102 replies to this topic

[Johnny Z]

GurpGork, on 23 May 2016 - 05:47 PM, said:

Based on the news that has come out over the years, there was an admission more than once, that being able to implement certain new features was impossible, due to original programmers having left MWO. It makes a lot of sense, because as mentioned very little "new" material has been produced (mechanics, changes to how things work, etc). It's all been maps, mechs, and playing with what amounts to spread sheet values.

And the time required for them to make the changes they do to the game ... takes forever. This has got to be some of the slowest development I've seen from an active funded development team.

It's very sad.

Someone already said and it has been said a few times, but the engine, Cryengine, is the original source code. "Technically" everything beyond that is "Modding". That goes for all the other games as well, and everything else, the models, textures, scripts, sound, could "technically" be described as "modding" for this topic.

The code has obviously been added to and changed and customized and is on going.

The original Crytek coder is said to live in a cave, high in the mountains, somewhere in Africa.

Edited by Johnny Z, 23 May 2016 - 06:44 PM.

[N0MAD]

Do you really compare the amount of developing you need to put a rather simple vehicle on a simple map, with developing a game with complex vehicles, NPCs, missions etc? really?

[Nightshade24]

Bishop Steiner, on 23 May 2016 - 04:26 AM, said:

If I recall correctly, they never tried to mech Quad Mechs work.

Why add fresh code, new and double the animation, new mechlab mechanics for such diminishing returns? The simple scarcity of chassis, especially in the time era chosen that would benefit from inclusion is so small, it simply makes no business sense to include. (And unless they all were turreted, the QQ by those trying to use them once introduced would be HUGE.....and that's before they realize they also lose 12 crit slots per mech...)

Russ has stated a number times on townhall and occasionally on twitter on the matter of quad mechs is that they have tried it back in early development but it was plagued with problems coding side that they decided to drop it.


Edit: however with the fact of the crit slots, and awkward aiming and limited chassis of this time period.

This is probably why they are not bothered with the fact that quads won't work, Unlike something more important say max speed cap and MASC.

Which they have made, found problems, and fixed MASC. increased our maximum speed to 170 and another one is apparently some time eventually. and so on...
Quad mechs? no attempts thus far to fix it.

Edited by Nightshade24, 23 May 2016 - 08:06 PM.

[Bishop Steiner]

Nightshade24, on 23 May 2016 - 08:04 PM, said:

Russ has stated a number times on townhall and occasionally on twitter on the matter of quad mechs is that they have tried it back in early development but it was plagued with problems coding side that they decided to drop it.


Edit: however with the fact of the crit slots, and awkward aiming and limited chassis of this time period.

This is probably why they are not bothered with the fact that quads won't work, Unlike something more important say max speed cap and MASC.

Which they have made, found problems, and fixed MASC. increased our maximum speed to 170 and another one is apparently some time eventually. and so on...
Quad mechs? no attempts thus far to fix it.

not denying that (though I honestly do not recall ever hearing that) but it also would have been "solved" if it were financially prudent to, the fact there was a minimal potential return means not worth excessive man power to solve. Still basic economics.

Edited by Bishop Steiner, 23 May 2016 - 08:07 PM.

[KahnWongFuChung]

OP the answer to your question is LOOK BELOW



And now you know why MWO fails.

[Nightshade24]

Bishop Steiner, on 23 May 2016 - 08:07 PM, said:

not denying that (though I honestly do not recall ever hearing that) but it also would have been "solved" if it were financially prudent to, the fact there was a minimal potential return means not worth excessive man power to solve. Still basic economics.

Yea, which is a shame there is such a huge lack of quad mechs in BT and the fact we'll probably never see them in MW: O.

[Troutmonkey]

Darth Futuza, on 23 May 2016 - 08:39 AM, said:

I don't understand all these naysayer comments, half of the people in this thread are acting like they're top industry c++ programmers with the code in hand who have just finished giving a code review of the mwonline source.

I work in C# with Unity with an indie company in Australia. I can't see PGI's source code but I can see the output, and the output is telling me there's lots of fundamental errors with how the code is structured and run.

[Darth Futuza]

Troutmonkey, on 23 May 2016 - 08:52 PM, said:

I work in C# with Unity with an indie company in Australia. I can't see PGI's source code but I can see the output, and the output is telling me there's lots of fundamental errors with how the code is structured and run.

I disagree, it's one of the more stable applications I've worked with. Crashes less often then League of Legends and the game works. There's definitely some areas that I think could use improvement, but it's not fundamental design flaws. eg: The zoom code is shoddy and slow, battle grid and ui for the menus etc is poorly implemented, the core game itself however is fine. Granted I haven't tried attaching a debugger to it, decompiling, and running test cases (and this would violate the terms of service - reverse engineering and all that) - so there might be some awful design decisions, but based on output I think it's done well enough compared to most games. Is it as stable as the linux kernel? Hell no, but for a game it's fine. Their netcode is also performing nice, never heard of or seen cheats that actually work because they don't authenticate through the client and game actions are entirely server side as should be. And yes, there are some lag issues, but it isn't usually too big a deal unless your from Australia playing on the NA servers. There have been a few smaller issues with the usual stuff like wall hacking etc, that of course is going to always be there, but compared to a lot of other cryengine games (ArcheAge, Warface, Crysis itself, etc.) this one does a phenomenal job.

Edited by Darth Futuza, 23 May 2016 - 10:30 PM.

[LordNothing]

Darth Futuza, on 23 May 2016 - 08:39 AM, said:

I don't understand all these naysayer comments, half of the people in this thread are acting like they're top industry c++ programmers

real game programmers use c.

[Troutmonkey]

Darth Futuza, on 23 May 2016 - 10:19 PM, said:

I disagree, it's one of the more stable applications I've worked with. Crashes less often then League of Legends and the game works. There's definitely some areas that I think could use improvement, but it's not fundamental design flaws. eg: The zoom code is shoddy and slow, battle grid and ui for the menus etc is poorly implemented, the core game itself however is fine. Granted I haven't tried attaching a debugger to it, decompiling, and running test cases (and this would violate the terms of service - reverse engineering and all that) - so there might be some awful design decisions, but based on output I think it's done well enough compared to most games. Is it as stable as the linux kernel? Hell no, but for a game it's fine. Their netcode is also performing nice, never heard of or seen cheats that actually work because they don't authenticate through the client and game actions are entirely server side as should be. And yes, there are some lag issues, but it isn't usually too big a deal unless your from Australia playing on the NA servers. There have been a few smaller issues with the usual stuff like wall hacking etc, that of course is going to always be there, but compared to a lot of other cryengine games (ArcheAge, Warface, Crysis itself, etc.) this one does a phenomenal job.

Zoom / Vision code should have never been server side. The syncing of the mechlab takes forever and should probably happen asynchronously in the background. Found lots of bugs when trying to break the UI by clicking buttons in various combinations quickly. ******* in game chat that STILL isn't fixed which has been broken since launch. Walking sounds that don't play on uneven surfaces like stairs. Warhorns not firing. Terrible menu scaling at resolutions that aren't 1920*1080. Lots of minor bugs that are easy to spot and fix but have never been. Just shows that there's a distinct lack of respect for QA and writing safe, clean code. Sure shooting works and the HSR is surprisingly really good, but pretty much everything seems like a hack that stops getting dev time as soon as it works, even if it doesn't work well.

[Navid A1]

When it comes to MWO, the standard is so low that i would be jumping up and down of happiness if they just let us use chat when we are searching for a match.

The sad thing is that the things they consider great features are super basic things introduced in the 90s and are pretty much standard in every game (like a polished and working UI)
It seems like that PGI's efforts are more on understanding the source code and cryengine rather than coding new things.

[Ryllen Kriel]

I still want infantry and mech mines and hovertanks and aerospace fighters! Even if the are old-school 2D animated sprites!

[lol lol lol lol lol lol lol lol]

Mister Blastman, on 23 May 2016 - 08:48 AM, said:

But that begs the question, much like the big unanswered ones...

"What is the meaning of life?"

"Why do I exist?"

"What is the Universe?"

"What is space?"

"If I think, does that mean I am?"

And at the bottom, here, on these forums, lurking in the back of our minds is the ever elusive burden dangling from our ganglia...

"What is PGI?"

Perhaps we should build a computer to solve it.

I bet if we built a computer to solve the answer to "What is PGI" on top of "What is life, the universe and everything" it'd be done quicker than the time it would take for PGI to finish making MW:O.

[Mystere]

Johnny Z, on 23 May 2016 - 06:29 PM, said:

Someone already said and it has been said a few times, but the engine, Cryengine, is the original source code. "Technically" everything beyond that is "Modding". That goes for all the other games as well, and everything else, the models, textures, scripts, sound, could "technically" be described as "modding" for this topic.

[Darth Futuza]

LordNothing, on 23 May 2016 - 11:32 PM, said:

real game programmers use c.

Cryengine, Unity, and Unreal Engine are all written in mostly C++ with a little bit of C, but okay...

Troutmonkey, on 24 May 2016 - 03:38 AM, said:

Zoom / Vision code should have never been server side. The syncing of the mechlab takes forever and should probably happen asynchronously in the background. Found lots of bugs when trying to break the UI by clicking buttons in various combinations quickly. ******* in game chat that STILL isn't fixed which has been broken since launch. Walking sounds that don't play on uneven surfaces like stairs. Warhorns not firing. Terrible menu scaling at resolutions that aren't 1920*1080. Lots of minor bugs that are easy to spot and fix but have never been. Just shows that there's a distinct lack of respect for QA and writing safe, clean code. Sure shooting works and the HSR is surprisingly really good, but pretty much everything seems like a hack that stops getting dev time as soon as it works, even if it doesn't work well.

Definitely agree about the vision/zoom stuff, also ui is very broken, the other stuff as well, but like I said those aren't the fundamental designs that are broken - it's small minor bugs that really shouldn't have ever been problems. You keep saying "safe" code, but these bugs don't make the executable unsafe or leak memory, or cause crashes.

I agree thought that it's probably safe to assume the observation of it not having very clean code is a possibility, since these bugs have existed for so long and their patches frequently 'fix' issues that cause new issues. It sounds very much like they have bugs that nobody wants to fix because that particular section of the code is hard to understand etc. That or they just don't care / are too lazy to fix it. Their UI is by far the worse thing though, dunno why, it ought to be easy with Cryengine.

Edited by Darth Futuza, 24 May 2016 - 09:35 AM.

[Troutmonkey]

Darth Futuza, on 24 May 2016 - 09:34 AM, said:

Cryengine, Unity, and Unreal Engine are all written in mostly C++ with a little bit of C, but okay...

Unity is C# or Javascript, but only chumps use JS

Darth Futuza, on 24 May 2016 - 09:34 AM, said:

Definitely agree about the vision/zoom stuff, also ui is very broken, the other stuff as well, but like I said those aren't the fundamental designs that are broken - it's small minor bugs that really shouldn't have ever been problems. You keep saying "safe" code, but these bugs don't make the executable unsafe or leak memory, or cause crashes.

Probably the wrong use of safe here. It's pretty easy to write "safe" code with an engine. What I mean is that a lot of the game states and buttons don't check for non-standard actions. Like if you triple click the find game button it would start searching but also show 2 "Matchmaking cancelled" prompts. You could "switch" mechs before a game start by pressing search and then quickly clicking a different mech before the search screen appeared. Other things like tool tips getting stuck on and not clearing. "Invalid Loadout" bugs in the mechlab after buying a new mech (got that twice from two purchases on Monday). The programmers / designers seem to have made one flow chart of expected use cases and then never checked what would happen if someone went outside of that.

Edited by Troutmonkey, 24 May 2016 - 04:26 PM.

[Mystere]

Troutmonkey, on 24 May 2016 - 04:25 PM, said:

The programmers / designers seem to have made one flow chart of expected use cases and then never checked what would happen if someone went outside of that.

You just reminded me of all the people I have given the boot throughout the years for insisting on doing such things.

[Troutmonkey]

I've got a YouTube channel filled exclusively with MWO bugs which reminded me of a few more
Random Command (Fixed?) - HOW? There isn't even a keybind for assume command!

Spectate Button (Fixed) - Shows they never checked non-standard use cases

Invisible Mechs and flickering objects - Bad culling

Paperdoll bug - No thought put into how the flashing is triggered

Stuck Compass Markers -Not sure how they messed this up. Possibly trying to save CPU cycles?

FPS drops when a player is killed after playing for a while (fixed) - Memory leak of some sort


and more
There would be more but I've kinda given up recording bugs as it take 30 minutes to upload 20 seconds worth of video. Screw Turnbull for ruining Australia's internet

[wanderer]

Quote

the core game itself however is fine

That would be because the core game was never coded by PGI in the first place.

[Escef]

CryEngine's source code was recently released. Now, I haven't done much in the way of coding since high school, a couple hundred or so moons ago, so I didn't personally look into it. But one of my livestream viewers did, and he told me me there are multiple instances of the code having comments in it, from the CryCoders themselves, that effectively said "no idea what this does, look into it later", and he even showed me an example of it. I just sat there looking at it, like, "WTF is this? Who the hell does that?" There's actual code in CryEngine that no one knows what the hell it does... That's just freakin' scary, it's a miracle it works at all, never mind that people can mod it.

Edited by Escef, 25 May 2016 - 12:27 PM.