Escape small island

[Deadly-Bagel]

Meh, there is a solution that could be implemented. My point is rather than bandaging the problem with an auto reload or landfill or whatever, if you're going to fix it then fix it. But I still don't think it's something to be particularly worried about.

Your statement are paradoxical.

We offer exactly "bandaid" solution that are easy because there is no need for dev to use alot of time to that particular issue. The problem is not the map generator as it's the first phase when starting a new game. The problem occur on second phase when we are placed on the map.

/stuck command is another good alternative.

Yes but your bandaid solutions don't really improve anything. Auto reload of the map? You still have to sit through another world gen except now you're wondering why. Dumping stuff in your inventory? 99.9% of games will never need or use it and most will just find it annoying. /Stuck? Probably the better solution as it's non-intrusive and allows you to keep the current map BUT:

You're saying the problem is that when this happens the game doesn't feel polished, right? That's why you're not happy with instantly restarting it. What I'm saying is that these bandaid solutions don't actually change that, so if you're going to fix it based on the reasoning you've given then you need to fix the problem itself.

[Jan11]

I know it`s not really the topic but It`s a big escape island.^^

I figured out how to get your big escape start island (no connections). (Most of the time)
and not the small 1x1 field

Water Frequency : Low / Size : Very big

Iron Freq : Very high/ Size: Very big / Richness Very good
... Copper, Stone etc etc all the same

and richness : very good

Start a new game and no island /restart / I started yesterday a map

If i need more space or minerals I connect my island with the land. ^^


Map Key

>>>AAAOABYAAAABBQYAAAAEAAAAY29hbAUFBQoAAABjb3BwZXItb3Jl
BQUFCQAAAGNydWRlLW9pbAUFBQoAAABlbmVteS1iYXNlAwUFCAAAAGl
yb24tb3JlBQUFBQAAAHN0b25lBQUFcxc7OYCEHgCAhB4AAwD8j8dY<<
<

[vtx]

Meh, there is a solution that could be implemented. My point is rather than bandaging the problem with an auto reload or landfill or whatever, if you're going to fix it then fix it. But I still don't think it's something to be particularly worried about.

Your statement are paradoxical.

We offer exactly "bandaid" solution that are easy because there is no need for dev to use alot of time to that particular issue. The problem is not the map generator as it's the first phase when starting a new game. The problem occur on second phase when we are placed on the map.

/stuck command is another good alternative.

Yes but your bandaid solutions don't really improve anything. Auto reload of the map? You still have to sit through another world gen except now you're wondering why. Dumping stuff in your inventory? 99.9% of games will never need or use it and most will just find it annoying. /Stuck? Probably the better solution as it's non-intrusive and allows you to keep the current map BUT:

You're saying the problem is that when this happens the game doesn't feel polished, right? That's why you're not happy with instantly restarting it. What I'm saying is that these bandaid solutions don't actually change that, so if you're going to fix it based on the reasoning you've given then you need to fix the problem itself.

Ok I'll play along and I thank you to play the role of opposition .

Identified the problem : With some settings the odds of starting on a tiny island without a way to escape increase.

Phase it occur : Most likely at spawning of player position. ( not the map itself )

Side notes : try sandbox scenario. There buttons you can click to activate all research and a box with some items in it, So all solution can be optional.
Solution must be not intrusive and count the fact that most of the time you didn't not need it.

Solution :

- Do nothing
--- Rely on restart of the map
PRO : Required zero coding from the dev.
CON : The problem still not fix and some map stay unplayable.

- Alter map around players :
--- Remove all water around X tiles of the player :
PRO : Reduce alot the odd of starting on that type island
CON : That mean you'll need to travel further to reach water, even if the "default" map have no island

- Alter player inventory :
--- Add landfill to player inventory
PRO : Player have a way to mannually escape the island. Escape it the way he/she want.
CON : Leave item in inventory that most player did not want and will have to mannually place in a box and shoot it to delete those items

- Alter player position :
--- Add /stuck command or a button "Are you stuck YES/NO" that will move slightly the player untill he was no longer on water.
PRO : Simpliest way to get out without terraforming or altering inventory
CON : Player will have to click NO most of the time, as the odd of this map is low if we use button approach but not occur for the command

- Add box with tool to escape that island :
--- Same as alter inventory but in a box
PRO : same as alter inventory
CON : you'll have to shoot the box to delete those items

- Scan map around players :
--- Search for enclosed small land around player and move it to the nearest land other side of that island.
PRO : did not required any interaction from the player
CON : Required more code. There a chance you end up on another island. So for each time you are teleported there need to validate the new position.

I think the most less intrusive and reliable solution will be scan the map.

Pseudo code : Scan X tiles to the left for water if no water = map valid
else scan X tiles down for water if no water = map valid
...
if you loop back = invalid map

[greep]

I honestly don't see why a map scan would be an issue. I think the factorio devs can figure out how to do a map scan for islands, it'd just be a simple recursive flood fill that backs out if it reaches X tiles or a pathfinding with similar effect. It could also count stone/iron/copper/coal/oil and generate a new map if it finds a moderate sized island with insufficient amount and display the rejection ala dwarf fortress to show why the creation is taking a while.

My guess is they recognize an easy solution and just don't see a point in implementing it in the beta version. What counts as "sufficient ore", for instance, if they're constantly revamping basic features of the game like science and adding new ores? And what if they decide to add swimming? Maintaining a good solution doesn't seem worth the effort for an unfinished product. And people posting silly island maps is like free press.

[vtx]

I honestly don't see why a map scan would be an issue. I think the factorio devs can figure out how to do a map scan for islands, it'd just be a simple recursive flood fill that backs out if it reaches X tiles or a pathfinding with similar effect. It could also count stone/iron/copper/coal/oil and generate a new map if it finds a moderate sized island with insufficient amount and display the rejection ala dwarf fortress to show why the creation is taking a while.

My guess is they recognize an easy solution and just don't see a point in implementing it in the beta version. What counts as "sufficient ore", for instance, if they're constantly revamping basic features of the game like science and adding new ores? And what if they decide to add swimming? Maintaining a good solution doesn't seem worth the effort for an unfinished product. And people posting silly island maps is like free press.

You miss the point of this topic. It's not about generate a new map automaticaly ( everyone can press ESC + Restart ). It's about making that map playable.

Indeed if they include swimming or the upcomming spider is accessible from the start the whole topic is useless.

In case you didn't know we allready see the finish line of this product aka release patch possible candidate 0.16.

[greep]

Well, either why, just replace "generate a new map" with "do stuff to map to make it playable" in my comment. I don't have much of an opinion on that part.

[Koub]

I think only the devs can really tell what are real "PRO" and "CON" in terms of development, any "easy" thing from our point of view can be a nightmare in the dev's point of view.

[vtx]

I think only the devs can really tell what are real "PRO" and "CON" in terms of development, any "easy" thing from our point of view can be a nightmare in the dev's point of view.

Of course it's my point of view for those PRO and CON ( player perspective ). Sorry if I offend the dev with that organisation of idea.

I know very well that what client request as an easy task for them, most of the time it's nightmare to implement and you end up asking yourself "Why didn't he ask that sooner in the project?!?".

[Deadly-Bagel]

- Scan map around players :
--- Search for enclosed small land around player and move it to the nearest land other side of that island.
PRO : did not required any interaction from the player
CON : Required more code. There a chance you end up on another island. So for each time you are teleported there need to validate the new position.

I think the most less intrusive and reliable solution will be scan the map.

Pseudo code : Scan X tiles to the left for water if no water = map valid
else scan X tiles down for water if no water = map valid
...
if you loop back = invalid map

Another solution might be having a selection of "paths" that are natural-looking land paths, whenever the game starts it checks a dozen tiles to the left and right and if both contain water it whacks a randomly selected path down in a direction. False positives don't have any effect other than maybe modifying a nearby water patch. This is a flexible solution, for example you could check water in one direction only and if found lay the path in the other direction. You could also test for water at the end of the path and if found then repeat, useful if islands are likely to be very far from land.

[ssilk]

I remember a comic called "DEN", where an ant-like race tried to fight a water castle. But ants cannot swim. But what they did was jumping into the water until there where enough ants in the water that the remaining ants could climb on. In the end they build a bridge from dead ants and reached the water castle.

So now my thought: The water castle is the isle where you are on. And the ants are the natives, attacking you.

How can you get so many natives attacking you? Let's assume there is a tree or a bush on the island (on every island) which holds fruits and if you cut it off you get a fruit and if you specially modify this fruit it smells in a way, that the ants will like and they jump into the water to get to the island.

Could be interesting to find the right point, where to stop this.

EDIT: I look back to the comic now (1 week later) and indeed it is so, that the ants carry earth to build a dam.

[Deadly-Bagel]

IMO no solution is needed. Sure it would be a nicety but it's such a rare problem that's so easy to fix and most of the players who encounter it find it funny anyway. If a solution IS to be put in place, it needs to require no player action, in fact be completely unnoticeable, and not impact performance of map gen. If it doesn't meet these criteria then we would be no better off than we are now.

This rules out automatically restarting, putting anything in the player's inventory, new recipes and console commands. What we are left with is detecting that the player is on an island (as long as it doesn't affect performance) and teleporting them to the mainland, automatically laying down landfill (can but doesn't have to check for island), or ensuring that no islands can exist in the first place.

[darkfrei]

Is it possible to delete all empty islands between map creation and player spawning?

Empty: or(no iron, no coal, no stone)

[vtx]

I decide to make this mod as I like the challenge of it.

Thanks to Deadly-Bagel it help me alot to find the best approach for this. I allready code the detection of small island only 2 step left. Also there a distinction between small island ( not viable at all ) easier to do and medium/big island that can be viable depending on ressource they have and require a different code.

I have one technical question about multiplayer that will lead the final step.

Did every players spawn at same location?

As this code must only run once and not at every players creation I think the landfill bridge ( automatically ) will be the best solution.

Unfortunately it's not exactly a mod that can be over the current base mod. It have to be inside freeplay scenario ( not totally true, but every scenario that are not freeplay have achievement disable ). Once I finish it I'll publish it in mod section forum to be evaluated by other modders and if good enought maybe it'll be added in base game .

[vtx]

Fun fact I discover while trying to work on that mod.

There allready a system in place that adjust your position if the position (0,0) normal starting place are water. Then it will search the nearest land.

The nearest land was a small island when I discover that as my new position ( 3.19921875,1.19921875 ).