License Change

[danbopes]

Optera!

I love LTN, and the amazing work you put in to this great mod. I use it pretty much in all my saves, and it really has been a life saver. After seeing a lot of your recent posts mention that you no longer play the game, have you thought about making your license a bit more permissive so that others may continue on/improve LTN even further? There have been times that I would have liked to have contributed or dug into the code, but the license is extremely restrictive, and makes future developers wary of developing anything for LTN.

[Optera]

Sorry, but I don't feel like letting go of a project I spent so many years on.

However LTN would require yet another rewrite from scratch to make use of all the current base features.
Even the premise of improving performance by cutting down the total number of trains seems to be wrong. I've heard current Factorio has worse performance with trains running empty 2/3 legs of the trip compared to parking full trains at loading station to be used directly from there (at which point I see no point in using a mod anymore, combinator logic and base mechanics would suffice).

[ssilk]

What I see is this:
LTN for bulk transports, move more or less material from A to B on a regular base: not a good idea. You can have much fun in doing this with the enhanced trains in 1.1 very well. Much more efficient, much faster (CPU & total time), much less possible errors.

LTN for one-time transports, like building outposts, move items for construction, irregular supplies of low number of items and so forth: yes! That’s the perfect usecase.

But for that the building of stations needs to be much, much simpler. Much more streamlined. I would reduce the possible configurations to one use-case. No further combinators needed. Requester station knows what’s needed (looks into requester chests and/or the constructions), provider station knows what is available (provider/buffer chests). Make misconfiguration impossible by reducing the options.
Add a train-station-builder and this is suddenly the simplest way to expand to new outposts.

But I doubt you will do that. that’s ok, I understand it.

[Optera]

What I see is this:
LTN for bulk transports, move more or less material from A to B on a regular base: not a good idea. You can have much fun in doing this with the enhanced trains in 1.1 very well. Much more efficient, much faster (CPU & total time), much less possible errors.

LTN for one-time transports, like building outposts, move items for construction, irregular supplies of low number of items and so forth: yes! That’s the perfect usecase.

But for that the building of stations needs to be much, much simpler. Much more streamlined. I would reduce the possible configurations to one use-case. No further combinators needed. Requester station knows what’s needed (looks into requester chests and/or the constructions), provider station knows what is available (provider/buffer chests). Make misconfiguration impossible by reducing the options.
Add a train-station-builder and this is suddenly the simplest way to expand to new outposts.

But I doubt you will do that. that’s ok, I understand it.

Remember the concept of LTN was made back in November 2016, back when Factorio was a very different game.
That's why I urge to start fresh with a different concept.

PS:
You where among my first beta testers: viewtopic.php?p=223899#p223899

[mmppolton]

My options on it is I still see a use for ltn for bulk delivery I have issues with train limits even use pre-made blueprint too many train in system and you get issues run low on ore and you get issues with not able to send ore / plate to mall first

[robot256]

I have experimented with another situation where train limits don't work at all, and adapting LTN for it would be fairly invasive. I'll accept that a rewrite is probably in order for many reasons. The details will come down to how much of the LTN "look and feel" and internal structure can be safely mimicked under the current license, for the parts where LTN has found the simplest way to do things. If I ever get time to work on it, that is.

[Qon]

I have experimented with another situation where train limits don't work at all, and adapting LTN for it would be fairly invasive. I'll accept that a rewrite is probably in order for many reasons. The details will come down to how much of the LTN "look and feel" and internal structure can be safely mimicked under the current license, for the parts where LTN has found the simplest way to do things. If I ever get time to work on it, that is.

You can make an exact replica of functionality, as long as you write your own code. You just need a new name and new graphics for the new mod. The mod has a license for it, not a patent for the idea of making trains into logistics robots.

The "internal structure" is going to differ at some parts if you aren't copying. The internal structure is going to match to some extent if the new mod does the same thing, but no one can say you can't have similarities. That's inevitable when it does similar things. Not sure what "mimicking internal structure" means, is it looking at the source code of LTN while writing the new mod because you don't understand how to write the new mod yourself? Then you are just copying source code, which is not ok. If you both just happen to find the best way then that's ok. It would be suspicious if the code looked exactly the same though for larger parts. Both mods are going to need to handle the same events, in mostly the same way, and store mostly the same information etc so many code similarities are impossible to avoid, which means they are completely ok to have.

I'm not a lawyer.

[Optera]

You can make an exact replica of functionality, as long as you write your own code. You just need a new name and new graphics for the new mod. The mod has a license for it, not a patent for the idea of making trains into logistics robots.

The "internal structure" is going to differ at some parts if you aren't copying. The internal structure is going to match to some extent if the new mod does the same thing, but no one can say you can't have similarities. That's inevitable when it does similar things. Not sure what "mimicking internal structure" means, is it looking at the source code of LTN while writing the new mod because you don't understand how to write the new mod yourself? Then you are just copying source code, which is not ok. If you both just happen to find the best way then that's ok. It would be suspicious if the code looked exactly the same though for larger parts. Both mods are going to need to handle the same events, in mostly the same way, and store mostly the same information etc so many code similarities are impossible to avoid, which means they are completely ok to have.

I'm not a lawyer.

I mostly agree with those statements.
However, coming up with a way of automatically dispatching trains for current Factorio shouldn't end up even remotely similar to what I came up with several years ago.
Trimming down some feature bloat and using base game train limits, half the data structure could be removed.

[Qon]

However, coming up with a way of automatically dispatching trains for current Factorio shouldn't end up even remotely similar to what I came up with several years ago.
Trimming down some feature bloat and using base game train limits, half the data structure could be removed.

Yes, since

a rewrite is probably in order for many reasons.

if the rewrite is too similar under the hood and especially in the places it couldn't be unless you copied then it's a total failure as a rewrite and an obvious copy job.

Really, if you haven't read the LTN source code thoroughly then you wont know which are

the parts where LTN has found the simplest way to do things

and you wont even have to worry about being too similar. If you don't read/copy the original then there is no risk of looking too similar because similarities are only really evidence of copying when parts that couldn't be similar by chance are similar anyways, roughly speaking.

The only tricky area is if you have studied the source thoroughly and can't really solve the coding by yourself so you might remember better than your ability to code from scratch and you can't really be trusted to know when you are writing down something you remember, maybe mistaking it for your own idea. It's a lot easier to read someones code and write your own original code after that does the same thing if you can write even better code. But why are you reading someone else's code then if you are that good at it already?

[ssilk]

What I see is this:
LTN for bulk transports, move more or less material from A to B on a regular base: not a good idea. You can have much fun in doing this with the enhanced trains in 1.1 very well. Much more efficient, much faster (CPU & total time), much less possible errors.

LTN for one-time transports, like building outposts, move items for construction, irregular supplies of low number of items and so forth: yes! That’s the perfect usecase.

But for that the building of stations needs to be much, much simpler….

Remember the concept of LTN was made back in November 2016, back when Factorio was a very different game.
That's why I urge to start fresh with a different concept.

PS:
You where among my first beta testers: viewtopic.php?p=223899#p223899

Of course I remember that.

This post was about how LTN could be seen and used (tested by myself) nowadays. It works, and what’s most important: it’s still fun to play it like so. Which means that direction is not wrong. I just wanted to give a helpful orientation. Which is in my understanding, that there is no big need anymore to automate bulk-transport. But (my opinion) there is a need for transporting many different things, making build-supply of outposts.

But I’m keen on what you seem to fabricate now.