How to improve pathfinding

please, hire more people and follow this example. you were literately the first thing i thought of when i read this.

http://blog.dota2.com/2015/03/various-pathfinding-fixes/

It literally took them months to find and fix that bug. The entire purpose of that blog was to show idiots that bugs are fucking difficult to find and fix unless they're extremely fucking obvious, in which case the bug probably would've been fixed before it went into production. There's no "example" to follow here, it's just one programmer giving a little insight into their debugging process and showing people why you can't magically fix shit overnight, which nobody who isn't in software development seems to be able to understand.

If you can't accurately and consistently reproduce a bug or find the single line of code in thousands where the bug occurs, you can't fix it. I work on systems with bugs that have been around for years, because they happen once every few months in multithreaded projects with hundreds of source files and thousands of lines of code and can't be reproduced. Unless you can reproduce the issue or get it to happen in a debug environment (which Valve essentially forced here by dumping so much debug info into the replays that they just needed a timestamp to get what was probably the equivalent of a complete memory dump at that exact time in the game), there's probably nothing you can practically do to solve it.

Also, adding programmers generally leads to a reduction in software quality, not an increase. Relic hiring more people would likely cause more problems.

Relic needs more testers, not more programmers.

There is a shitload of common bugs in the game that are still around for ~reasons~ considering it's fairly easy to reproduce them.

And pathfinding is shit as a combination of bunching mechanics and EVERYTHING moving and sticking to cover, including tanks.

it did take them months. but they did it. relic doesn't have the manpower to do stuff like this even with the trello board pointing out much simpler things. even worse, half the time relic fixes one thing they break something else, often with no relation between the two.

Why do you think manpower would suddenly fix bugs? Valve didn't fix this bug because of manpower. They fixed this bug because they added a fuckload of debug material to their replay files and were able to use that material to find the source of the issue once they got their hands on a replay that reproduced the bug. Neither of those are very manpower-intensive operations; the entire process was probably done by a single programmer.

http://en.wikipedia.org/wiki/Brooks%27s_law

Adding manpower, especially to maintain code, more often than not results in more problems and does very little to increase the rate at which current problems are solved. You need to train your programmers, let them become familiar with the codebase, and make sure they don't fuck something up because they don't understand why something was done a certain way. It takes an incredible amount of time and effort to get a programmer up to speed with people who have been working on a product since pre-release, and more often than not they're going to introduce new issues while they're ramping up because it's impossible to understand every single nuance of a complex codebase until you've worked with the codebase for an extended period of time.

http://www.giantbomb.com/dota-2/3030-32887/credits/

Those are the credits for Dota 2.

http://www.mobygames.com/game/windows/company-of-heroes-2/credits

Those are the credits for CoH2. CoH2 has a whole lot more manpower dedicated to it than Dota 2 does. Manpower means jack shit.

Oh come on! not only does Company of Heroes have less money to thow around then DOTA. Pathfinding is also much more complicated with multiple entities within units, but also vehicles which act completely different pathfinding as well. Plus all that cover and trees on the field that makes things difficult as well.

The two are completely different games and there's more to it then just throwing more people at it.

relic doesn't have the manpower right now to keep up on anything, much less set people aside for obscure bugs or serious overhauls of things like pathing. i'm not sure how big their team is but they're trying to hire like 6 people right now (they need more than that) and almost half are high level positions...

Are you purposefully ignorant? Because it's clear you haven't read a word I've said. And since when are Art Directors, Senior AI Artists, Data Analysts, Executive Producers, Financial Controllers, HR Coordinators, Associate Brand Managers, Cinematics Artists, Communications Coordinators, Community Managers, AI Programmers, or Tools Programmers responsible for gameplay engine bugs? There is a single Relic job posting right now related to general programming. Please get your head out of your ass.

Dota 2 has a team of 28, including designers, programmers, and artists; they have access to far less raw manpower than Relic does. More people are working on CoH2 than Dota 2 right now. Tell me again how important manpower is to bug fixing?

Don't get mad Inverse xD You are indeed right tho, Seb would confirm that 10000000000000000000000 times about the COH2.ORG code

pathfinding in general is pretty tricky to get right. There are quite a few games where it's an issue where you have unit collision enabled.

with regards to CoH2 the maps could be designed better to take the pathfinding issues in consideration so you minimize the weird pathing vehicles take when they get near an object.


Credit to Ginnungagap for the picture

NM

And since when are Art Directors, Senior AI Artists, Data Analysts, Executive Producers, Financial Controllers, HR Coordinators, Associate Brand Managers, Cinematics Artists, Communications Coordinators, Community Managers, AI Programmers, or Tools Programmers responsible for gameplay engine bugs?

i would consider all those to be at least somewhat related to solving gameplay engine bugs. there's also the issue of whatever those people are supposed to be doing either not being done or cutting into someone else's time.

dota 2 is also (as far as i can tell, having played ~~60 hours of it) in much better shape, balance and bug wise, than CoH2 and (again, as far as i can tell) has much better analytic tools available to the people working on it. theoretically, once you reach a certain point with a game you only need one or two part time balance devs and occasional time from a programmer to keep a game balanced. RET had only bC on it for several months and it was *mostly* fine. as time goes on the amount of effort required for fixes should decrease as the slop is taken out of the system...

the coding problems, pathfinding in particular, are fairly complicated but the balance issues mostly come down to making changes to a spreadsheet and then testing said changes with a heavy does of judgement involved. more manpower means that changes can happen faster and they're easier to test test, provided everyone involved has decent judgement and is properly supervised.

if anything more people means more bureaucracy and things are much slower to fix.

for a relatively small company Relic have enough bureaucracy and red tape that would give even a German nightmares.

i would consider all those to be at least somewhat related to solving gameplay engine bugs. there's also the issue of whatever those people are supposed to be doing either not being done or cutting into someone else's time.

dota 2 is also (as far as i can tell, having played ~~60 hours of it) in much better shape, balance and bug wise, than CoH2 and (again, as far as i can tell) has much better analytic tools available to the people working on it. theoretically, once you reach a certain point with a game you only need one or two part time balance devs and occasional time from a programmer to keep a game balanced. RET had only bC on it for several months and it was *mostly* fine. as time goes on the amount of effort required for fixes should decrease as the slop is taken out of the system...

the coding problems, pathfinding in particular, are fairly complicated but the balance issues mostly come down to making changes to a spreadsheet and then testing said changes with a heavy does of judgement involved. more manpower means that changes can happen faster and they're easier to test test, provided everyone involved has decent judgement and is properly supervised.

You think HR and community management are related to solving code bugs? Umm...

Dota 2 never had a team in excess of 50 people; even at its largest, Dota 2's team was smaller than CoH2's. Again, you think more manpower is going to help CoH2 because...why, exactly?

Regarding the state of Dota bugs: http://dev.dota2.com/showthread.php?t=127919

You linked to a blog about a bug fix and used it to complain about Relic not fixing bugs. Now you're talking about balance, as if the two are related at all in terms of who works on what. You're fairly clueless aren't you?

You think HR and community management are related to solving code bugs? Umm...

Dota 2 never had a team in excess of 50 people; even at its largest, Dota 2's team was smaller than CoH2's. Again, you think more manpower is going to help CoH2 because...why, exactly?

Regarding the state of Dota bugs: http://dev.dota2.com/showthread.php?t=127919

You linked to a blog about a bug fix and used it to complain about Relic not fixing bugs. Now you're talking about balance, as if the two are related at all in terms of who works on what. You're fairly clueless aren't you?

yes, not directly, but yes. HR is fairly indirect but communities manager is pretty obvious: they compile the issues and take them to the people who deal with them, ideally sorting them by importance as best they can. i don't think that's the choke point here but it does relate.

because they either don't have enough people to do the job or someone(s) is incompetent. or possibly both.

bugs and balance are not the same but they often have the same outcome to the player. bug also have an affect on balance. the people dealing with bugs don't necessarily care a whole lot about

the balance but the balance guys damn well better care about the bugs.
my point about the blog was that a very obscure bug finally got fixed; CoH2 has a shit ton of bugs that are not obscure and have not been fixed, some of which were introduced seemingly at random.

In your original post, you told Relic to "follow [Valve's] example" and "hire more people" so they can, as implied by the blog post you linked to, fix bugs like Valve fixes bugs. Except Dota 2 has less people working on it than CoH2 does, so your entire argument that more manpower = better bugfixing has no factual basis in reality, since the accomplishment you're using as your main piece of evidence was achieved with less manpower, not more. When this was pointed out to you in clear terms, you continued to defend your baseless argument despite having zero supporting evidence beyond the blog post you initially linked, which is actually evidence against your position as I pointed out above.

http://en.wikipedia.org/wiki/Brooks%27s_law

Please inform yourself before you pretend to have a clue about how software development works.

Why am I thinking this is the sort of conversation Inverse has wanted to have with at least one boss in his time but can't?

Provoke the frustrated coder at your peril, you may be Pointy Haired but you are not his Boss

Haha nah, my bosses are great, and so is the team I work with. It's just funny because the evidence given in the OP actually directly contradicts the argument it was meant to support. He's saying "Relic needs a bigger team, and to prove my point I'm going to show you work done by a smaller team than Relic currently has."

pathfinding in general is pretty tricky to get right. There are quite a few games where it's an issue where you have unit collision enabled.

with regards to CoH2 the maps could be designed better to take the pathfinding issues in consideration so you minimize the weird pathing vehicles take when they get near an object.


Credit to Ginnungagap for the picture

Lol this picture is fantastic!

my argument was actually not an argument at all. it was "look, valve solved an obscure but with dota while relic introduces bugs and doesn't fix existing ones, even when they seem to have easy fixes. also, relic's team is clearly to small for what they're trying to do so hire more people."

at this point i don't care enough about the game to have a serious argument about it; just enough to think the march patch might make the game better.

clearly my apathy meant that i have a poorly worded OP; i am not at all surprised and don't care.

it's also clear that you are irritated about the subject (which i understand) or you would not have responded with 10x the words to my shit post.