Fix the bugs in the game!

"

Jixa87 wrote:

"

bennisen wrote:

I meant fixing bugs, not implement new features or change major things.

They aren't mutually exclusive. New features are code, bugs are code that isn't working because of a new feature or a strange interaction between two features.
What if you need to change major things to fix those bugs?
The bugs occur when a specific action, triggers this function, which relies on a function that's likely meant to handle multiple different functions. and in most scenarios it returns the correct value, except if there is packet loss
So then the bug occurs because of a networking issue
There are bugs that occur because of a specific interaction between graphics drivers intefacing with different versions of graphical languages with an incredibly diverse range of graphics card hardware.

It's extremely complex, which goes back to your point

"

bennisen wrote:

Many dont understand how difficult game development is

For some bugs the main problem is simply the lack of understanding what causes the bug. For these bugs there is no organizational obstacle to an outsider helping with bug finding/tracking other than trusting them with (parts of) the source code. They wouldnt have to "fix" any bug in the sense of writing/contributing code. Incidentally in one of your above examples it was likely such a bug/stage: If you need to reproduce the problem then you are automatically at a stage where you dont have this insight yet.

(Not that I believe GGG would be accepting any such a thing. IME thats incompatible with how larger contemporary companies operate.)


Edit: For additional explanation see this from https://www.jokebuddha.com/joke/Retired_Engineer :

"

Jixa87 wrote:

"

bennisen wrote:

2. You could learn programming, master it (what could take a few decades) and help them for free.

I agree with points made, but just want to highlight how impossibly difficult this statement is.
GGG would have an internal process, developers work on sections of the code, then continually merge into the main block of code. If there are conflicts, they deal with them internally. All updates need to be tested, to an extent given the timeframe.
It's extremely difficult to have some third party agent "helping" write code. It then just adds another layer of complexity when code merges happen and additional testing.
Then you have the accountability, if you came in and wrote some code and it doesn't work and you're not 100% committed by funding and available to fix every block of code you write. Then it just negatively impacts the team cause someone else who was working on something else now has to work on this or decide to discard it.

It's a horrendous scenario, best explained by too many chefs can spoil a broth.
The other common phrase is "If it takes 1 developer 10 days to write this code, surely we could just get 10 developers to write it in a day"
Doesn't work, never seen it work, not sure in what scenario this could work.

Literally hire more bugfix devs. Spend the time and money. It isn't hard. Fix the engine if you have to. Literally just computer science.

They are operating with sub 50 people by some accounts. Increase that number and be amazed at how much can get done.

Like some problems are just throwing more talent at it. If people need more education, pay for it.

"

roundishcap wrote:

Literally hire more bugfix devs. Spend the time and money. It isn't hard.

Now you're talking about recruitment and stating it's not hard? Have you tried to recruit software developers in the past 6 months? it's an absolute nightmare. People come and people go, the new people are never immediately as good as the people who left. Trying to find the right fit is incredibly hard. I don't know how GGG recruit, it's not often you see their careers page light up.
It's not like you can walk out on the street and pull in some niche specific coders that can start being productive straight away, takes months.

Point is, fixing bugs is hard, getting developer's is hard.
I can only assume the company is doing it's best to fix bugs, hire new staff, deploy updates, maintain poe1, advertise and develop poe2
What is easy, is to sit on the outside and point at the company and demand they do better on the assumption that internally they're not trying to do that already.

There are no "bugfix devs" lmao.

"

ArtCrusade wrote:

There are no "bugfix devs" lmao.

I dunno, this reddit thread seems to believe there are.... whilst not common or an official title it can be part of their job description... the part that the employer wants them to focus on?
https://www.reddit.com/r/ExperiencedDevs/comments/10qx21r/is_there_such_a_career_path_as_bug_fixing_guy/

Im not sure if the 20000+ fire resistance chests from standard got mentioned in here already.

But i might do it as well to give a shout out to the dev fixing bugs: its still in the game and there are mirror services provided around it here is a trade link:

https://www.pathofexile.com/trade/search/Standard/pZrnMjeC0


Together with the fire mastety passive 1 life regen per overcapped fire resistance it solves all problems with degens - forever. The increased damage per overcapped fire resistance with Replica Dream Feather

"1% increased Attack Damage per 450 Armour" breaks this bug item quite hard.

In standard the abadoned child there are whole builds based around a bug item
Its like



Replace the idea of software projects with builds based around a bug item and you get the idea.

the problem isnt fixing known bugs 99.9% of the time

its finding the bug and reliably reproduce them

if you can reproduce them you can start narrowing down what exactly caused it,this can be a lot of work with how many things happen in poe at once and how fast they happen

"

ciel289 wrote:

the problem isnt fixing known bugs 99.9% of the time

its finding the bug and reliably reproduce them

if you can reproduce them you can start narrowing down what exactly caused it,this can be a lot of work with how many things happen in poe at once and how fast they happen

In regards to software some bugs are simply not reproducible.

Some bugs are simply not reproducible but cause huge issues see a quote from Chris Wilson on reddit regarding bugged item mods like 20000 fire resistance chest:



"It appears that once every billion item saves or so, stats are being corrupted in some way. This has been going on for a year or so, and we haven't yet worked out exactly where it's occurring. We have been adding stuff to detect the items so we can pinpoint the cause, so we will find it.

In the meantime, we fix them as they crop up. It's a big priority to us, but due to the infrequency of when it occurs (which is biased towards when we have the most players online, like now), it's a very slow process to track it down.

As an example of what this may be, we have been trying to track down another rare problem and eventually determined that memory copies across the 2gb memory boundary can become corrupted under certain circumstances due to a bug in software that isn't ours. While we are pretty sure that isn't this the same problem (as none of the server processes in this chain use more than a gig or so of memory), we're certainly interested to see if this item corruption becomes fixed once that issue is fixed.

Edit: it's probably tens or hundreds of billions of saves required!"

Source: https://www.reddit.com/r/pathofexile/comments/84ekvo/is_ggg_going_to_comment_on_these_bugged_items/

Turned out it was a C Library bug that caused some mods to roll like this 1 in a tens or hundreds of billions of time. The C Library bug got fixed by now but the items remain.


The post explicitly is saying they are hunting for those bugged items with a high priority. I assume from the link provided to official trade side it should be easy to track the trade lanes and find the account with the original bugged item and fix all instances of it.

Edit: Mobile Phone Typos

"

zzang wrote:

Im not sure if the 20000+ fire resistance chests from standard got mentioned in here already.

But i might do it as well to give a shout out to the dev fixing bugs: its still in the game and there are mirror services provided around it here is a trade link:

https://www.pathofexile.com/trade/search/Standard/pZrnMjeC0


Together with the fire mastety passive 1 life regen per overcapped fire resistance it solves all problems with degens - forever. The increased damage per overcapped fire resistance with Replica Dream Feather

"1% increased Attack Damage per 450 Armour" breaks this bug item quite hard.

In standard the abadoned child there are whole builds based around a bug item
Its like



Replace the idea of software projects with builds based around a bug item and you get the idea.

+1

"

ciel289 wrote:

the problem isnt fixing known bugs 99.9% of the time

its finding the bug and reliably reproduce them

if you can reproduce them you can start narrowing down what exactly caused it,this can be a lot of work with how many things happen in poe at once and how fast they happen

This is why it is necessary to rewrite your code from the ground up. It is simply too painful to blindly try and recreate something.
QA should be part of the process and switch to/ work with UX whenever their workload is light.

Trying to rebuild something from the ground-up is also the only way to know your code inside and out, hence making bug finding multiplicatively easier.

Rewriting code-base ground-up every 5-10 years just makes sense. The more you rely on imported stuff, the more you can stop relying on imported stuff during this process, and the team will be more proficient every 5-10 years hence a better code-base with better documentation.

Get the dual QA/ UX guys to help with documentation so it's a better experience to read through.