Industry locations, spacing, topography

[sdog]

In my last large game, i forgot to set the industry spacing to a reasonable value -- that is between 200 and 400 for a large map -- and ended up with industry unusuably close to each other. It was not so severe, as it is mostly a pax game. However, since this happened to me more than once i think i already suggested a while ago to set industry spacing as a fraction of the map size, instead of an absolute value.

In experimental we have an excellent algorithm creating and sizing towns in a very realistically looking way, based on the topology. Viz. settlements are attracted by water and preferably build clusters.

I wonder if it is feasible to use this method also for industry placement. An example: Coal power plants for example should be attracted by water and settlements but repulsed by each other. Collieries on the other hand should be attracted by each other, by high ground and be repulsed by water.

At the moment coal mines get rather evenly spread over the map, power plants too and are often interspersed. Low industry spacing and interconnection settings makes coal transport a very local bussiness, high industry spacing values causes a rather absurd effect: Locations A and B be distant from each other. The power plant near A spawns a colliery near B, while the power station at B spawns one near A. The player can shuttle a train between A and B now always full with coal.

PS.: The city generation code was rejected for standard as having to much computational load, in my experience however, intercity road connection takes up the most time in map generation. Has there been some testing? If no, we shall do it and suggest inkelyads code again!

[jamespetts]

SDog,

ahh, those are interesting ideas. They seem good in principle, although how exactly one defines different characteristics for each industry seems to need further consideration, as it might otherwise become either arbitrary or horrendously complicated.

A note of caution also: this may well have to join a rather long queue of potentially desirable but non-urgent feature improvements, as I have already planned out in approximate outline the sorts of things going into the next two major versions, as well as being nearly ready to release the next minor/bugfix version.

Of course, if somebody else were to code it, it could easily make the next major version

As to the city generation, you are correct that, in my experience at least, it does not seem to take an excessive level of computational load. I don't know whether any formal testing has been done, but, if it is thought desirable in Standard, it might be worthwhile.