79,221
edits
Project:General Discussions/Archive: Difference between revisions
Project:General Discussions/Archive (view source)
Revision as of 07:13, 29 November 2016
, 29 November 2016Archive a few other completed discussions, other "Recent Discussions" seem to have outstanding issues
No edit summary |
(Archive a few other completed discussions, other "Recent Discussions" seem to have outstanding issues) |
||
| Line 840: | Line 840: | ||
We need to get this documented on the Global Problems template docs. [[User:Kolano|Kolano]] ([[User talk:Kolano|talk]]) 08:24, 6 October 2015 (CEST) | We need to get this documented on the Global Problems template docs. [[User:Kolano|Kolano]] ([[User talk:Kolano|talk]]) 08:24, 6 October 2015 (CEST) | ||
=== Performance Addition to Game Wiki Pages === | |||
As Kolano said in his edit, performance is not the goal of this wiki. The default settings are already the best balance of speed and compatibility, the wiki exists to help people know what those settings are and what they mean and to catalog problems. Users can break things on their own juuuust fine! While I'm at it, why did you place notes throughout the page? You aren't intending to have that in the live version are you? - [[User:MaJoR|MaJoR]] ([[User talk:MaJoR|talk]]) 17:11, 26 April 2015 (CEST) | |||
:It's a WIP, or RIP really (revision in progress), that I will propose along with the other proposal I am working on. The goal is to make pages easier to read and to appeal to both audiences; people who want performance over emulation accuracy '''and''' people who want emulation accuracy over performance. --[[User:Wildgoosespeeder|Wildgoosespeeder]] ([[User talk:Wildgoosespeeder|talk]]) 22:14, 26 April 2015 (CEST) | |||
::I think you are misunderstanding the approach the wiki is taking. Our job is not to balance speed and performance, but to inform the user of accuracy issues and problems and let them make their own decision on how to balance their own needs. This is the only way the wiki can remain objective - "balance" would vary depending on the user's system, the game and problem at hand, and their own personal opinion. If you make a post saying that a certain setup is the best balance, and I disagree with that, who's right? No one is! It's impossible to resolve, since it's all subjective. But if it's accuracy related, we can test it, confirm it, and apply it. By sticking to accuracy it gives us a reasonable way to make a reliable wiki, and the users can be informed about what a problem is, how to avoid it, and the costs it will take; allowing them to make their own decisions on whether those costs are worth it to them. That said, we do try to inform them of the costs, especially if it is something nasty like Single Core. - [[User:MaJoR|MaJoR]] ([[User talk:MaJoR|talk]]) 08:36, 29 April 2015 (CEST) | |||
::::I think you misunderstanding what I am proposing. I am not proposing accuracy loss. I am proposing an additional section that documents performance tweaks as well as preserve the section for accuracy. To my understanding, the current way pushes accuracy as the forefront while performance is an afterthought. The performance tweaks are there, yes, but are "camouflaged". Somewhat harder to see. Look at [[Interactive Multi Game Demo Disc 2001-10/sandbox|the sandbox]] again. Maybe you will understand what I am suggesting. Most people will want to play their games with minimal distractions, not see how accurately Dolphin can emulate their favorite games. Some do want an accurate emulator. I'm proposing satisfying both parties, not just one or the other. I am not suggesting a middle ground either. --[[User:Wildgoosespeeder|Wildgoosespeeder]] ([[User talk:Wildgoosespeeder|talk]]) 09:28, 29 April 2015 (CEST) | |||
::I'm somewhat supportive of adding a performance tweaks section, but there would need to be some fairly strict guidelines on the sorts of things we'd capture there. For instance it may be nice to work out games that don't use EFB, and can safely disable EFB effects (something it might be nice to be able to effect via the game ini's actually). Having some coverage of issues raised by various performance tweaks might also help avoid those getting mixed in as reported bugs. There's also the subject of visual tweaks, where we might cover things like compatibility of AA / Antis-tropic / Force Filter settings (another thing it might be nice to be able to disable via ini) and/or provide a spot for HD texture pack details. | |||
::Per concerns already raised, we should take this slowly though to try to avoid it becoming messy. [[User:Kolano|Kolano]] ([[User talk:Kolano|talk]]) 09:48, 29 April 2015 (CEST) | |||
::Honestly I don't really see much point... Games that need DSP LLE are marked as such, those that need EFB to Ram are marked as such, so saying "this game is fine with EFB to Texture" is just redundant. This is why we removed config entries that recommended EFB to Texture and created the best accuracy rule in the first place. Well, that and people filling out a page with their favorite settings irregardless of how it broke things! The accuracy rule gave us a standard of objectivity that kept favorite settings and things that can be reproduced away from the wiki, and helped it be as useful as it is. Games that are ok without EFB is a valid point... if users could disable efb copies anymore! Most of the dumb settings like that have been removed. :P And if antialiasing or AF or something breaks something or is needed, it's already mentioned (see skyward sword). "Camouflaged" it may be, but all of the information is there, so honestly I think all of this is covered already by existing conventions. If a game needs something that's intensive, it's mentioned, and if it doesn't, use what runs fastest (which is already the default...) - [[User:MaJoR|MaJoR]] ([[User talk:MaJoR|talk]]) 02:51, 3 May 2015 (CEST) | |||
:::On a final note, any performance additions to this wiki should follow the objectivity rule: they would have to measurable and reproducible. Part of my resistance to this idea is just how badly things were with performance in the past, because people would put things up saying it's "faster" and then we'd kill it for being dumb and they'd put it back '''insisting''' it is better and must be on the page! We were only able to clean that up by implementing the accuracy and objectivity concepts. So I'm definitely scared by all of this performance talk, as it is trying to bring back some of the things that let crap enter this wiki. While I am not very open to it, at the very least it would need to follow objectivity standards, and have a plan prepared ahead of time for dealing with "but these are the best settings!" people. ...and there is also the fact that we can't test every game and every post that comes through, and this performance thing could let users sneak more weird settings in since it's harder to understand performance impacts since it varies so much with hardware. Ok yea I'm still against it. But my point is, objectivity rules and preparations against ignorant users must be involved in this! - [[User:MaJoR|MaJoR]] ([[User talk:MaJoR|talk]]) 02:58, 3 May 2015 (CEST) | |||
::::One thing I noticed that Dolphin doesn't do is have a reset button or something to restore default settings. Sure, there is dialog for the recommended setting if the user is unsure of the default setting, but no convenient button. If a button were implemented, you could use that to further enforce the importance of first testing with default settings vs. them deviating from the default settings where the game had bugs running under Dolphin. That way real bugs are filtered away from false bugs induced by the user. --[[User:Wildgoosespeeder|Wildgoosespeeder]] ([[User talk:Wildgoosespeeder|talk]]) 04:40, 10 May 2015 (CEST) | |||
:::::In general and in my opinion, Dolphin and its support pages aren't the most user-friendly I have used. It seems to be geared towards enthusiasts. I can use it and understand what I need to know to make the emulator work, but I don't know if the average user will understand like we do. That is the root concern I have and I am trying to address it one problem at a time and see what you guys think of the solution I came up with. I have a lot of practice with this in general. I hope to be of good help by providing valuable feedback! --[[User:Wildgoosespeeder|Wildgoosespeeder]] ([[User talk:Wildgoosespeeder|talk]]) 04:40, 10 May 2015 (CEST) | |||
::::::Well... the goal of the wiki is to educate people, and we tell them how to do things, how to fix problems, etc etc while letting them make their own decisions on techniques and performance. And the data is kept objective so it can be repeatable and sustainable long term. Not every user is going to want to know why this or that does this and why it fixes this bug, but the emulator has many support mechanisms to help those users already. The wiki serves as the last part of a chain, not the first. GameINIs catch most things, the forums/irc/reddit and other support catches the rest, and then we inform the enthusiasts that provide support as well as users who want to learn and grow on their own. There are many layers of support, and we are the most technical. '''And that's ok!''' It is what wikis are best at: education. The wiki serves as a long term database of technical information regarding the impact of features and changes in the emulator and ways to work around various issues, something that none of the other support mechanisms can do effectively. I'm totally open to improving our ability to educate, but you need to understand our resistance: what we have works, and it was the first (if not the only, I don't know of any others) emulator wiki to achieve this kind of reliability. The Dolphin wiki has succeeded because of the users willing to learn and share that knowledge, but also because of the curators who have worked hard to develop rules and structures that foster the usefulness of the wiki and encourage users to contribute. The decisions we admins have made over the years, such as the objectivity/reproducibility standard, were made to create something that is reliable and informative to end users but sustainable to wiki editors and admins. It is a very careful balancing act, but it's working! So there is going to be some resistance when a single new person comes out of the blue and proposes sweeping changes. While that may be a little frustrating for you, please understand that we're trying to listen and understand the new ideas so we can incorporate things that will improve the wiki, but also protect the systems that have worked so that the changes improve the wiki and not make it worse. To that end, you have repeatedly said that you have worked in other sites that work better, but you have never shared them. If you could show them, it would really help us understand where you are coming from, and help us pick up good ideas from them on our own. Would you mind sharing those references? - [[User:MaJoR|MaJoR]] ([[User talk:MaJoR|talk]]) 09:51, 10 May 2015 (CEST) | |||
:::::::I never mentioned I contributed to other websites the way you are implying. I have contributed, yes, but never pushed dramatic changes or ideas to how something could work better for a website. Usually it's already good as-is. The Dolphin website is the first time I felt I could do it more confidently. As for making things more user-friendly, they are mostly related to designing interfaces for sample programs I have made that I never made the source code public. I do have them, but I haven't really bothered to go through my various projects (big or small). I don't have a website to showcase them. I find designing a website from scratch cumbersome. If I have predesigned assets to work with and an easy drop-in system, like using Visual C# 2008/2010, then I can design something rather easily. It's like building with LEGOS vs. drawing a picture for me. LEGOS come easier for me. I have the pieces here on this website, just not the ability to create them. I just find the pieces are not being used most effectively. --[[User:Wildgoosespeeder|Wildgoosespeeder]] ([[User talk:Wildgoosespeeder|talk]]) 10:30, 10 May 2015 (CEST) | |||
=== Perfect Compatibility? === | |||
I'd like you to take a look at the SSB:M compatibility page. This game is rated 5 stars as "Perfect: No issues at all!". Yet, there seems to be one or two minor, ''minor'' errors listed on the page. | |||
So what is the specific criteria for a perfect rating? Would no known bugs, period, be too out of the question? Or is "perfect" emulation accuracy unreasonable anyhow, considering it isn't cycle for cycle accurate and won't be anytime soon. Basically, can we define the word perfect? | |||
[[User:MirandaStreeter|Miranda]] ([[User talk:MirandaStreeter|talk]]) 01:16, 24 October 2013 (CEST) | |||
Perfect should mean there are no known issues/defects with the reproduction. It's unclear why Melee wasn't knocked back to 4 stars earlier, since it clearly has open issues. As you note it doesn't necessarily mean there is an exact 1:1 reproduction, as we're dealing with high-level emulation in many places. There should not be noticeable image defects/gameplay issues though. So for instance, even if a character starts to talk a tenth of second too late in some title; it could still be rated perfect, if that doesn't effect gameplay and would generally be unnoticeable without a stopwatch. [[User:Kolano|Kolano]] ([[User talk:Kolano|talk]]) 04:13, 24 October 2013 (CEST) | |||
:SSBM isn't the only offender. A number of games seem to be 5-star happy yet have known issues listed without fixes. These include: | |||
::* Aggressive Inline | |||
::* Harvest Moon: Magical Melody | |||
::* Ikaruga | |||
::* Sonic Adventure DX: Director's Cut | |||
::* Tales of Symphonia | |||
:Yet more are listed at 5 stars without any version compatibility reports or test. None at all. These include: | |||
::* Donkey Konga | |||
::* Dora the Explorer: Journey to the Purple Planet | |||
::* Hitman 2: Silent Assassin | |||
::* Hudson Selection Vol. 3: Bonk's Adventure | |||
::* Mr Driller: Drill Land | |||
::* Pac-Man World Rally | |||
:You'll also find some 5-star games that require DX9 backends as a fix, despite it being entirely dropped in latest revisions. These include: | |||
::* Hot Wheels World Race | |||
::* Sega Soccer Slam | |||
:While all of those are obvious fixes, others aren't so much. Many of the rest either have conflicts between the reporting graph and actual report entries, list 4 stars in the graph while reading 5 stars in the official list, are listing compatibility results without any details, or haven't had a test entry since v3.0 and previous. These would have to be looked into on a case by case basis. I'd also check and see if a specific user/IP is marking 5 stars prematurely, or if it's a collective problem. --[[User:MirandaStreeter|Miranda]] ([[User talk:MirandaStreeter|talk]]) 21:10, 30 October 2013 (CET) | |||
----- | |||
Whoa now. This is exactly the problem. ALL GAMES HAVE SOMETHING WRONG WITH THEM. A tiny lighting bug. Tev derp. SOMETHING. By the criteria of "no known problems" there should be no five star titles. And even if we keep the ones that have no known bugs, it is just a matter of time until someone grabs one of those tiny global problems and there we go. The "perfect" rating concept is inherently flawed. | |||
As was discussed in the rating changes below, it is the opinion of the devs (minus skidau) that 5 stars ''should not mean absolute perfection''. 5 stars is something we should use, and since the emulator isn't perfect then nothing would be five stars. The criteria that was discussed is below in the Rating Changes section. That criteria is "almost perfect". Very very very good, with allowances for very minor bugs. Melee is a perfect example. Everything is perfect, except for some very very tiny bugs that no one bug someone who goes back and forth from console and dolphin a lot would notice (JMC47 and the netplay people in this case). Under the "almost perfect" criteria Melee is five stars. | |||
We need to work this out. We've been putting off this rating thing forever, and now that demotions are starting to happen based on the old ratings, it's time to get this settled. Until this rating issue is settled please do not demote any more 5 star games. Kolano, if you revert 5 star pages I won't get into an edit war with you, you pretty much handle all the boring tasks like this and I respect that. But anyone else? I'm reverting all five star demotions until the ratings issue is settled permanently. | |||
For the record, I promoted Melee to five stars based on encouragement from the devs that it qualified for "perfect" status regardless of the little bugs (again, skidau objecting :P). - [[User:MaJoR|MaJoR]] ([[User talk:MaJoR|talk]]) 22:27, 30 October 2013 (CET) | |||
== 2016 == | == 2016 == | ||