A more intuitive tech tree

[CaptainKipard]

It is based on Comunity Tech Tree mod. Although I suggested wider root of tech tree, that will alow even more customization for various moders. But even without more nodes in career start looks promissing.

I thought that many people missed those threads and could find it usefull.

In the early days of the Community Tech Tree when it was still being planned I tried to steer them towards something like we're discussing here, but unfortunately what they wanted to do was specifically to expand the stock tree, not improve it. We can leave them to their work.

Over 700 posts in less than 24 hrs, on a reddit post asking to "work on the tech tree." Wow. I just glanced. Moving parts around, "it should started unmanned," "No, it's the Kerbal space program," etc.

Edited for visibility.

Also I don't have a reddit account but if anyone reading this does, then please drop the link to this thread and maybe mention a few of the ideas from the OP.

Edited May 6, 2015 by Cpt. Kipard

[grosser_Salat]

I got a quick question:

How do you make the lines go from the the downside of a node to the upside or vice versa?

Parent
        {
            parentID = Stability_T2
            lineFrom = RIGHT
            lineTo = LEFT
        }

Which other values than LEFT or RIGHT can be used?

Thanks in advance!

[CaptainKipard]

I got a quick question:

This is not the forum for modding questions. Ask here.

[Kegereneku]

Will test, so far it look like it's going in the right direction.

Question : Were you capable of doing this one because of the "easier modability" of 1.0, or were you waiting for 1.0 to deliver ? (truth, I want to know if I should keep being so critical against the apparent lack of change in 1.0)

[Yemo]

Summary of the thread

The points may conflict. I tried adding all of the most discussed or supported ideas.

1. Rework the tree to give players an open experience. Rough outline below. Let players follow their own progression.
   
2. No need for a starting node. Give players some science to spend initially. more
   
3. Branches divided by technology.
   
4. Try to put a single part per node, or only strongly related parts. Keep it intuitive.
   
5. Have successive nodes with larger/more advanced parts. more
   
6. Allow modders to dynamically append nodes. more
   
7. Create new part variable for determining part type/function. Useful for filtering. more
   
8. Little or no node interdependency.
   
9. Make really low-tech parts like ladders or thermometers available early. more
   
10. Create an in-game tech tree editor
    
11. The ability to define tier placement per node. more
    
12. Customisable tech nodes / Custom images.
    
13. Customisable tech branch backgrounds / Customisable branch labels (see image below).
    

My opinion regarding those points, for an open discussion:

ad1. The more open a a tree is, the less newbie friendly. It also Drastically increases the chances of dead-ends. When you research planes and rovers a lot but then lack the means to get to the higher science biomes.

ad2. I prefer a starting node so that I know that players have at least some basic parts to get off the ground. Though I like the idea of some starting science, but players can set that already.

ad3. Agreed

ad4. Definately disagree. Most importantly it simply is not practically possible with modded games. Also it would be a clickfest and there is no real reason why a quadcoupler should be in a different node than a tricoupler.

ad5. Yep.

ad6. Not practically possible, would result in chaos and incompatibility, if every part mod adds its own nodes.

ad7. From the tech tree side, that makes sense, from a part mod side, it leads to chaos.

ad8. I prefer the contrary. For me it is the difference between progression tree and science sandbox.

ad9. Totally agree, and girders.

ad10. Would be quite useful. Of course it would lead to a lot more people changing the tech tree, which will in turn result in compatibility problems.

ad11. Science and positions can already be defined per node, tier is just the visulization.

ad12. Implemented as far as I understand.

ad13. I like this very much, I d love to have some simple tools like textfields or colors to bring some more structure into the green buttons with green arrows on blue ground

[CaptainKipard]

If you're going to try to go against the consensus in the thread, you could at least put an appropriate amount of effort into it. Some of it looks like you didn't read the links next to them or even the point itself carefully, some are just empirically false, most of it is unspported by any argument.

I'd like this thread to not be a place to cast subjective votes, but a place to get a consensus, and that's not going to happen if you do what you just did. Make an actual case.

[Kegereneku]

On behalf of Cpt.Kipard (which is assuming too much about what we think) I would like to answer a few points :

I'm testing right now This tech-tree which used bundle, but intelligently.

Although this isn't how Kipard formulated it, I always discussed this suggestion as this. The Bundle used [by the stock-tree are seemingly made to be so "newbie friendly" that choice do not matter. The objective here was to make choice matter in a way that is fun to play, without a need to grind for any part]

ad1 : Ideally as long as you get the basic component to get to any planet, you don't risk a deadlock anymore. It will just be less efficient and more costly than hoped. Having to abuse cheaper probes when you wanted a base...etc

Also, I don't think anyone in Kerbal SPACE program will research too much spaceplane before realizing rocket are as awesome.

ad8 : I think we would all settle on it following a more logical progression (based on technology). The objective is to allow you to research something that fit your (immediate) need without being forced to grind for unrelated garbage in between.

Hence "Little to no stupid interdependency"

ad11 : As I understand, right now Tier (as in building Tier) is based on Research-point cost. This is not ideal when you want a Tier2 part that isn't as costly than a tier1 costliest part. (though, the game can still be well balanced around it)

Edited May 25, 2015 by Kegereneku
why did a portion of my message dissappear !

[Mulbin]

Single-part-per-tech seems nice in theory, but with 260 odd parts in the stock game it just makes a horrid mess of a tech tree.

Time for me to be smug about my tech tree I don't think it's a horrid mess and most nodes contain one part. Not inteneded for mods though as it uses no original nodes... I'll probably right some MM for key mods if requested though.

[Kegereneku]

Time for me to be smug about my tech tree I don't think it's a horrid mess and most nodes contain one part. Not inteneded for mods though as it uses no original nodes... I'll probably right some MM for key mods if requested though.

http://i.imgur.com/K9GL60m.jpg

Gotta say, even if I don't support much anymore the 1part-1node philosophy I'm wondering how it would play out. So far I think the potential problem to be aware of is the risk of deadlocking a player into a bad combinations of parts.

[Mulbin]

That is quite a good effort - visually it's much better than what I'd expected. Will you be modifying the R&D building upgrade costs, or ignoring the restrictions entirely?

Mod support is I think the biggest headache for tree mods.

Since this is a discussion thread, I'll mention one of the problems I have with this approach - fuel tanks. Early game it's fine to pay more for every larger tank size of the same diameter, since you have the 30 part limit on building level 1. But later on (and as tech costs increase) I think it becomes redundant - there is no benefit over stacking two smaller tanks aside from aesthetics.

I will be including a config for sarbians custom barn mod to lower the science limits. Firt tier will only go up to 30 instead of 100.

Mod support isnt really an issue as the purpose of mine is primarily to support stock, although I will be publishing various modules to add mod functionality to stock parts such as adding mechjeb to stock pods, scansat for stock scanners, life support for stock tanks, remotetech rebalance for stock parts only. Basically a set of module manager files that allow mods to function if desired while still having a completely stock tree. This is something I'va wanted for a long time as I enjoy shzring and downloading stock craft.

regarding fuel tanks, I will be keeping them fairly cheap at the later stages so there is little sacrifice needed to buy them. I have also made it necessary to buy a certain number of one size before you can access the next or progress to fuel filled hulls.

[Sky_walker]

1. Rework the tree to give players an open experience. Rough outline below. Let players follow their own progression.
   
2. No need for a starting node. Give players some science to spend initially. more
   
3. Branches divided by technology.
   
4. Try to put a single part per node, or only strongly related parts. Keep it intuitive.
   
5. Have successive nodes with larger/more advanced parts. more
   
6. Allow modders to dynamically append nodes. more
   
7. Create new part variable for determining part type/function. Useful for filtering. more
   
8. Little or no node interdependency.
   
9. Make really low-tech parts like ladders or thermometers available early. more
   
10. Create an in-game tech tree editor
    
11. The ability to define tier placement per node. more
    
12. Customisable tech nodes / Custom images.
    
13. Customisable tech branch backgrounds / Customisable branch labels (see image below).
    

1. NO. It can't be too open either. Tech tree is very helpful to guide a new players through the game. It should be build in mind of being helpful for them, especially at a very beginning. The one-part-per-node is a complete opposite of that idea. Sometimes you might want one-part nodes, eg. with LV-N, but it cannot be a fundamental principle of the tree.
   
2. NO. I totally disagree with that. Starting node is an essential part of the game experience. Eg. with the tech tree by Sherkaner you'd be able to "loose" the game before even building a single rocket by just making bad decisions at a beginning - it's unacceptable. Not to mention that building a tech tree with no starting node would be far more problematic than otherwise.
   
3. Yep
   
4. NO (as explained in point 1)
   
5. Yep
   
6. Yep
   
7. Yep
   
8. NO. Not only interdependency makes sense in many cases but also adds to the requirement of otherwise very powerful and too easily accessible parts (eg. LV-N, though that one IMHO should be an end-game tech so more people would know how to make an interplanetary missions without LV-N addiction, cause it's perfectly doable but noone is bothered having such a capable engine in a mid-late part of the tree)
   
9. Yes, but only if it doesn't mean spamming player with dozens of useless parts very early on. Things like large grinders should be somewhere in a middle / mid-late part of a tree when player can actually build large rockets, has a capacity to make a good use of them.
   
10. Perhaps... but IMHO it's a waste of time. I seen an arguments raised and I think it's not worth the time spent.
    
11. Yep
    
12. Yep
    
13. Yep
    

In general I think that everyone agree that we need a more intuitive tech tree. But the very first post in this thread is spoiled by some very controversial ideas that really, shouldn't be there in a first place.

Tech tree schematic by Sherkaner

This tree is designed around existing parts using their stats as a guide for their place in the tree. It's not an endorsement of the parts themselves, which do need to be overhauled as well.

Development thread for the tree mod based on the idea below.

http://i.imgur.com/5PKkHI3.png

Jesus, that tech tree makes as little sense as a current one.

To give you a few examples:

• Monorprop tanks seems to be thrown all over the place with no logic.
  
• Engines are always in a separate nodes than fuel tanks - cause making life more difficult is a primary objective (as far as with rocket engines you can sort of use existing tanks, though it's a very, very bad idea with 1.0.x aerodynamics - with ion engine and xenon tank it's a chore and very unfriendly for the first playthrough).
  
• RCS block available not only before linear RCS, but also requiring you to upgrade to the level 2 to get it. lol
  
• FL-T100 -> Oscar B while at the same time FL-R10 -> FL-R25.
  
• 24-77 as a starting engine? Really? With LV-T45 being 2 techs away?
  
• Flea
  
• All of girders and beams in a tier 1
  
• More winglets and fins than ideas what to do with them. Also: starting with canards and no elevons.
  
• External command seat as a starting tech
  
• ...I can't be bothered to keep on going. It's a mistake on top of the mistake. Why do you even promote that?
  

Edited May 28, 2015 by Sky_walker

[sherkaner]

Jesus, that tech tree makes as little sense as a current one.

To give you a few examples:

I didn't claim it was perfectly balanced, and I think you're kind of missing the point in me creating it. A lot of those objections specifically have to do with the caveat that you quoted from me yourself: "It's not an endorsement of the parts themselves, which do need to be overhauled as well." What I mean by that is that many of my orderings have to do with some concept of what the parts should be, not what they are, and would require huge retooling of parts specs and contract flow to make sense. So yes, my tree has enormous balance issues; it was meant as a proof-of-concept to spur discussion and hopefully to show that you could create a mostly per-part tree and it wouldn't be a nightmare.

Unfortunately some still consider it a nightmare, but I am glad it's spurring some discussion finally. Generally the overall objections seem to be around the idea that a player could end up with "bad" combinations of parts. There is some truth to this -- I just see it as a feature, rather than a bug. I think with some good tutorials, and maybe some trial and error (I don't see why a player's first career has to be guaranteed to be successful), a player should feel empowed by the fact that they selected an R&D path of their own that suited their program and wasn't pre-ordained to guarantee a "good" flow.

That said, I do think that if I were to do it over again, I would probably condense some of the nodes a bit. There really isn't a good sensible reason to not group some tanks with some engines for example. I just wanted to start off by "going all the way" on the one-part-per-node path to see how it would go.

I think probably the most important thing is keeping totally unrelated parts out of combined nodes. The stock tree is awful about this, and is the source of the biggest headaches I believe. I would like some others to have a look at the SETI tech tree created by Yemo (who posted here recently, but I think was unfortunately run off by the defensive response to his post). It doesn't my any means follow the single-part-node philosophy, but I do think he's managed to pull off a massively improved tree that maintains a more guided flow, but puts things in a much more sensible order. The big thing is that he actually keeps major technologies in a "line" so that the player can guide their R&D path, which is very much along the lines of what we have discussed here. And it's very friendly to mods (which admittedly is a big issue with my tree -- I was very much thinking stock game). Plus, critically, he's actually rebalancing parts to make them work well with the tree and a retooled contract flow, which is absolutely necessary to have a solution that truly works.

I'll have to check out OpenTree as well sometime.

Edited May 28, 2015 by sherkaner

[Mulbin]

My tech tree is now published and can be found here if anyone wants to take a look - http://forum.kerbalspaceprogram.com/threads/123417-1-0-2-SuperStock-Ground-UP-Rebuild-of-Stock-Tech-Tree

[sherkaner]

My tech tree is now published and can be found here if anyone wants to take a look - http://forum.kerbalspaceprogram.com/threads/123417-1-0-2-SuperStock-Ground-UP-Rebuild-of-Stock-Tech-Tree

Just a surface-level look so far, but looks good! I thought about a tree that went both to the left and right to declutter a bit, and seeing your tree, I think it's probably the right choice.

[Mulbin]

Just a surface-level look so far, but looks good! I thought about a tree that went both to the left and right to declutter a bit, and seeing your tree, I think it's probably the right choice.

It wasn't so much about de-cluttering as the limitation of ksp! You only have a limited height to work within and my tree still nearly reaches the edge. Solution worked well though I think.

[sherkaner]

It wasn't so much about de-cluttering as the limitation of ksp! You only have a limited height to work within and my tree still nearly reaches the edge. Solution worked well though I think.

Yeah, I just barely managed to get my per-part tree within the height limitation and that required a ton of tweaking (and even script-writing for bulk movement of nodes without hand-editing) to get it right.

[CaptainKipard]

NO. It can't be too open either. Tech tree is very helpful to guide a new players through the game. It should be build in mind of being helpful for them, especially at a very beginning.

And it is. Happy days. It's grouped by technology, and the parts get more advanced as you go along. What's difficult about that? Nothing.

NO. I totally disagree with that. Starting node is an essential part of the game experience.

Because...?

Eg. with the tech tree by Sherkaner you'd be able to "loose" the game before even building a single rocket by just making bad decisions at a beginning - it's unacceptable.

You say "e.g." as if there's an obvious connection between this and your previous statement. You can lose right away? So what? You can lose any game as soon as you start. This is a management game, ostensibly. You don't need your hand held as much as you want. It'll be fine. Trust in yourself. Don't make any of those "bad decisions".

You know, like buying a KSC upgrade right when you start. That would be stupid, and you can already do that, but no one complains. And the reason is that it's not a problem anyone faces, because people aren't that stupid, or they simply quietly learn from their mistake and start again. You wont lose much progress in the first 10 minutes. This is a non issue.

Not to mention that building a tech tree with no starting node would be far more problematic than otherwise.

Again. No reason is presented.

NO (as explained in point 1)

TBH I don't think it's realistic myself but still the nodes shouldn't be nearly as crowded as they are now. As arbitrary as science is it still seems a huge waste of resources to research a whole node full of parts just for that one thing you need. When I play, half the parts aren't even bought; they just sit there.

I've updated that point a little.

NO. Not only interdependency makes sense in many cases but also adds to the requirement of otherwise very powerful and too easily accessible parts (eg. LV-N, though that one IMHO should be an end-game tech so more people would know how to make an interplanetary missions without LV-N addiction, cause it's perfectly doable but noone is bothered having such a capable engine in a mid-late part of the tree)

Not to worry. "Little to no" means "less than it is now". The current tree is not even a tree. It's a web. And it's really silly.

Yes, but only if it doesn't mean spamming player with dozens of useless parts very early on. Things like large grinders should be somewhere in a middle / mid-late part of a tree when player can actually build large rockets, has a capacity to make a good use of them.

Ask yourself how many games allow your to use all of your old upgrades/weapons/spells/technologies. I can't think of a single one where it's a problem. Parts become obsolete. Just deal with it. Without a complete overhaul of the parts system (which has been suggested to no avail) this point is completely moot.

Perhaps... but IMHO it's a waste of time. I seen an arguments raised and I think it's not worth the time spent.

Squad already have one internally, so obviously it is.

In general I think that everyone agree that we need a more intuitive tech tree. But the very first post in this thread is spoiled by some very controversial ideas that really, shouldn't be there in a first place.

I don't think "controversial" means what you think it means. This thread wasn't originally mine, but I read every post in it since it started, and the response to Sherkaner's tree and all the ideas in that list have been overwhelmingly positive. They're not all mine by the way. They are the result of consensus. I update the post when something new emerges.

There've been some naysayers but no one really bothers explaining their ideas in an objective way. Opinions are fine, but back them up with something substantive.

Yemo (who posted here recently, but I think was unfortunately run off by the defensive response to his post).

Hey... now... I asked them to backup their points instead of machine-gunning opinions. That's a good thing.

Edited May 29, 2015 by Cpt. Kipard

[sherkaner]

Hey... now... I asked them to backup their points instead of machine-gunning opinions. That's a good thing.

I don't think he knows the warm heart behind the gruff Kip exterior like we do...

[CaptainKipard]

I don't think he knows the warm heart behind the gruff Kip exterior like we do...

I've updated the OP with a little note encouraging readers to engage in a more productive way.

I've also added notes to the points that have been at least partially done.

[CaptainKipard]

Well I've found this discussion productive and useful, whether it adheres to one person's views on the "laws of discourse" or not. Personally I'd much rather spend my time working on a mod that improves the tech tree, than writing a short essay justifying the reasons and philosophies behind it.

OK? A few of those points are directly or indirectly related to improving the modability of the tech tree. You really need to actually write the "essay" to get that ball rolling though. The points have links next to them. Do I have to start making the OP comically obnoxious? I don't know what you hope to achieve with mediocre tools at your disposal. Full mod support for example is technically possible, but not with the current state of the game. The tree system needs to be a whole lot more dynamic. All of this info is already in the thread. I don't understand why I'm forced to repeat it.

There is nothing wrong with a short response if it can be easily understood.

Nice strawman. Please don't do that. That's really frustrating. If it's unintentional then go back and make sure you've understood exactly what I tried to get across.

I can't say the same about entering a mod release thread and criticising the mere existence of a mod and its author's methods while contributing nothing useful to the discussion of the mod itself.

Oh I got confused about who's making this. It's just you Oneiros? Somehow I got the impression it was Fraz86 because he decided to speak for you for whatever reason. Some of what I said is not even pertinent now if he's just speaking for the sake of it, and not actually tied to the mod in any way.

I think that better fits the description of "machine-gunning opinions" than what has been put forward here.

OK, wow.... A machine gun has a sustained high rate of fire. It's not a single argument. I made mine. I have no idea what you're even talking about here. Can we please just move on already for god's sake.

If an open discussion is what you want, you need a round table and not a podium.

Again, nice strawman. You have got to stop.

[passinglurker]

Just how important is openness to all this anyway? I mean if the tech tree went the other way and was just a single linear path with all tech organized into a single node per tier how would the gameplay suffer? you'd never get a dead end buying the wrong thing, balancing and rebalancing would be trivial, and adding crazy engine/solar panel mod combo parts would be very simple.

[Kegereneku]

I think 'openness', intuitiveness and game balance are the main axis, with the ideal being a balance of each.

- Absolute 'openness' is basically having each part buy-able in no order

- total intuitiveness would be better but can allow combination which, although intuitive(~realistic), isn't balanced to the game.

- And game balance can make lead to an obtuse result by making bad choice impossible. I would say it is were is the current tech-tree.

[tater]

Many people have intuitions that are not correct, fwiw. Knowing something about real life informs my intuition, and as a result when I see parts appearing late in the tree that I know are coincident with far earlier parts, I might understand there is a "game" reason for the location, but it's not at all intuitive. I said it way up the thread, but almost everything in this tree was in play in RL before 1965 (effectively coincident, not a real progression).

I suppose this puts me in the camp of not having a starting node per se, and having a massively parallel tree structure, letting the player decide. I would add that if this is not about replacing the stock tree, anyone modding their game likely doesn't need total noob spoon-feeding.

Edited May 29, 2015 by tater

[Vanamonde]

Please keep the discussion about the topic rather than getting into fights with each other.

[passinglurker]

Many people have intuitions that are not correct, fwiw. Knowing something about real life informs my intuition, and as a result when I see parts appearing late in the tree that I know are coincident with far earlier parts, I might understand there is a "game" reason for the location, but it's not at all intuitive. I said it way up the thread, but almost everything in this tree was in play in RL before 1965 (effectively coincident, not a real progression).

I suppose this puts me in the camp of not having a starting node per se, and having a massively parallel tree structure, letting the player decide. I would add that if this is not about replacing the stock tree, anyone modding their game likely doesn't need total noob spoon-feeding.

so the challenge then is to find a way to organize the parts that is intuitive and doesn't disrupt game ballance? what about organizing the nodes based on contracts that the contents therein are meant to fulfil? so for example if you want to to fulfill space station contracts the space station node and its path of dependencies that you follow to reach it would unlock parts like rcs, docking, hitch hiker cans, solars, etc...

One of the biggest frustrations with that current tree is that it at times doesn't equip you to handle the contracts you unlock by progressing through the tree. Organizing the nodes and parts by intended contracts instead of schools of technology could solve that while intuitively informing you of what you need for a type of mission and intuitively cueing you in on what contracts you are about to unlock by researching a node.