Project priorities (was: Control Panel replaced)

Lightning Hunter

I still don't understand why so much time is being spent on replacing things in Doomsday that worked perfectly to begin with... So much attention is needed elsewhere - like support for more user-made maps (Boom support and support for Mapinfo, etc.). I installed the latest version of Doomsday today after not having used it in a year or so. I was hoping Doomsday would be more stable and have support for more maps out there, but instead found a new layout that I had to learn all over again. One huge difference I have noticed between the control panel and this new layout is that if I'm changing the gamma correction, it doesn't show me what the actual in-game gamma looks like anymore as I change it (the control panel went invisible to show you what it looked like). Also, maybe I'm blind - but I can't find any options on detail textures in the new menu.

The new menu looks nice and all, don't get me wrong. It just seems that Doomsday continually takes several steps backward before going forward. I'm looking for the day when Doomsday gets back to version 1.9.8 status in the way of stability, and increases support for user-made maps.

Anyway, I'm not trying to put a damper on Doomsday progression. I'm just perplexed as to why Doomsday has taken this direction the last 7-8 years or so (or has it been longer than that?) I've been using Risen3d the last few years, because it never seems to fail me. It loads 95% of all the user-made maps out there, and is basic - which I've come to realize I prefer. It has taken years of me following Doomsday to realize that. Doomsday is no longer simple... Every time I download a new build of Doomsday, I find as many new bugs as I do fixes. I will still follow the progress and continue to hope that Doomsday becomes stable and reliable, but I think that may take a long time, if it ever happens.

Gordon

boom support is not top priority as it should be until the framework is in place to do that stuff. would you want an engine that has the support but no framework to support it?

skyjake

Lightning Hunter wrote:

why so much time is being spent on replacing things in Doomsday that worked perfectly to begin with [...] It just seems that Doomsday continually takes several steps backward before going forward.

The fact is that much of Doomsday's code has become outdated. It may appear to work "perfectly", however the way it has been implemented is actually preventing us from moving forward to the direction we want to go.

I'm interested in playing the long game: if one takes a wider view of the computing ecosystem, one can see large shifts happening. I don't want Doomsday to end up in an evolutionary cul-de-sac. Therefore it is crucial that we migrate away from obsolete technologies like OpenGL 1.x. However, the price of this change is that we must tear down / replace old code with better, modern code. In many cases this kind of changes have a deep impact throughout the code base. Also, over the years I and DaniJ have become better coders, and can now recognize many design flaws and inflexibilities in the old code — things that have become or will become roadblocks for further progress.

Lightning Hunter wrote:

So much attention is needed elsewhere - like support for more user-made maps (Boom support and support for Mapinfo, etc.)

Improving support of user-made maps has been one of DaniJ's top priorities for the past several years. Particularly in 1.12 we took significant steps forward with this.

Lightning Hunter wrote:

I have noticed between the control panel and this new layout is that if I'm changing the gamma correction, it doesn't show me what the actual in-game gamma looks like anymore

This could be known issue on Windows: http://tracker.skyjake.fi/issues/1107

Lightning Hunter wrote:

I can't find any options on detail textures in the new menu.

Detail texture settings are now in the Renderer Appearance sidebar editor.

Lightning Hunter wrote:

Doomsday is no longer simple... Every time I download a new build of Doomsday, I find as many new bugs as I do fixes.

Please do submit bug reports to our new tracker, if they are new issues. Obviously our goal is to maintain (or even improve) stability as we progress, however bugs are unavoidable.

This year the stable releases have focused more on new functionality. It is true that we would benefit from a stability/bugfix release, however that kind of work is difficult to focus on for a long period of time.

Vermil

Firstly, my apologies if the below comes across as presumptuous.

The vision for Dday 2 sounds and is awesome. I also fully respect the work nesscerry to achieve the vision.

But that is because I personally know something of it and the time has been taken to explain it to me in ways I understand (i.e very simple).

I really do not mean this to sound rude and I know the roadmap and wiki exist, but for whatever reason, the wider Dday community has no real clue as to what the vision is and never has. About the only real declaration of the vision by Deng Team is the little 'technology' box quietly slipped onto the bottom of the 'engine' page on the website during the overhaul of the site earlier this year.

The user base see a lot of work done on systems they don't understand the impact of, don't know that they are things they might not see the benefit of for a long time yet or, most importantly, things not directly relating to the games themselves; 99.9% of the community are neither coders or modders.

Improved support for limit removing maps, improved rendering hack support, camera vignette and the texture sharpness aside, there has been no real change in the games themselves since 1.8.6.

I don't feel it's my place to detail the vision (or what I know of it) on Deng teams behalf.

Of course, there’s the other side of the coin, that the vision is still a way off coming together and to go into detail about it could be considered hyping with nothing to show.

DaniJ

Unfortunately it would appear that many of the new features and improvements made over the past few years have gone unnoticed by the community at large.

As you say, Vermil, much of the friction regarding user expectations appears to be a result of miscommunication. With Doomsday our goals extend far wider than the "DOOM source port" definition. Perhaps we could have communicated better but I can comfortably say that ours is perhaps the most open of all source ports being actively developed today. We publish all future plans, a long term roadmap and provide frequent devblog updates.

I hear the argument that it seems that we are focusing on things the user doesn't care about. However, the reality is that is exactly what we are doing. Much of what we do may not appear to affect the game directly but dig a little deeper and you can see the logic behind how and why tasks are prioritized the way they are. It is important to bear in mind that with Doomsday we aim to provide a complete solution for playing all the games we support. By that I mean a solution that depends on no third party applications, utilities and/or system specific crutches. At the same time our ultimate goal (as ever) is the realization of an engine capable of supporting all 2.5d FPS games. Furthermore, as Doomsday began on the Windows platform it should be expected that much of the work done since 1.8.6 is not of interest to Windows users. Particularly if all one cares about is playing fullscreen DOOM.

Take for example the UI work which has featured in the past couple of releases. One might intuitively think this has no bearing on the games themselves however this is not the case. In a project as large as Doomsday it is crucially important that the high level architecture is clean (we subscribe to the don't repeat yourself principal of software development). So, rather than have a dozen or separate subsystems and definition languages for the Doomsday taskbar UI, game menu, game HUD, intermission/interludes, etc, etc..., our goal is a single UI framework that supports all these and more.

In such an integrated architecture it is not always immediately obvious to the end user how any given set of changes will ultimately affect the games themselves. If one is not prepared to read the development materials we publish then it is only natural that our work schedule seems something of a mystery.

skyjake

Vermil wrote:

the vision is still a way off coming together and to go into detail about it could be considered hyping with nothing to show

That is a very good point. At least I have tried to be very careful when communicating the future vision, as 1) the details may change and 2) it will take a long time to get there, due to the nature of the project and the size of the dev team. It is better to underpromise and overdeliver than visa versa.

We should try to continue improving the way the stable releases are documented, though, so that all changes can be understood and put in the proper context.

Lightning Hunter

It is important to bear in mind that with Doomsday we aim to provide a complete solution for playing all the games we support. By that I mean a solution that depends on no third party applications, utilities and/or system specific crutches. At the same time our ultimate goal (as ever) is the realization of an engine capable of supporting all 2.5d FPS games. Furthermore, as Doomsday began on the Windows platform it should be expected that much of the work done since 1.8.6 is not of interest to Windows users. Particularly if all one cares about is playing fullscreen DOOM.

You know, I never really thought deeply about what the goal was of the Doomsday developers in recent years - but now that I read DaniJ's post - I have come to realize that the goal is steering away from what I am looking for in a Doom engine. Doomsday, which started as Jdoom - was, and still is, the best looking engine around. However, as DaniJ pointed out - the goal now seems to be to combine all fps games in to one platform that runs without dependencies. This is obviously why there seems to be nothing that the average user cares about. The average user is looking for stability, reliability, ease of use, graphical improvements, and map support. All of these things will not be worked on for a log time, so it seems. For this reason, I will stick with Risen3d until the day comes that these things are worked on. With such a large goal, however - this may take years and years. Through experience, I have come to realize that small, basic projects are often far more successful than large projects. Intricate plans are rarely ever realized. I hope the goals of the Doomsday developers are met someday, but I just don't have the time anymore to download updates, find new bugs, get used to new layouts, and have no support for half of the user-made maps out there.

I wish you the best of luck, Skyjake and DaniJ - but I think I will view the project from afar until I see that it is more stable.