Bay 12 Games Forum

Please login or register.

Login with username, password and session length
Advanced search  
Pages: 1 [2] 3

Author Topic: State of Optimization (Returning Player)  (Read 3807 times)

Shonai_Dweller

  • Bay Watcher
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #15 on: February 24, 2021, 06:23:16 pm »

 
Quote
I think people's attitudes change quite a bit when they pay for a product (as opposed to donating to a project you want to support), and I do think new players will be less forgiving of the game's quirks than us bay12ers
You know the game is still free right?
Steam payment gets you a tileset and some music.

But, yes, Toady may choose to give up the incredible development plan and make a fortress defence clone to rake in some cash. A shame, but understandable. It's a tough world. Hopefully Steam sales will make up for the loss of those who donate to support the vision of the project.
« Last Edit: February 24, 2021, 06:26:40 pm by Shonai_Dweller »
Logged

PatrikLundell

  • Bay Watcher
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #16 on: February 25, 2021, 03:16:40 am »

Quote
I think people's attitudes change quite a bit when they pay for a product (as opposed to donating to a project you want to support), and I do think new players will be less forgiving of the game's quirks than us bay12ers
You know the game is still free right?
Steam payment gets you a tileset and some music.
:

The comment is very valid. People who pay for the Premium version pay for it regardless of it being available for free as well, and if you pay for something you expectations are different than if you support a hobby project. Sure some of the new players may opt to get the Classic version, but if the era of pirating is any guide, not actually having paid for something that's available for sale doesn't do much to lower the expectations (and voices) of those who didn't pay.
Logged

seht

  • Bay Watcher
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #17 on: February 25, 2021, 01:57:59 pm »



But, yes, Toady may choose to give up the incredible development plan and make a fortress defence clone to rake in some cash. A shame, but understandable. It's a tough works. Hopefully Steam sales will make up for the loss of those who donate to support the vision of the project.
[/quote]

What do you think the Steam money is going to do to change the way the game advances? This post is about performance, which has been an issue, in some peoples' opinion, for far longer than the announcement of the Steam release. Do you think improving it is even possible?
Logged

Nopenope

  • Bay Watcher
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #18 on: February 25, 2021, 02:40:37 pm »

He said there was quite a bit of low-hanging fruit he could address, but didn't get around to yet. The expectation is that paying customers on Steam will have a different relationship with the game than Toady fans, so there may be pressure to fix the more prosaic stuff and avoid a deluge of disappointed reviews (you know, the reviews that say "the game is promising but don't buy it yet"). But maybe my worries are unwarranted and he'll be showered in profits from the get-go, who knows. If I could predict the future I wouldn't post on a niche game forum, I'd be relaxing on my yacht bought with the payout from Gamestop's stocks.
Logged

PatrikLundell

  • Bay Watcher
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #19 on: February 25, 2021, 04:22:00 pm »

:
If I could predict the future I wouldn't post on a niche game forum, I'd be relaxing on my yacht bought with the payout from Gamestop's stocks.
Are you sure you wouldn't just relax in front of DF on your yacht when not posting on the forum :P
Logged

spinnylights

  • Bay Watcher
  • Wolf Leather Dress
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #20 on: February 28, 2021, 06:23:14 am »

This post is about performance, which has been an issue, in some peoples' opinion, for far longer than the announcement of the Steam release. Do you think improving it is even possible?

The game is in alpha after all

Given how complex DF is, I think it performs great considering how far it is from being finished. Having spent a lot of time optimizing software, I'm 100% positive that there are ways the game could be sped up here and there, but it's not necessarily a great idea to spend a bunch of time and effort optimizing half-finished, tentative, or prototypical features. Not only does it tend to make those parts of the codebase harder to work on afterwards (they get pretty arcane) but you might end up throwing away all your hard work if you need to change the feature later. Barring common sense algorithms + data structures decisions, it's typical to work on heavy-duty optimization once you're otherwise more-or-less finished. You don't even know where the worst bottlenecks are going to show up until then anyway.
Logged
Our little Kinsmen — after Rain / In plenty may be seen, / A Cross and Grumbly multitude / The soiled Ground upon // A needless life, it seemed to me / Until a little Roc / As to a Tavern saucily / Advanced and breakfasted. // As I of He, so Armok, me / I pondered, may have judged, / And left the little Mountain-folk / With Wariness enlarged. —–Emily the Poetic Thorns

Nopenope

  • Bay Watcher
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #21 on: February 28, 2021, 08:15:07 am »

The game calculates paths for every single creature on the map at every single frame. Do we really need this? Is it that big of a deal if sometimes the game gets pathing wrong?

Temperature calculations are kind of neat and make for a nice press article ("oh wow the game simulates temperature") but in gameplay terms it only matters for magma, fire, ice freezing/melting and extreme biomes. Can't the calculations be simplified a bit? Am I forced to choose between "magma is like water" and a FPS drop?

Weather calculations are also neat except for the fact that you can't ever see the clouds or the fog and wind only matters for windmills. Can't the calculations be simplified a bit? Am I forced to choose between "no precipitations ever" and "your CPU is hogged by computations that won't ever matter for your game"?

Some people will balk that I bring up such ancillary and plebeian notions like "gameplay", "play experience" and "concrete effects on the game" instead of standing in contemplative awe before the glorious invisible simulations while repeating the all-powerful mantra "the game is only in alpha" but maybe, just maybe, people who haven't supported Toady for the last 10-15 years and still expect the game to play like an actual game instead of a MoMa piece won't show such adoring displays of charitability?
Logged

PatrikLundell

  • Bay Watcher
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #22 on: February 28, 2021, 08:38:16 am »

The game calculates paths for every single creature on the map at every single frame. Do we really need this? Is it that big of a deal if sometimes the game gets pathing wrong?
:
Does it? There are frequent claims that this is the case, but if so, what are the path vector in the unit for? Why is pathing so stupid when blocked that it pathes around back to where it was blocked only to turn around and continue from there if it gets blocked in a long single tile corridor? The FPS hit by the infamous "cat pathing blocked by locked door" IS caused by the critter trying to path and fail every tick.

:
Weather calculations are also neat except for the fact that you can't ever see the clouds or the fog and wind only matters for windmills. Can't the calculations be simplified a bit? Am I forced to choose between "no precipitations ever" and "your CPU is hogged by computations that won't ever matter for your game"?
:
Wind mills are affected by latitude only, not by weather.  You'll get a permanent wind power output that's high, medium, or absent based on latitude.
Have you measured the extent of the "CPU hogging" of weather? Do you really want Toady to waste time on a "simplified weather" simulation, rather than developing new stuff/fixing old bugs/improving the UI? It's true that currently the progression of evil weather outside of the fortress (or area where the adventurer is) isn't visible or available in the game, but it would be possible to use the current framework to provide "weather forecast" warning about evil weather approaching, and I guess, normal weather forecast/divination for adventure mode (probably not useful at the fortress time scale, though).

:
Some people will balk that I bring up such ancillary and plebeian notions like "gameplay", "play experience" and "concrete effects on the game" instead of standing in contemplative awe before the glorious invisible simulations while repeating the all-powerful mantra "the game is only in alpha" but maybe, just maybe, people who haven't supported Toady for the last 10-15 years and still expect the game to play like an actual game instead of a MoMa piece won't show such adoring displays of charitability?
Thus far DF has been Toady's private game tinkering project that's full of tangents and possibly overly developed parts (with other parts being absent for years). Going commercial will probably be a very bumpy ride, with a new audience with different expectations than that of the current free/donation base.
Logged

Nopenope

  • Bay Watcher
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #23 on: February 28, 2021, 08:45:03 am »

To be clear I am not advocating that Toady simplify a specific part of the game, I just want to be able to have a normal fortress with default features that lasts more than a couple of years. Because we don't really know what really causes FPS drops beyond tentative heuristics I just put forward ideas but at the end of the day I don't care how he does it as long as the performance gain is there.
Logged

spinnylights

  • Bay Watcher
  • Wolf Leather Dress
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #24 on: February 28, 2021, 09:02:51 am »

To be clear I am not advocating that Toady simplify a specific part of the game, I just want to be able to have a normal fortress with default features that lasts more than a couple of years. Because we don't really know what really causes FPS drops beyond tentative heuristics I just put forward ideas but at the end of the day I don't care how he does it as long as the performance gain is there.

I really do feel like any aggressive optimizations right now (especially ones with design implications, but even purely technical ones) are probably premature given how far away the game is from being done. DF just isn't a normal game—most games don't take decades to develop, nor are they as ambitious in their mechanics (most games are not even CPU-bound, after all). Toady could easily spend a year just trying to grease the wheels everywhere and then end up tossing large portions of that code away after a few updates—I don't see why that would make people satisfied in the long run.

The only thing I feel frustrated by in all this is that the game is closed-source. If the codebase was public, it would be easier for other people (myself included) to work on making it run faster without Toady having to worry about wasting time on optimizations that might turn out to be useless—he could work on things with design implications and it would be okay if our tweaks got removed later. It would also be easier for people to either add extra configuration options or patches that would enable some of the changes you described without it being as big of a deal in terms of the impact it could have on the canonical game.
Logged
Our little Kinsmen — after Rain / In plenty may be seen, / A Cross and Grumbly multitude / The soiled Ground upon // A needless life, it seemed to me / Until a little Roc / As to a Tavern saucily / Advanced and breakfasted. // As I of He, so Armok, me / I pondered, may have judged, / And left the little Mountain-folk / With Wariness enlarged. —–Emily the Poetic Thorns

Manveru Taurënér

  • Bay Watcher
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #25 on: February 28, 2021, 09:20:32 am »

I really do feel like any aggressive optimizations right now (especially ones with design implications, but even purely technical ones) are probably premature given how far away the game is from being done. DF just isn't a normal game—most games don't take decades to develop, nor are they as ambitious in their mechanics (most games are not even CPU-bound, after all). Toady could easily spend a year just trying to grease the wheels everywhere and then end up tossing large portions of that code away after a few updates—I don't see why that would make people satisfied in the long run.

The only thing I feel frustrated by in all this is that the game is closed-source. If the codebase was public, it would be easier for other people (myself included) to work on making it run faster without Toady having to worry about wasting time on optimizations that might turn out to be useless—he could work on things with design implications and it would be okay if our tweaks got removed later. It would also be easier for people to either add extra configuration options or patches that would enable some of the changes you described without it being as big of a deal in terms of the impact it could have on the canonical game.

Well, as long as their livelihoods depend on it, making it open source doesn't make much sense. But hopefully the steam release will be successful enough they won't ever have to worry about money again (which iirc was talked about as possible allowing that to happen).
Logged

spinnylights

  • Bay Watcher
  • Wolf Leather Dress
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #26 on: February 28, 2021, 11:28:59 am »

Well, as long as their livelihoods depend on it, making it open source doesn't make much sense. But hopefully the steam release will be successful enough they won't ever have to worry about money again (which iirc was talked about as possible allowing that to happen).

At the risk of straying off-topic, I don't see why that would be true—after all, the game is already available for free, and people still fund its development. It's not as if having the source would change anything about that; just because other people could contribute patches to it doesn't mean its core authorship would change. They would still be in control of what made it into the codebase—as they should be, since its their project. I wouldn't want it any other way.  ;D

They wouldn't be an isolated case, either; for example, the core developer of the DAW Ardour is supported full-time by its users to work on it despite it being free software (as in beer and as in freedom—he actually makes close to $100k/year). As far as games go, I know the developer of the upcoming RTS EVERGLORY is working on the free engine full-time. And, perhaps it's worth a mention, I myself work for a living on games for which both the source and the media are freely available to anyone who wants them; if anything I would say we make more as a result than if we didn't do that because it gets people excited, not that I'm in it for the money really.

In any case, I definitely do hope the Adams brothers do very well from the Steam release—they have enough to worry about without being scared of medical bankruptcy or the like, and I admire them so much for making DF free as in beer; aside from it being one of my all-time fav games it's also just a rare thing to do in the game world. If a big Steam release helped them feel more comfortable sharing the game's code that would be lovely icing on the cake.
Logged
Our little Kinsmen — after Rain / In plenty may be seen, / A Cross and Grumbly multitude / The soiled Ground upon // A needless life, it seemed to me / Until a little Roc / As to a Tavern saucily / Advanced and breakfasted. // As I of He, so Armok, me / I pondered, may have judged, / And left the little Mountain-folk / With Wariness enlarged. —–Emily the Poetic Thorns

PatrikLundell

  • Bay Watcher
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #27 on: February 28, 2021, 11:51:25 am »

The source code issue isn't only about keeping a livelihood, but also about control of the development. Once the source code is released there will be people who will insist on having their favorite modifications incorporated into the game, and as far as I understand becoming a project administrator rather than a game developer is not an attractive change to them: I doubt they'd waste time on more or less shoddy submissions (just look at the bug tracker: there are quite a few reports that are rather clear on why things fail and what should be done to fix them). Also, once there are multiple variants there will be lots of bug reports incorrectly flooding DF rather than the copy generating the bug.

Hiring or otherwise having to deal with project administrators and the associated bureacrazy doesn't seem to be high on their wish list either, although they've taken the step to work with Kitfox and the artists: that may potentially cause them to find that this coordination overhead might not be quite as horrible as feared.
Logged

Nopenope

  • Bay Watcher
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #28 on: February 28, 2021, 12:22:09 pm »

Their wanting to control the development is only contingent on their livelihood not being threatened. I may be wrong, but I don't think they'd mind people developing an alternative forked "Dwarf Fortress" if it meant a seven figure annual income for them to keep working on whatever they want.

In the situation of their opening the source code, I don't even think they'd be even "administering" the game development: rather, a team of trusted volunteers would probably do that, fix the bugs, accept patches and merge the branches accordingly while Tarn would just keep churning on whatever features he likes.

Of course this all raises the matter of a guarantee of said annual income or team of kind-hearted volunteers, as real-life constraints exist, drama in open-source communities is a thing, and people working for free do have a habit of disappearing with no notice or justification. But the more time passes and the community shows unwavering support, the more confident the Adams brothers can be. I could understand their reservations 10 years ago when the project was still pretty much in its infancy and Tarn's reputation wasn't that widespread, now he's carved a solid enough niche for himself that he could take more liberties and be more trusting imo.

In any case, the Steam release will prove crucial in the direction of future development. Ideally it'd be a resounding success, set them for life, so that they feel comfortable opening up the source or parts of it. In the worst case there will be backlash as new players stumble upon the game's quirks and unfinished state and performance issues, requesting refunds and forcing Tarn to delay his pet features for more bugfixing and polishing, which is tedious work and kinda demoralizing. More likely, it will be something in-between.
Logged

Starver

  • Bay Watcher
    • View Profile
Re: State of Optimization (Returning Player)
« Reply #29 on: February 28, 2021, 02:22:27 pm »

Their wanting to control the development is only contingent on their livelihood not being threatened. I may be wrong, but I don't think they'd mind people developing an alternative forked "Dwarf Fortress" if it meant a seven figure annual income for them to keep working on whatever they want.

I think you're wrong in at least three key aspects, there.

But that's just my impression. Obviously far from yours. Likely we'd both be considered mistaken, but in different directions.
Logged
Pages: 1 [2] 3