Topic: wishlist: preferrably split selcted ware among warehouses

GunChleoc wrote:

If we were a small company with, let's say, 5 programmers & artists working on this full time, Build 20 would have been out 10 years ago. Our main constraint here is lack of time.

Yes, lack of time- yes. But 10 years ago... It would be difficult (3 years for 20 builds?, or my assumption of time is wrong?)

Also the second problem is number of ideas. Many people have many good (and bad) ideas. Trying new possibilities take lots of time. That can't be done even with 100 people involved.

I can see that on this forum there are 3 people who are talking much (me, kaputtnik and king_of_nowhere). Also two contributors (GunCheloc and Tibor). Of course there is more people involved. :). It is definately too little people to run the project very fast. (more programmers for sure)

einstein13
calculations & maps packages: http://wuatek.no-ip.org/~rak/widelands/
backup website files: http://kartezjusz.ddns.net/upload/widelands/

Top

Quote

king_of_nowhere wrote:

anyway, i only suggest modifying conditions already existing, which should not touch other code. the code should be something like "if f(x)=y, then do this", or "z=z+f(x)", where x, y and z are gaming variables. changing a specific condition should turn it into "if f '(x')=y', then do this", which should require no more than changing f(x) to f '(x') without altering any code anywhere else. generally, all modifications i propose are of the kind that would only require those specific and limited modifications. or so i assume, correct me if i'm wrong.

Well if it was such simple as replace '2' with '15' somewhere in the code it would be done by now. But what is problem is that computers are too stupid. I completely understand what you want, but I can not tell this to computer in few human words. You must be absolutely explicit when wanting something from computer. Usually simple ideas have unspoken prerequisites or constrains that seem logical and obvious to human - not worth to mention - bud coder has to think about them, identify, code them, test, re-think, re-code, test and so on.

Anyway if you or somebody want to spend few hours doing some non-coding work I would suggest participating in making AI difficulty levels - I mean there is new code (not merged now) that brings new flags limiting number of buildings per difficulty level. All that is needed is to edit lua conf files for buildings and test it. And again and again. Till it works reasonably. I have some settings for barbarians done - only for barbarians, so a volunteer can completely prepare these settings for other tribes and review also barbarians... Also, no compilation would be needed for this job.

Second job, quite simple is to review function review_wares_targets() that increases target of wares but in very linear and stupid way:

target_quantity = default_target_quantity * (productionsites_count + ports_count * 5) / 50;

it is quasi rounded to one decimal point, so for 75 buildings all default_target_quantity are multiplied by 1.5, all rounded down. It would be nice to add some logic. Here some very basic coding skill could be needed, the function itself is very primitive.... But it is separate piece of logic so it should be easy to start with

Top

Quote

staustelladam, you'r right. But i can understand king_of_nowhere ... it's sometime frustrating to have good ideas for improving the game, but the proposals do not get in the game in a short time.

This is the common situation for all(!) opensource software. The reason is always the same: Lack of developers. This is sometimes hard if you use an open source os: You find a nerving bug, post it on official bugtrackers but nothing happens.

Posting bugs or make proposals is frustrating if nothing happens in a short time. One is dependent from a few people while doing the best for the project. This is a bad situation for the one who want to make the project better, but the developers couldn't do anything against it. They're doing also the best for the project and are also in a bad situation: While trying to make the project much better by spending much of their spare time for coding, they have to take care about the users and has to be patient with them. Answering questions and give help shrinks time for coding.

There is no solution for this dilemma

Top

Quote

king_of_nowhere wrote:

So, I'd like to have a team of programmers who do it because they love it, but nonetheless are paid to do that as their full time job, but without corporate directives.

That sounds like a Kickstarter indie, or something like Frogatto & Friends

Busy indexing nil values

Top

Quote

Tibor wrote:

One technical problem for suggestion - wares on the way to a warehouse. I mean we know how many wares are stored in warehouse, but we dont know how many of them are on their way there. So in case of traffic jam we can end up sending many wares to the same (let say currently empty) warehouse.

You know two warehouses: W1 with 1 piece of X, W2 with 0 pieces of X. Engine can keep sending wares to W2, and there can be 10 pieces of X on way there...

yes, but from a practicall point of view, that dooesn't matter as long as there is at least one ware in each of the interested warehouses. you still have that ware available where you want. So that's just a good reason to get a good buffer; also an incentive to make good roads, but there are already enough of those.

that's also the reason why the "minimum target" approach is preferrable to the "always split evenly" approach. The first will require lower traveling distances once the minimum target is met.

Top

Quote