[Darzil]

Personally the way i tend to justify things is to pretty much handwave Personas as unfakable as others have suggested.

For the rest, I think of the matrix pretty much this way (partial disclosure - day job for last 20 years has been IT, currently dealing with Cloud, new to Shadowrun in 5th) :
Extension of the "Internet of Things" principle of wireless internet for everything, most areas end up getting so many wireless devices that you end up with more bandwidth connecting them to each other than bottlenecking them through routers. This is the "Matrix".
Grids are largely just a shorthand for traffic priority, public grid always gets least priority, though they use something like current day VLAN technology to separate the traffic. Hackers jumping Grids are effectively just hacking the 'VLAN tag', which is why it isn't too challenging. Going between grids does involve going via some kind of router, hence the reduction in effectiveness hacking between grids.
'Noise' is either signal degradation reducing Bandwidth, lack of devices reducing Bandwidth, or amount of traffic saturating bandwidth, all reducing connection efficiency.
Again with the "Internet of Things", there is so much spare processing power with all these devices around going to waste. The ability to allow other devices to use this for their own processing is huge, allowing devices to perform better with less heat output and lower power requirements. There are downsides, though, such as having to be wireless to be anything other than dumb on many devices (pretty much the way that voice control of your smart TV is processed on a server somewhere, not in your TV).
Of course, this also impacts GOD. How do you track down and control the Matrix when illegal activity is distributed? Essentially you can't tell GOD everything that happens, but devices report if they think they are doing something that could be illegal (they won't know, as they are processing only some of the instructions). So once GOD starts seeing several devices reporting in a smallish geographical area, it starts investigating, and narrowing things down. This is why Overwatch score takes time and activity to build, and why GOD don't have records of everything that happens everywhere.

Without going into detail, that's pretty much how I justify it to myself for the basics.

[Jack_Spade]

Give every Stealth RFID chip the ability to run a persona through a modification with two packs (cannibalized from other RFIDs) and you suddenly have dozens of  stealth commlinks.

They're still tags. What you would need is a persona to load into each and every one of them.

Is a tag still a tag if you could run it like a commlink?
Otherwise, if they automatically filter out the tag, you could use it to commit digital crimes without gaining overwatch score.

[Herr Brackhaus]

wrapper always could be beaten by matrix perception.

While that is true, wrapper states that you have to know to look for something to see through it. So if I see a swarm of 20 commlinks, is a single hit on an MPT enough to reveal all of them as wrapped RFID tags? Or, as the text seem to suggest, do you have to pick randomly which icon to look at.

And that is the power of Wrapper and tags. Nothing in the Data Trails entry seems to counteract this, which makes a 250 nuyen program way too valuable since you can effectively conceal your presence in a cloud of digital chaff, for lack of a better term.

And it gets even worse if, like Jack suggests, you make the tags run as a device that has the capacity to host a persona. At that point, it is for all intents and purposes a mini commlink.

Or, you buy 10 burner links and you're no longer talking about tags from a hardware perspective, but actual commlinks. Wrap them as simply higher end devices, and how does your opponent know the difference except to look at each one.

[Novocrane]

Is a tag still a tag if you could run it like a commlink?
Otherwise, if they automatically filter out the tag, you could use it to commit digital crimes without gaining overwatch score.

A device shows up as a device when it's being run as a device.

The fact that the device has a user overrides the device’s normal icon status, turning it into a persona.

There is no default where you can use a tag to commit crimes and still have it show up with a device icon.

While that is true, wrapper states that you have to know to look for something to see through it. So if I see a swarm of 20 commlinks, is a single hit on an MPT enough to reveal all of them as wrapped RFID tags? Or, as the text seem to suggest, do you have to pick randomly which icon to look at.

And that is the power of Wrapper and tags. Nothing in the Data Trails entry seems to counteract this, which makes a 250 nuyen program way too valuable since you can effectively conceal your presence in a cloud of digital chaff, for lack of a better term.

And it gets even worse if, like Jack suggests, you make the tags run as a device that has the capacity to host a persona. At that point, it is for all intents and purposes a mini commlink.

Or, you buy 10 burner links and you're no longer talking about tags from a hardware perspective, but actual commlinks. Wrap them as simply higher end devices, and how does your opponent know the difference except to look at each one.

This is why I believe bug scanners are more common in secure locations than is otherwise presented. It's the simplest way for security to pick out suspicious clumps of wireless devices and pinpoint what they need to inspect.

[Jack_Spade]

Well, you can slave your stuff to tags to benefit from their 3 points of sleaze without forming a persona on them.

[Gratuitous Boom]

Instead of wrapping stealth tags to look lite personas, why not wrap your persona to look like a stealth tag?

[Marcus Gideon]

Rather than Wrapper all the Stealth Tags to look like anything, why not just Wrapper the few pieces of Gear you and your team have to look like Stealth Tags? Then you have 110 "Stealth Tags" to sort through, with 10 of them being real stuff.

Speaking of which, lemme ask the class a question.

When someone sets a device to Silent Running... do you think ANY of the device's functions or features should be visible / detectable before you make the focused Matrix Perception check to scan it?

Since the book says if there are several Icons nearby and they are all Running Silent, you have to scan each one at random to reveal it. Which would suggest that it's impossible to say "show me the hidden Icons that are not Tags" to filter out the chaff. Until you scan each one, they all look like featureless grey blobs.

If you do let players see features, then it completely negates the point. Why bother trying to Run Silent and hide from detection, if someone can say "Show me the hidden Icons that are Cyberdecks" and suddenly there you are.

[Kincaid]

Wrapper is a fantastic tool, especially if you know the host's iconography ahead of time.  It's not too hard to defeat if you do something suspicious (which you'll probably do at some point), but it can give you a decent amount of lead time.

And yeah, bug scanners are weirdly powerful.  All my fly spys have them.

[Novocrane]

Since the book says if there are several Icons nearby and they are all Running Silent, you have to scan each one at random to reveal it.

"If you know at least one feature of an icon running silent, you can spot the icon" is the order of the day. Most commonly this will be "I know there are icon(s) running silent" followed by "let's go through them at random!", but if you have certain knowledge of an icon's features other than the fact it is running silent, you can work from that starting point instead. This is where bug / radio signal scanners step in.

[Sendaz]

When someone sets a device to Silent Running... do you think ANY of the device's functions or features should be visible / detectable before you make the focused Matrix Perception check to scan it?

Since the book says if there are several Icons nearby and they are all Running Silent, you have to scan each one at random to reveal it. Which would suggest that it's impossible to say "show me the hidden Icons that are not Tags" to filter out the chaff. Until you scan each one, they all look like featureless grey blobs.

If you do let players see features, then it completely negates the point. Why bother trying to Run Silent and hide from detection, if someone can say "Show me the hidden Icons that are Cyberdecks" and suddenly there you are.

This goes back to the fluff bit about Demi-gods seeing through stealth tags. 
How are they able to screen through the horde of blobs that is a pile of running silent icons to find the actual working devices in in such short order?

Remember when a normal device is running silent, it is still interacting with the Matrix, just in a very limited way.
The stealth tag is just a silent icon sitting there, it has a presence but not much else. 
Your wireless gun running silent mode is still pulling in data or requesting weather updates although in a stealthier manner.

Maybe the Demi's are not looking at the individual silent icons so much as seeing who is pulling and putting out a certain level of data off the matrix.
So to the Demi the stealth tag looks like a rock, yeah it's there but its input/output is pretty static while your gun in silent mode is requesting and getting a weather update every 0.1 seconds so it gets a tiny pulse every so often.
They don't know what that data packet was, but that alerts the Demi this is not just another stealth tag.

Still doesn't answer a lot of things though.

[Novocrane]

How accurate would you consider this 4e setting info in 5e?

When a wireless device needs to pass information to another device in mutual Signal range, it simply sends the data. If the destination is not within this range, for example when you are in the UCAS and trying to speak to Mr. Johnson in Lisbon, the information travels from device to device in a process called routing.

When information is routed between devices, it is non-sequentially sliced into a number of pieces and sent to the recipient via multiple paths; this makes it almost impossible to intercept the traffic except within Signal range of the sender or the receiver, the only places the information is in one readable piece.

The routing functions of a device are handled by a separate component of hardware than the other functions of the device. This makes the routing process invisible to the user, and allows the device’s node to connect to the Matrix even when it is operating in Hidden mode.

[Marcus Gideon]

How accurate would you consider this 4e setting info in 5e?

When a wireless device needs to pass information to another device in mutual Signal range, it simply sends the data. If the destination is not within this range, for example when you are in the UCAS and trying to speak to Mr. Johnson in Lisbon, the information travels from device to device in a process called routing.

When information is routed between devices, it is non-sequentially sliced into a number of pieces and sent to the recipient via multiple paths; this makes it almost impossible to intercept the traffic except within Signal range of the sender or the receiver, the only places the information is in one readable piece.

The routing functions of a device are handled by a separate component of hardware than the other functions of the device. This makes the routing process invisible to the user, and allows the device’s node to connect to the Matrix even when it is operating in Hidden mode.

I thought we'd already established in the other thread, that SR4 was based on real networking, which is why that description of Routing sounds sensible, logical, and basically how Routing really works. SR5 has gone completely off the reservation without looking back. The books say that distance based Noise is based on the antenna of your device, to the antenna on the target. So there is no more Routing, no more hops. It's straight from you to them.