I think I have a decent handle on the 5th edition matrix rules, but I've got a few questions, particularly with sprites. I suspect a lot of these don't have an official answer right now.
At the moment, the matrix can be thought of as a massive "grid space" full of device icons that roughly correspond to the location of devices in the real world. There are other icons as well: file icons appear in/on the device icons, but can not exist in the grid alone, host icons are massive and typically floating in the sky, although some (like the host for each local Stuffer Shack) appear lower to the horizon corresponding to their real world counter part's location. There are also Persona icons for active matrix users, agents, and sprites, which zip around interacting with icons and coming and going out of host systems.
Device icons are tied to the location of the device they represent, but persona icons can move around. Most persona icons replace the device they are based on, so a decker or agent running on a cyberdeck replaces the deck device icon with their persona. Living Personas and sprites do not have a device, so they are pure persona or nothing. Living Personas do still have a physical location, while sprites do not.
You can detect non-silent running icons for devices within 100 meters of your physical location. The rules are explicitly device icons, and it's measured from your physical location, not where you persona has zipped off to.
Unlike astral projecting characters, there is no movement speed for Personas. If you succeed on the matrix perception test to locate an icon beyond 100 meters, or automatically locate a host, you just sort of zip straight to it (or maybe pull it to you), but you don't get to see or interact with any of the other local device icons that may be near that icon because it's more than 100 meters away from you physical body... unless of course you spot them or already had them spotted, as you can always see/find icons you already have spotted, especially those you have marked or are the owner of.
Question 1: If you located the commlink icon for the target you intended to extract, you wouldn't be able to look around at the local grid space to see icons for his body guards or security on his building (unless you already had them spotted,) although you could attempt to look for such things with a matrix perception test?
Question 2a: A decker decides to attack a rival's persona. They aren't within 100 meters of each other. The decker makes a matrix perception test to locate & spot his rival (lets say in gridspace, not a host) and throws a dataspike at him. A successful attack action alerts the target on a success, but do not point point or describe the attacker. Even though our attacking decker isn't running silent, the victim of the attack doesn't know who attacked him or where it came from?
Question 2b: If our attacking decker had tried to mark his victim with brute force instead (maybe to get more punch of the dataspikes), the victim could look at the mark placed on him, and if he recognized it, could then theoretically make a matrix perception test to locate/spot his assailant and then fight back?
Question 2c: Two deckers sitting at oposite ends of a bar decide to have a matrix duel. Although they are within 100 meters of each other, unless they had previously spotted the other's persona icon, they can't automatically see each other in grid space because persona icons replace device icons, and the 100 meter rule only applies to device icons?
Question 3: Sprites have no physical location, so therefore they can't automatically detect icons within 100 meters ever?
Question 4: Sprites also are their own persona with their own marks/etc so they don't automatically share spotting information with their owner?
Question 6: Agents are software run on a device (such as a Deck, RCC, Drone, or Vehicle) and therefore have a physical location and can spot icons within 100 meters? An agent program on a drone could wander a building scanning for hidden icons, for example?
Question 7: A registered sprite can be orders to use a sprite power on an icon, and this counts as 1 task which lasts until it's instructions are changed. Sprites don't sleep, and registered sprites don't accrue an OS score simply for existing, so the sprite will continue it's task indefinitely? I mean, eventually the device icon might go offline, the sprite might be attacked and destroyed, and even asking the sprite to change between hidden and public mode would be a "change," but in theory, it could go for a LONG time.
Question 8: Icons are things on the grid, not nodes you can be inside (like they were in 4th edition) so if you ordered a sprite to maintain a power on a device, it would always be next to the device icon in grid space, but because they have no physical counterpart, this doesn't make it possible to spot them within 100 meters, and because they have no device, you can't spot them with signal scanner at all. In short, they are only possible to spot if you specifically go look for a specific one with a matrix perception test, or if they are in a host system with you?
Question 9: Sprites lack a physical form so they can't ever use the direct connection exploit to help gain entrance to a host system?
Question 10: Device icons in a WAN are located inside a host, and are not accesible from outside the host, so for a Sprite to interact with one at all, it would need to enter the host normally (getting a mark on it one way or another) or if it's registered, fade back into the resonance, have you (the technomancer) enter the host, and then recall it from the background resonance... or of course be compiled in the host in the first place.
Question 12: Tracking or having GOD descend on a sprite gives away the location of the technomancer that compiled it. Because sprites don't run programs, the Stealth cyberprogram can't prevent this. Does a Technomancer with the Echo of the Stealth program gain some protection from this, or does that only apply to track/convergence on him and not his sprites?
Question 12: You can automatically spot all the device icons within 100 meters of your physical space, but it's limited to device icons, so you can automatically see the commlink of the guy sitting on the bus three rows ahead of you, but you can't see his file icons unless you make a matrix perception test to look for them, because they aren't devices.
Question 13: The programs running on a device you use to generate your persona are still visible on your persona (with a matrix perception test), otherwise how could they be targeted for a the Crash program action?
EDITED HERE
Question 14a: It's been said that technomancers can't gain the benefits of a VCR or Datajack because they use a living persona, but is that actually spelled out in the rules? As long as they are generating their "Living" Persona from their brain, and not a device like a Cyberdeck, Commlink, or RCC, shouldn't they be able to make use of implants they paid essence for?
Question 14b: Assuming the answer to 14a is "no where in the rules is it banned," couldn't they also therefore use a Trode-Net & fiber optic cable to get a DNI direct link to a device they are trying to hack?