Solution to many problems and general improvements. Issue 1.

[Goury]

So I played Factorio for thousands of hours and started to realize some design problems that is easy to fix and now looks obvious.
It wasn't obvious to me few years ago, so it may be not as obvious for one who haven't experienced what I experienced.
Sure I greedy bastard and want my share, but I grant permission to use those Ideas on terms of CC-BY-SA 4.0 license.
Should devs wontfix this, feel free to develop a mod.

The problem: weapons are useless.
Sure new flame thrower feels good and ramming spawners with tank feels fun, but let us face the truth: compared to turret creep it's all useless. Combat shotgun can be kinda useful if upgraded to the top, but pistol and automatic rifle is just waste of time and resources.
Adjacent problem: robots are overpowered.
The solution: revisit robotics.

Complex description:
Have you ever wondered why player requires tools to cut trees and robots can do it magically in instant? I did.
The same thing with turrets: they rock just because there is many of them and you are all alone with this tank.
So no matter how cool guns will be, they will never be as useful as something you can have a lot of.

Simple change to robotics can make game much more balanced.
Firstly, let's address the problem that robots is endgame feature and personal ports even endgamer.
Robots should not require flight, ground robots is a thing too.
So let player have very basic robots since early state and let upgrade this slow and useless robots to something more smart and useful later.
Also robots should be able and required to use tools.
There should not be diversification for robots to be combat, logistic or whatever, instead there should be robot parts that allow them to move faster or have more armor or more arms to use more tools or heavier frame to carry more cargo. And if you want to keep robots endgameish feature — AI should be upgradeable, requiring advanced materials for robots to perform advanced stuff.
Thus weapons can be made much more realistic (at least balance-wise) and this will make sense and will be interesting, because now you can equip ten very simple robots with pistols and have chance to use this pistols at long last. And for making rifles you can require advanced machining, because long barrels are harder to produce and more powerful cartridges require firearm to be made out of different materials.
And since robots are using tools, there won't be insta-deforest at the beginning of the game and there will be point to have more advanced tools than steel axes.
Same with vehicles: no need to make tank overpowered, just let there be tech for using AI cores to allow having dozens of tanks.
And make just generic turrets with slots for equipment instead of just three types of turrets.
Make it so turret upgrades are its rotation speed, amount of slots for equipment, hit points and whatever else, but not just raw speed and damage. Let speed and damage be property of weapons and ammo. (Realism-wise: do not upgrade all ammo on researching new tech. Obsolete ammo that produced with ancient tech should remain weaker than new one. Same with guns.)
So imagine turret with slots filled with 120mm tank cannon, few lasers, heavy energy shield, few roboports and factory producing ammo, so you can supply it with raw materials instead of ammo crates.

This way endgame can look like you having army of combat robots (not those little flies, but heavy armed motorized tank-like ones) armed to the teeth, equipped with shields, clearing biters as it should be.
Also this way biters can evolve even more without making it impossible to compete in some situations.

Another issue is roboports nearly useless before powerarmor and fusion reactor can be solved if batteries be allowed to recharge from any power source and not just ones installed in armor.

To be continued in next issues:
Wooden walls and energy walls — because walls is too hard to build quick and not enough for endgame
Robotic factories, allowing more complex compact endgame production by creating intermediate products inside the same factory — because endgame things should not be just disabled by tech-wall
Maglev trains with electric locomotives and armor wagons — because I want to equip energy shields and guns on train too
Fising and foresting — because food is required to live and wood is building material and fuel
Ballistic missile launchers and artillery — because some times you don't want to deal with this big nest but just nuke it
Alien artifacts require live captured aliens — because somebody may want to win the game without killing living beings and because of extra challenge
And many more

[ssilk]

Please avoid all the colors.

And the weapons will be (much more) balanced with coming 0.15. See one of the Januaries FFF's.

[Goury]

Why there is colours if I should avoid it?
Is there a problem disabling the feature that should not be used?

Also January fff's have nothing to do with the balance problem that is in the fact that there is no limit on how many turrets you can build and no matter what weapon you use, it's just single thing.

[ssilk]

Why there is colours if I should avoid it?
Is there a problem disabling the feature that should not be used?

You answered yourself: avoiding doesn't mean fobidden. Using colors makes sense. Sometimes.
As moderator for example I use them to explain changes I need to make in posts.
And it cannot be turned off for users only, I'm sorry.

And why should you avoid it: to improve readability. why do newspapers do not use lot of colors? And cause I explained in the sticky articles in this board.

(Btw: you didn't read them, did you? This is one of those anti-articles that cannot be discussed, cause there are too many ideas in it)

Also January fff's have nothing to do with the balance problem that is in the fact that there is no limit on how many turrets you can build and no matter what weapon you use, it's just single thing.

The point is - and I just want to be helpful with your suggestion - that with those big balance changes this suggestion might become obsolet.

[Goury]

Okay, colours thing sounds like Factorio team don't know howto set up phpBB properly. Should they hired me last year, I’d solved this problem. But I'm not available now.
I'll try to help remotely about this later.

As for sticky articles, I haven't read them all yet.
Some day I'll read them all.
Maybe few days after this day I'll repost this issue (and some more issues) in more readable form.
Or may be I'll post it in my blog and repost it here somehow.

BTW, is it okay to use colors for smiles?
Like this:
／人◕ ‿‿ ◕人＼

[daniel34]

Okay, colours thing sounds like Factorio team don't know howto set up phpBB properly. Should they hired me last year, I’d solved this problem. But I'm not available now.
I'll try to help remotely about this later.

The problem with colors is not a technical one, you could use them as much as you like and it wouldn't break the board. The issue - as ssilk mentioned - is readability. We not only have the default factorio (dark) theme on the forums, you can also use the prosilver and subsilver2 (light) themes and I know some people prefer them. IIRC in your post you used a very light purple-ish color, which would be hard to distinguish from a light background. Using bold text instead ensures that it's readable no matter which theme the reader uses.

Compare your smiley when using different board themes, do you see the difference in readability?

smiley-factorio.png (2.25 KiB) Viewed 3485 times

smiley-prosilver.png (2.42 KiB) Viewed 3485 times

smiley-subsilver2.png (2.42 KiB) Viewed 3485 times

[Goury]

Aww, themes.
Okay.
There is an easy way to solve this by using css styles instead of colors. I've managed forums with pitch-black, plane-white, gothic-red and pure-pink themes and managed to manage colours just fine.

[TI-89]

Ground robots I think has potential, but because it represents a substantial change to the game I don't see it as viable outside of mods at the current time. But I have thought something similar for robot research. I think modular component based research has huge potential once we get 0.15 changes like infinite research and research which varies with difficulty.
One really nice thing ab roboports is they ease the learning curve. It's easy to set something up and have it function, and start getting an idea of how the game works. And it also suggests potential optimizations for a factory -- if you constantly see a solid line of robots carrying a particular resource across your base, it's hard not to think 'hmm I should really do that with a belt/train.'
For beginners robots can function approximately as they do now. But once you've figured out trains and have some idea of best practices/blueprints for station design, you can start a new game at the next level of difficulty. Now (for example) robots will have research for battery capacity, engine power, and energy efficiency. Upgrading components can have complex interacting effects similar to using modules: power increases speed and energy consumption, efficiency reduces speed but increases flight distance, etc. This could also allow tiers of bots which can be used in different networks for different purposes, and could help eliminate some of the things where drones are currently far too powerful. E.g. If you want to use large scale lognet for message passing you get speed and capacity for long flights, but don't need all that much power; conversely if you want to ship around barrels of crude oil (imo most broken current application) your bots need large amounts of power -- up to you if you research efficiency or just eat the power cost.
While my example above starts to move towards micromanagement a bit much, I think the concept of modular research is really cool for showing the trade offs between different approaches to the game. It leaves freedom for the player in designing their factory, and if tiers of items are well structured (not that it wouldn't be complicated) would allow comparison with factories which function in vastly different ways.
For example there could be similar research for trains, where speed and power are customizable with research to achieve desired performance. One player could decide that he only wants to use one locomotive for any train. This will be possible, but will require X levels of research (with a precisely calculable investment), and will eventually require solutions like fueling trains with rocket fuel for them to be able to make it any meaningful distance. Another player could decide that space isn't as big a concern as fuel efficiency and build a ten engine train that could travel for miles on a single piece of coal. But to do so requires Y levels of research to obtain the same throughput as player 1.
Where this gets really cool is if we get more interesting resource availability/ biome generation settings with more meaningful tradeoffs. In worlds where resources are very sparse, efficiency is a must. Then if combat gets tuned up could make things like efficiency is crucial in the desert so as to not be overrun, but not needed in forests/valleys if you clear a biter perimeter.

[kyranzor]

If you want armies of automated, heavily armoured and armed killing robots that go along the ground and don't need player interaction, look at my mod Robot Army.

[Engimage]

I would suggest you to play some Bob's mods.
There all robots require robot frame, logic board and a tool to create. There are also several tiers of bots so you do upgrade them. Combat robots there also require said robot frames so most of your ideas are already nicely implemented.
About ground robotics - a Robot Army mod does the trick.

I suggest you to search for your ideas on the forum. Most of things you mentioned (like artillery) were heavily discussed already and you might add up to those threads.