Matrix Games Forums

wow..... thanks for all the advice on this, appreciate it!

Been awhile but was curious if I could generate pre-filled hex layers based on GIS intersections and filters. Turns out you can!

QGIS 3.24.1

Vector maps from here https://www.naturalearthdata.com/downlo ... l-vectors/

First image is standard vector display (if the attachments order properly!)

Second image is 20km hex grid overlay after joining to the lakes layer and filtering out the smallest lakes.

This may sound a bit negative but I think that we have reached the stage where our ability to source map data is really good. What we need now is the ability to more accurately depict that data on the screen. Being able to show coastlines, lakes, ridge-lines, fortifications, and a host of other things on hex-sides would be a major leap forward. Showing more types of rail-lines would be nice too. Terrain.... There are a lot of extra types of terrain that could be added and as the scale becomes smaller that number expands exponentially.
I know that at the moment the programming priority is on leaders and lines of command and I fully agree with that.
But a bit of extra topographical chrome would be nice. My son who is a professional programmer tells me that it would probably take about 30,000 lines or six weeks for an average professional programmer.
At least think about it anyway.

I think everyone agrees those additional hex properties would be wonderful to have. We could also have a Python tool that takes known source GIS data, a map projection, map boundaries, and hex grid size and have it auto generate a map that imports into TOAW directly. Incredible time savings by tweaking, as opposed to building from scratch.

"My son who is a professional programmer tells me that it would probably take about 30,000 lines or six weeks for an average professional programmer."

Hate to say it but this is a nonsensical statement. It has nothing to do with lines of code, or time estimates from a programmer who has never seen the codebase. It has everything to do with integrating new features, into a 30 year old system that has seen multiple upgrade cycles, in a way that doesn't break everything, including all existing scenario balance. This *could* be six weeks, full time. It could also be year full time, or in TOAW world, 9 years part time! Anyone outside the codebase has no way of knowing.

30,000 in six weeks is 1,000 lines per day!

https://skeptics.stackexchange.com/ques ... rogramming

This says the average is 10 lines per day.

Personally, I've done about 40,000 lines in six years. About 25 lines per day.

I was a bit worried I might have got that wrong so I went back to the original question I asked my son. His actual answer was on a large project about 4,000 quality lines a week. So a bit more than six weeks. Sorry about that.
As for the 10 lines a day back in 1975. I started programming in 1980 and writing a functioning program was like trying to land a man on the moon. So ten lines.. I can agree with that.
I dropped out of programming for a couple of decades and only got back into it a couple of years ago. My son calls my programming "caveman stuff". But between only writing part days and dealing with cancer I have written 60,106 lines of code this year completing a functioning map making program. That doesn't include code that was deleted during debugging.
The difference between that and what Curtis is doing is that mine is an original program and he is updating an existing one. I just have to debug what I have written. He has to try to identify all of unintentional side effects on the existing program. Let alone play testing.
So I was a bit unfair on my programming comments. I apologize.

biddrafter2 wrote: Fri Apr 22, 2022 2:07 pm in a way that doesn't break everything, including all existing scenario balance.

This has already been done and is being done at an even larger scale at this very moment. It is an old argument that's been regulated to the Museum of Prehistoric Rhetoric.

Lobster wrote: Wed Apr 27, 2022 1:35 pm This has already been done and is being done at an even larger scale at this very moment. It is an old argument that's been regulated to the Museum of Prehistoric Rhetoric.

When I was working on the old Diamond Mind baseball game I created code that would centralize multiple season sims, or idealized controlled tests, into simple statistical tests to check for significant changes based on system code changes. A very complex integration test for a system built without unit tests.

Would be excellent if TOAW had something similar. If the computer opponent could play itself well and reliably 10k times via the engine code (not UI) and store the results, then sensitivity in each version to scenario balance changes could be compared. Idealized, simplified scenarios could be created to test specific features. Perhaps this already exists?