OpenMW's Roadmap and Future

Anything related to PR, release planning and any other non-technical idea how to move the project forward should be discussed here.
User avatar
AnyOldName3
Posts: 2668
Joined: 26 Nov 2015, 03:25

Re: OpenMW's Roadmap and Future

Post by AnyOldName3 »

My opinion on MWScript-MWSE is that it's dangerous to even say we might support a subset of it. I know it makes changes to the original parser to allow constructs the unmodified original CS didn't allow, so if we're going to support the whole thing, it's definitely going to require significant work with significant oversight from core OpenMW developers, even if it's just to prevent and repair regressions it causes. As MWScript is such a fragile beast in the first place, and we've been forced to replicate many of its quirks, it's pretty easy for any minor change to break edge cases everyone's forgotten about in ways that won't be noticed for a long time and will require a lot of effort to fix which won't come from a drive-by developer who originally caused the regression. Obviously, there'll be bits that are so simple to add that they don't have these consequences, but it'll make the anything-that-doesn't-require-MWSE-or-MGE-XE-will-work statement less clear, and once people see we've got partial MWSE support, they'll want it to move more towards full MWSE support.
User avatar
Greendogo
Posts: 1467
Joined: 26 Aug 2011, 02:04

Re: OpenMW's Roadmap and Future

Post by Greendogo »

From the beginning with Nicolay Korslund and much of the time with Zini OpenMW was just supposed to be a 1:1 recreation of the Morrowind experience with the ability to play all vanilla compatible mods.

Having said that, this isn't written in stone. Of course the most important value we can share is group cohesion and to create a space where we can share in an exciting creative process.

I agree that if someone wants to add a feature such as MWSE support that would be great, however it should not be a version 1.0 blocker as it isn't part of the original game. Feel free to argue the point, additional well made features only make the engine more fun. But remember, it isn't that important in the grand scheme of things if it comes into the engine now or in several years.

Another thing, it isn't important if OpenMW 1.0 comes out this year or in another 10. The game is mostly playable and 1.0 is mostly just an acknowledgement that we have accomplished OUR original goal, which is to create a 1:1 recreation. There is little to no value, to me, to release a "1.0" if the label rings false.

On that note, thanks for pointing out the bounty page. I strongly believe that the Gamepad GUI Mode belongs in version 1.0. I know hardly anyone agrees with me on this, but it has been my longstanding desire to see this feature get some love. I now have far more money than time to develop it out myself, so I feel like this bounty system is perfect for that. ;)
User avatar
psi29a
Posts: 5356
Joined: 29 Sep 2011, 10:13
Location: Belgium
Gitlab profile: https://gitlab.com/psi29a/
Contact:

Re: OpenMW's Roadmap and Future

Post by psi29a »

Greendogo wrote: 14 Jul 2020, 09:50 From the beginning with Nicolay Korslund and much of the time with Zini OpenMW was just supposed to be a 1:1 recreation of the Morrowind experience with the ability to play all vanilla compatible mods.
That pretty much sums up things and while we've been at it, we've gone above and beyond that in some areas.

It would seem that the most ardent Morrowind players and modders tend to think of Morrowind now not as vanilla, for them vanilla is now Morrowind+MCP+MGE:XE+MWSE-lua; that is a perception thing.

So if we stick to our guns, then we can plant the flag on 1.0 when our blockers are resolved.

I'm fine with leaving MWSE-1 unimplemented; but I do believe there is value in some MCP and MGE:XE parity, some of which are quality of life improvements. It would be nice to have the MCP wiki page , for example, filled out with reasons why we won't implement certain things that way we can refer to in the FAQ and Launcher help. Same with MGE:XE.

The question is, do we make those blockers or just... yeah, we'll get to them eventually and not delay the "1.0" flag planting ceremony?

To address another topic that akortunov and anyoldname3; short-term fixes really are a wack-a-mole at this stage of development. Which... oddly enough, was how Morrowind was developed as late-game changes had unintentional consequences in the engine. So having a long-term plan to guide development is ideal. We have zini's post 1.0; though some has changed at this point. I'll go over it again with him to get more insight.

How do we handle user response for fixing 'issues' they seem as bugs (forgotten features) that others consider to be cheating, oversights and actual engine bugs that are exploited? We seem to see-saw on some MCP issues for example and most recently about level progression when reading books about skills we already have maxed out. The tool tip itself already says something to the effect: maximum proficiency has been reached when you hit 100; so there is no more experience to be gained from reading. It could be argued that such an oversight is integral to level progression in Morrowind.
https://gitlab.com/OpenMW/openmw/-/issues/5516
User avatar
akortunov
Posts: 899
Joined: 13 Mar 2017, 13:49
Location: Samara, Russian Federation

Re: OpenMW's Roadmap and Future

Post by akortunov »

Greendogo wrote: 14 Jul 2020, 09:50 From the beginning with Nicolay Korslund and much of the time with Zini OpenMW was just supposed to be a 1:1 recreation of the Morrowind experience with the ability to play all vanilla compatible mods.
Yes, and the plan was to introduce Morrowind's behaviour by default and provide modmakers a way to change it during post-1.0 dehardcoding.
It is the main reason why Zini and scrawl usually declined changes in game mechanics, even if they came from MCP.
The same thing for large non-Morrowind features, which affect gameplay and game balance. So no ammo recycling, reverse pickpocketing, Dark-Souls-style combat and so on.
User avatar
psi29a
Posts: 5356
Joined: 29 Sep 2011, 10:13
Location: Belgium
Gitlab profile: https://gitlab.com/psi29a/
Contact:

Re: OpenMW's Roadmap and Future

Post by psi29a »

akortunov wrote: 14 Jul 2020, 17:05
Greendogo wrote: 14 Jul 2020, 09:50 From the beginning with Nicolay Korslund and much of the time with Zini OpenMW was just supposed to be a 1:1 recreation of the Morrowind experience with the ability to play all vanilla compatible mods.
Yes, and the plan was to introduce Morrowind's behaviour by default and provide modmakers a way to change it during post-1.0 dehardcoding.
It is the main reason why Zini and scrawl usually declined changes in game mechanics, even if they came from MCP.
The same thing for large non-Morrowind features, which affect gameplay and game balance. So no ammo recycling, reverse pickpocketing, Dark-Souls-style combat and so on.
That's playing fast and loose with the facts, Zini didn't have have a problem with MCP fixes, especially when they fixed bugs and changed game mechanics in Morrowind. That was a judgement call he made early on even before scrawl came on board. The more controversial things got put behind settings though but were still allowed in. Yet other things weren't necessary and yet others was considered just for modding.

So the "introduce Morrowind's behaviour by default" statement is close but still not accurate. The idea is to reimplement Morrowind without the flaws, bugs and oversights. Take it back further, Korslund just wanted to play Morrowind (and others) on Linux. :lol: (I sympathise, me too.) So yes, OpenEngine was a thing for him.

There are multiple competing sides to this, which doesn't make OpenMW consequent. The the desire to make OpenMW great, as seen here: https://wiki.openmw.org/index.php?title=Wishlist which was first laid out by Lgro and Zini and then updated by others, including scrawl. Then the other side was the push-back saying that some features should better of be left to modders/mods and then yet another side that we should stick as close to Morrowind gameplay as possible.

There are features now in OpenMW where by it was questioned if it wasn't best if left as a mod, but the work was done and work on mwscript+ wasn't yet begon, let alone openmw-lua. There is some nuance here. Do we leave a particular feature up to a modder or do we add it to OpenMW to enhance Morrowind? It's a case by case situation.

I wouldn't be surprised that once openmw-lua is in place that we start cutting out and removing code and pushing them into mods. There are other 'features' that were added but no one cares to maintain and so they rot away: 360 degree screenshots. It might get to the point that code (and features) begin to be removed as well if they get in the way.

As for roadmap, I'll let Lgro and Zini's own words sum that up from here: https://wiki.openmw.org/index.php?title=Talk:Roadmap
The issue-tracker is the roadmap. As anticlimactic as that is.

So if we want to move OpenMW into "1.0" that those issues labelled as such are our blockers for release. The rest are nice to haves.

Update: I have also have feedback from Zini about the roadmap and I'll be adjusting it accordingly as well.
User avatar
raevol
Posts: 3093
Joined: 07 Aug 2011, 01:12
Location: Caldera

Re: OpenMW's Roadmap and Future

Post by raevol »

Sorry quick reply to say this is awesome and thank you for posting, I will read the whole OP and thread soon!
User avatar
psi29a
Posts: 5356
Joined: 29 Sep 2011, 10:13
Location: Belgium
Gitlab profile: https://gitlab.com/psi29a/
Contact:

Re: OpenMW's Roadmap and Future

Post by psi29a »

Another conversation I've had was in regard to support of Oblivion and Skyrim NIFs, after talking about how it would help modders in that they could use more modern tools (and later versions of Blender) to create content for OpenMW, this was Zini's response:
I remember someone saying that they also want to go into that direction. Not
what I would have done, but I understand the reasoning. And the point
about blender is certainly valid. If anything, that seems like a
priority to me.
Keep in mind, this is just for NIF support for now pre 1.0. This doesn't mean that OpenMW will support playing Oblivion, Skyrim or anything like that aside from just being a walking simulator. OpenMW's focus up to 1.0 is still being able play Morrowind and create games like Morrowind.

But I'm sure that someone will fork OpenMW and add /app/openob and continue on that way if they wanted. Encouraging collaboration would bring in more interested developers, far more than those interested just in Morrowind which is quite niche to begin with. This dove tails nicely in post 1.0 de-hardcoding which opens the doors back up to something "OpenEngine" like.
User avatar
raevol
Posts: 3093
Joined: 07 Aug 2011, 01:12
Location: Caldera

Re: OpenMW's Roadmap and Future

Post by raevol »

psi29a wrote: 15 Jul 2020, 21:34 I wouldn't be surprised that once openmw-lua is in place that we start cutting out and removing code and pushing them into mods. There are other 'features' that were added but no one cares to maintain and so they rot away: 360 degree screenshots. It might get to the point that code (and features) begin to be removed as well if they get in the way.
Since I've actually been writing code professionally now (for like a few months, don't take me too seriously), I think this is the most important part of this whole discussion. For my $0.02, anything that isn't a 1:1 vanilla engine re-implementation item or enabling functionality in a Lua/shader/mod API is completely tangential to 1.0. Multiplayer would be the exception there because it's so integral to the architecture.

That doesn't mean that people shouldn't work on non-1.0 features if they want to, but in my humble opinion, our entire goal here should be to make the engine as moddable as possible, and to enable as much functionality to be implemented outside of the engine as possible. Some day Morrowind itself should just be a "mod" that OpenMW can run.

I think another big architecture thing that we're not discussing here is the fixed function/programmable graphics pipleline. AON3 can you weigh in on this?
User avatar
akortunov
Posts: 899
Joined: 13 Mar 2017, 13:49
Location: Samara, Russian Federation

Re: OpenMW's Roadmap and Future

Post by akortunov »

psi29a wrote: 15 Jul 2020, 21:34 That's playing fast and loose with the facts, Zini didn't have have a problem with MCP fixes, especially when they fixed bugs and changed game mechanics in Morrowind.
When I implemented a fix for pickpocketing formula from MCP, he declined to merge it with such motivation:

"This change seems to be outside of the scope of the OpenMW. It should be up to mod developers to make changes of this kind. I expect OpenMW to gain the necessary modding capabilities relatively quickly once we enter the post-1.0 de-hardcoding development phase. Hopefully 1.0 isn't too far off anymore (we made some substantial progress in the last couple of months)."

This is Scrawl's attitude towards pickpocketing system:

"I don't see the current system as broken, it's just intentionally hard."

It was a year 2017, by the way. Fow new gameplay features Zini used the same approach:

"As mentioned this feature can affect the balance and there is at least the possibility that it could break some particular content. That is not something for the user settings and definitely not for 1.0."

So from what I can tell, generally the attitude was the "Morrowind's game mechanics is not broken, it is just questionable".
psi29a wrote: 15 Jul 2020, 21:34 The idea is to reimplement Morrowind without the flaws, bugs and oversights.
By modern standarts, half of Morrowind are flaws, bugs and oversight. The issue here is that most of possible improvements are either subjective or break some content (sometimes both). Basically, there is a two kind of issues:
1. OpenMW behaves as Morrowind
2. OpenMW does not behave as Morrowind

In the first case you can just say "OpenMW is Morrowind's reimplementation, so it behaves as Morrowind does". In the latter case you obviously can not do a similar thing since most of deviations are questionable and many players treat them as bugs, fill bugreports and/or complain about them on forums and social networks. In such cases your politics was either to eliminate such deviations or hide them behind of options (disabled by default) with plan to remove them during de-hardcoding. Also keep in mind that there is only one Morrowind's way to do X, while there is an unlimited number of non-Morrowind's way to do X. As for me, it is simpler to implement one way to do X and allow modmakers implement the rest than try to hardcode multiple ways to do X just to spend an additional time to de-harcode it.
User avatar
Zini
Posts: 5538
Joined: 06 Aug 2011, 15:16

Re: OpenMW's Roadmap and Future

Post by Zini »

Wow, some heated discussion. Let's clarify a few things.

Yes, we (both Nico and I) started out with the goal to make OpenMW 1.0 an exact copy of original MW, in regards to functionality at least. And then the plan was to introduce more modability that would allow mod authors to deviate from the original behaviour.

However this goal was quickly squashed by the reality of what mess original MW is and since then we have been operating under the principle of aiming for a nearly exact copy of the original. A prime examples of this is of course the scripting engine, which is much more capable than the original. That actually broke Morrowind.esm at some point, because Morrowind.esm relied on some scripts to fail in a certain way. We implemented a workaround for this special case, but the situation remains unchanged for other content files and probably should stay that way.

Generally we deviated from the original in two cases:

1. Sticking with the original would have caused substantial harm to our plans for future improvements.
2. It was obviously broken, the fix was of minor complexity and fixing it would have been somewhat unlikely (but not impossible) to break existing content.

As for implementing script extenders, I originally rejected these outright for two reasons:
1. Security concerns
2. Issue #1 from above.

Point 1 is still valid of course. Point 2 less so (since we are going to discontinue MWScript), depending on what subset we want to implement. It would require a developer with a lot of spare time and some expertise. And it would require lots and lots of testing. This is a task that can be undertaken, but it shouldn't be undertaken lightly. The concern that such a change could break things is certainly valid.

I hope this cleared things up a bit.
Post Reply