Factorio Forums

TL;DR

A mode to manipulate Named Logistic Group

What?

Naming operation:
Input
-Named Logistic Group
-Signal(s) and/or strings
Output
-Rename the group to the input string with/without signals as formatted.

Write operation:
Input
-Named Logistic Group
-Signal(s)
Output
-Update the existing signals with the input values. Ignores signals not present in the Named Logistic Group. Signals with null values are still in the group (just like they are currently).
-Multiple write (collision) to the same Named Logistic Group is resolved by summing up the values of writes for each signal
-Does not change import origin (import from) or minimum custom payload

Payload operation:
-Named Logistic Group
-Target signal (note the singular)
-Minimum payload (constant or a value of a signal)
-Origin (one of the qualified planet)
-Update signal (a signal that is not the same as the target signal with a value > 0)
Output
-Target signal outputs a constant value of 1 if setting is successfully updated and a -1 if not. The update happens only when the value of the update signal is > 0 and must be present for at least 60 ticks.

Why?

There is a lot that can be done when the values of the signals in the Logistic Group can be manipulated...Interplanetary signalling, intra-planetary signalling, manipulation of constant combinator, true channels, a new type of memory, and many many other ingenious use.

That would be a nightmare to manage.
I like your reasoning (partially) and it think it's (partially) reasonable enough but I disagree with the proposal.

• Interplanetary (including interplatform) communication: Strong +1
• True channels: +1
• A new type of memory: Why?
• Manipulating constant comb: Strong -1. They are named "constant" for a reason

See also:

• 121255: signal sharing between planet and orbit
• 121440: Advanced Circuit Network Functionality to Solve Space Platform Logistics

Hares wrote: Fri Dec 27, 2024 10:02 pm
[*] A new type of memory: Why?
[*] Manipulating constant comb: Strong -1. They are named "constant" for a reason

The reason it's a "new" type of memory is because the constant combinator always output whatever values it has. So when some signals in a logistic group are manipulated by some selector combinators, such change would percolate through anything that uses that particular group, and if it happens to be a constant combinator, well, you get the picture? These are just side-effect of logistic groups being accessible in constant combinator. Of course, I agree it doesn't have to be the case...

If they take up this idea of allowing some manipulation of the logistic group, they might want to isolate the group from constant combinators if they truely want them as "constants"...I have a suspicions that behind the scene constant combinators are really not immutable. They're just acting as a "vessel" of constants for the sake of the gameplay. Code-wise, manipulating them probably weren't threaded because they were meant to be "constant". Likewise, the logistic groups data life cycle were probably modeled after the constant combinators.

I'm under no illusion that this is a small ask... It would be nice to not have to make a mod for this functionality.

On the up-side...think of what you can do with a "constant" combinator whose values change...Oh my, one could re-encode the factory that is under control of a set of logic...Parameterization of logic in real-time

Fatmice wrote: Sat Dec 28, 2024 2:12 am

Hares wrote: Fri Dec 27, 2024 10:02 pm

• A new type of memory: Why?
• Manipulating constant comb: Strong -1. They are named "constant" for a reason

The reason it's a "new" type of memory is because the constant combinator always output whatever values it has. So when some signals in a logistic group are manipulated by some selector combinators, such change would percolate through anything that uses that particular group, and if it happens to be a constant combinator, well, you get the picture? These are just side-effect of logistic groups being accessible in constant combinator. Of course, I agree it doesn't have to be the case...

You don't get it. I was asking "why?" and not "how?" nor "what?". It is obvious based on the original post on how can this be a new type of memory.

Speaking of "constanticy", they are definitely not immutable as changing shared named group changes all instances of it. The real application is the fact the only way its output can be changed currently are user interactions – that's why they are "constant". And you can in fact have similarly behaving "memory cell" – at least within the surface – and only single channel unless you're doing wacky bit manipulations – by using radars.