microbots. If a smaller proportion of the bots are recovered, his personality will be degraded upon reconstruction and will take some length of time to restore itself to maximum values. That last part is pretty generous, since I had considered ruling that if fewer than 20% of the bots are recovered he may only be reconstructed as a beta or gamma fork, depending on the amount, but that seems too punishing. I prefer the idea that the ego can slowly resolve itself back to 100% from a remainder of at least ~5-10% of the bots. Everything should be evenly distrubuted although I suppose player preference could put emphasis on "every bot is equal to every other" vs. "I have specialized bots in the swarm."

Well if the dev could take the 30 seconds necessary to answer what could have been in the book in the first place, it would be nice too :)

A swarmanoid does not consist of nanites. There isn't a 'hive'. The approximate size of individual units in the illustration looks to be on the order of a centimeter or two. In order to NOT give swarmanoids any benefits or deficits which aren't listed in the books, I tend to rule that a 'dead' swarmanoid actually is 'too few units to process coherent thought', so the remainder shut down. The 'stack' is 'grab a handful of what's left of Bob and let's get out of here.' So there's no 'It will take you a quarter hour to sift through all of the mechanical bugs on the floor to find the queen', etc. Similarly, if swarm elements get cut off from the main (no ability to signal) for more than a few seconds, they instant-wipe themselves, so there's no 'The bad guys just captured 10% of Bob and can now reconstitute him.' A blast door sealing off 10% of a swarmanoid is equivalent to a normal morph losing a limb to same blast door, etc...It doesn't include a stack. The way I handle swarmanoids and flexbots, neither of which have particularly clear or coherent rules, is that they have the ability to string themselves out and extend their field of view across multiple rooms, down corridors, etc. Effectively your morph is a bunch of tiny little drones that must remain in connectivity with each other (or larger drones for flexbots). They can pass through gaps that a full-size transhuman can't. The biggest distinction I make between the two is that the swarmanoid can't manipulate items, and flexbot units can. So you sort of have a 'distributed attention' that lets you be in multiple places at once, within limits, but other than that I don't give them any benefits/deficits that aren't RAW. I'm flexible within reason for things that 'make sense', but as far as cortical stacks, et al., the less fiddliness, the better, IMNSHO.

Having units auto-wipe themselves if they get lost is a glaring security hole. Have an effect that fools them (like an EMP destroying their comms) and the morph wipes the ego... Having units not wipe themselves is also a security hole, since now enemies can capture part of you and get a copy of your ego to interrogate at their leisure. I think the solution is cryptographic secret sharing - the ego is encrypted, and you need a certain number of correctly signed swarm unit processors to run it. I would also suggest that some swarm units probably do have small backup cortical "slivers" that allow reconstruction. But maybe this is the expensive model: you can get a stack-less cheaper, and an insecure one really cheap...