What changes would you like to see?

Dear Fellow Developers,

With the potential dawn of a new era for the Multiverse engine, I think that it would be a good idea to discuss what we actually want to do with the engine, if it were to go Open Source.

So, what changes would you like to see with the engine? What additions would you like to make?



Go!

There are a number of modifications/bug fixes that I have been hoping to have fixed

Bugs-
Client crashes after changing instances many times (I believe that it does not unload the models from memory for the different instances)
Client does not always close all the way leaving the player logged in and forcing the player to close from task manager (pretty rare) but sometimes still leaves player logged in until server reset (also rare)
Detaching the camera while the player was in first person mode permanently hides the player until camera is released to the player and player goes into third person
When close to steep terrain slope camera shoots upwards for a top down view instead of zooming into first person mode
Applying a scripted animation(nodetrack) to an object automatically resize the object to the default setting
Hiding all of the submeshes of an object in the worldeditor causes the object to be visible (Have a work around but annoying to have to label each time that it should be hidden)
Put in a delay to show frames after player clicks so that it doesnt automatically think the player clicked on something in a frame right under the cursor
Fix sockets so if the models size is changed the sockets are moved appropriately

Server
If a mob is following a player and the player pause the mob will momentarily point to the origin
Too many regions on top of each other causes significant slow down to trigger onEnter event (even if regions are in different instances)
When updating a quest every quest in the game is resent making the game stop (Have a fix but would need to update mars code for everyone)


Changes:
Client - Update the rendering engine to the latest version of Axiom
Client - Update the IronPython interpreter to the latest version
Client - Require/Package a more recent version of dotnet - not a huge concern
Client - 2D Texture animation doable in IronPython but can be slow on certain computers
Allow script access to the collision calculations


Add ons (Long term)
Change server from xz coordinates to xyz
Allow for some handling of physics on the server for pathfinding purposes
Allowing for multiple world manager servers
Allow for some more indepth physics
Allow for mounts

Those are the ones I can think of off the top of my head

I didnt incorporate updates to the toolset
World Editor
Add the ability to use the shift key to select items in the list instead of having to hold ctrl and selecting each object
Change so copy and pasting an object keeps the rotation/orientation of the previous object
Add better orientation controls...currently there is a missing axis
Create a focus command similar to what they have in maya where you can hit a button and view the object (current one does not work very well)
Add the ability to select a whole collection by selecting the collection and moving it all as one
Fix the error that hitting the delete button does not delete a terrain decal

It would also be nice to be able to export a scene from maya or max and have it output each model individually and create a world file

Fix an error in the terrain editor where changing the height sometimes goes crazy
Add the ability to move objects in the terrain editor (it would be nice to have all of the controls of the world editor ported over to the terrain editor)


Some sort of gui editor would be awesome
Similarly a quest editor would be nice but each team would probably need to either make their own or heavily modify it

Some great observations, thanks Delurin!

I agree with all of them, of course, but I agree even more with your points about:

- Copying an object in the World Editor doesn't copy the orientation, which is a real pain, especially for things like wall segements.

- I personally feel that the World Editor and the Terrain Editor needs to be merged. If, whilst adding objects in World Editor, you realise that a hill would look better over there, or that hill is too high, or the river is far too wide for my bridge model, you have to go back into Terrain Editor and change it. This causes disruptions in workflow, and is generally a pain! There should just be two tabs in the World Editor (terrain is part of the World, after all), and be able to switch between the two seamlessly. I'm assuming that it would be harder to completely merge them, but is a potential program to release in the future.#

I know the Islands of War team uses some form of central database for creating things like quests, factions, mobs, and things like that - I'll ask them about it when I'm on their website's chatroom next... It's just whether they'd be willing to basically hand over their hard work. Some persuasion is required, methinks!

Great so far, keep them coming if there's more!



Tristan

I remembered some of the GUI changes that would be nice
there are some tags that are not implemented like toplevel, resizable, movable
those would be nice to implement
A little harder might be to make every frame into a window or container so that if you have two frames overlapping each other they will not mix (currently there are only a few different levels for the frames to exist on)


Add an event to be triggered when the player hits the x button on the top right corner
similarly trigger an event when the player alt tabs/ move cursor outside of the screen (I believe that if you are rotating and you alt tab or move outside you will continue to spin until you hit an arrow key small detail but makes it look unprofessional)

It would also be nice to implement the OnKeyEntered for text boxes currently I dont believe it does anything so it cant override arrow keys,delete or backspace keys or trigger events when they are pressed

I've been looking at some posts on the old forums, and noticed that there were some features that were requested, but of course weren't implemented. From what I can see, the company were going to implement some of these, so I think it would be interesting to put them in.

Zoning, so that you don't have to create one massive world map, which from what I've read can create some serious performance issues.
Built-in support for FPS-type combat. So, you don't need to specifically target characters in order to shoot them.
A Character Creation program as part of the Multiverse Tools package - I talked about this in this post
A Terrain Editor minimap - in the future, I hope we'll be able to actually combine the World Editor and the Terrain Editor into one, but as a feature for the combined program, a minimap would be useful. If you're making large worlds, it's difficult to know where you are editing in relation to everything else.
Built-in Heightmap Editor - Similar to what Terragen uses. Some people like to model the initial stage of the terrain using heightmaps, so that they get a feel for the overall area.
Update the SpeedTree assets/plugin - from what I understand, the free SpeedTree assets we get as part of the package are out of date, and could cause performance issues. If we can find a way to update the plugin and the assets, that would be great (just from a performance perspective). If we can't update it, how about trying something like this: LTrees
Support for Japanese/Korean/Chinese Input - this is something I saw on the old forums, and is rather interesting. It would also attract Asian developers (maybe)
Support for 4D maps - this is an incredibly interesting idea, and is worth looking at. BlackHawke brought it up back in 2007 here

Can't think of anything else right now, but I'll add some of the ones we've already discussed to a section on the Wiki when I get a chance.

I would also recommend changing the server so that it checks the players instance before checking the players location for perception updates...As it stands it updates everything with in range that a player has come in range, it does not check if the player is in the same instance as the object it updates (this also applies to server regions). This will probably be a major concern for multiple instance worlds with lots of players running around sending messages.

So have any of these changes been looked at thus far I've just downloaded the repository and Im considering using Multiverse for my mmorpg. 2 things I'd like to know

- Problems that exist still that could cause significant problems i.e. lack of zoning or most importantly crashing issues.

- I am assuming the 10% Royalty etc still is in effect I mean I would hope so anyways considering what is offered here. Is it still mandatory for the financial side to be handled by your company also I believe I've seen that it's no problem for Projects to have there own launch screen not including the Multiverse world selector etc as that would also be important.

Aside from that Im very excited to compile Multiverse and take a look at everything aside from the face that it's basically a complete mmo which is great the tools mean alot to me and from testing the map editor to start I was quite happy with how it worked.

I believe that the 10% no longer applies. Currently you would be responsible for billing etc. We might be setting up a billing system but it probably wouldn't be for a while. You do not have to go to the multiverse world selector (although it currently requires recompiling and making your own installer but that will hopefully change soon)

Most of the crashing issues are taken care of there are only a few extremely isolated cases that I have run into but others have not, which obviously I am working on trying to fix.

For the zoning currently multiple worlds/instances are possible so you can still divide up your world but it would not be seemless. Alternatively you can make one giant terrain (2000 x 2000 km I believe) and only the parts that a player can see are sent to the player.

With regards to the Billing issue, as delurin said, the 10% fee no longer applies, and all billing will be handled by the studios from now on.

However, I have been looking into billing systems aimed at MMOs, and I've come up with a few good solutions, including an Open Source one, which would be good from our point of view. I'll keep you updated on that.

It would be nice to get zoning included in the code, as it will improve gameplay and performance. Delurin's suggestion of splitting them up would work well for games that have lots of big islands, so when you get to a particular distance from the short of one island, the opposite island island is loaded. You could then have a very low poly version of the island terrain to replace the original, to save on the FPS.
But if you don't have any islands, it'll be much more difficult to implement.
Again, this is something we've added to our milestone list as a feature to implement, but that will happen a bit further down the line, once we've sorted out the current issues, such as the Axiom engine and various bugs etc

Alright sounds good thanks for the very quick reply to be honest this forum seems a bit unpopulated right now but it's suprising considering the toolset your offering and especially that you chose a true opensource license in the MIT, I've posted in on an mmorpg maker forum I frequent every now and then hopefully some more people will give this a try. Only issue I've run into in the source is the SpeexWrapper project seems to be missing which in turn stops 3 or 4 things from being built.

Hi Treva,

Thanks for your interest, by the way!
(delurin - if you could look at the SpeedWrapper issue, that would be great)

This platform was originally closed source, and was doing very well, but something happened at the company that create it (Multiverse Inc.), and the platform was but on a very low-powered backburner.
Eventually, I got them to make it open source a few months ago. However, most of the developers who were active in using it originally have disappeared, as they became illustrated at the lack of updates - the last closed source updated released by the company was sometime in 2008, so everyone gave up and lost the inspiration they needed for their projects. Big shame really.

So my aim, (and now the aim of delurin and scfischer, the programmers in the Open Source team), is to get the Platform back up and running, and better than ever before.
I'm in the process of putting together a marketing campaign to get more studios/developers using the platform, but I'm trying to find the cheapest (or free) way to do that, whilst getting maximum coverage. It may take a long time, but I'd like to see it back to its former glory, with the possibility of being better than current paid MMO platforms on the market at the moment.



With the billing bit, the way that the project will create revenue is via the Marketplace that I'm currently working on - sellers will give us a small fee from each sale, which we'll put into hosting fees etc. We may also offer billing support, and run a small billing service ourselves (which I would head, depending on how much time it would take)

Alright well I wish you luck I'm hoping to find some success in getting things running and be able to use it for my project was really happy to see it was opensource under MIT, the GPL obviously scares everyone away because you can't make a commercial project and release your source it's too open to hacking.

Anyways the issue was Voice.dll wasn't built so I added Voice library to the solution (visual 2010 btw) however the reference SpeexWrapper didn't exist under lib SpeexWrapper I open the solution which is one of 3 files there and it says the project is missing so I guess there in lie's the problem the other issue's all stem from Voice.dll not being built

One thing I'd reccomend revive all the mmorpg development forums Multiverse tabs let them know it's open source under MIT and I'm sure you'll generate a good bit of interest. As soon as I get the chance to dive into this and take a look at it I would be happy to post a review on those sites as Ive for quite some time browsed most of these sites.

Edit - Ok it seems SpeexWrapperVS2010 project file didn't exist but the project file there was able to be converted up fine it's jsut a single file project so we may not have an issue here apart from a naming thing.

I havent had any trouble compiling the code. But I also use the makefile (it is a lot easier than trying to get Visual Studio to compile)

ah I've never really gotten into make it's all good now as I edit'd just 3 libraries weren't in the solution and 1 solution had to be remade the project wasn't linked right, works great now. Going to Install fedora and test out the server hoping for success there and we'll see if I can't connect and see how things look.

I believe that if you follow the make/make dependencies it will move all of the files into the appropriate folders

TrevaH,

I'm added Visual Studio 2010 projects and solutions for most everything in the SourceForge trunk. If you pull the most recent versions you should get the new projects. However, you may encounter an Exception when running under Visual Studio 2010. Below are the instructions to overcome.

They should compile fine. However, when running the apps in debug mode the developer may encounter an Managed Debugging Assistant (MDA) Exception, LoaderLock. To bypass this exception do the following;

1) Go to the Debug menu
2) Expand Managed Debugging Assistants
3) Scroll down and uncheck Loader Lock
4) Click Ok

Take Care,
scfischer7

TrevaH wrote:

Alright well I wish you luck I'm hoping to find some success in getting things running and be able to use it for my project was really happy to see it was opensource under MIT, the GPL obviously scares everyone away because you can't make a commercial project and release your source it's too open to hacking.

Anyways the issue was Voice.dll wasn't built so I added Voice library to the solution (visual 2010 btw) however the reference SpeexWrapper didn't exist under lib SpeexWrapper I open the solution which is one of 3 files there and it says the project is missing so I guess there in lie's the problem the other issue's all stem from Voice.dll not being built

One thing I'd reccomend revive all the mmorpg development forums Multiverse tabs let them know it's open source under MIT and I'm sure you'll generate a good bit of interest. As soon as I get the chance to dive into this and take a look at it I would be happy to post a review on those sites as Ive for quite some time browsed most of these sites.

Edit - Ok it seems SpeexWrapperVS2010 project file didn't exist but the project file there was able to be converted up fine it's jsut a single file project so we may not have an issue here apart from a naming thing.

Having had a test server set up and working over the summer, then reorganizing the team to come back to find multiverse was no more was a surprise. Finding out that it was moving to open source was actually a pleasant surprise.

However, with this moving to open source there are a few things I would like to see:
Migration - Since the server and the client will be updated periodically, there should be some utility to 1) determine if our current implementation can be migrated and 2) allow for us to include the client updates in our manifests to send down to the users. This is, of course, if we decide to upgrade to the new features.

Expanded Terrain Options - This one may be just wishing but I'm listing it just in case. I already use ProFantasy tools to create worlds and maps but cannot successfully get them rendered into multiverse. This has caused several early stalls in our progress. This can also be caused by our lack of experience. In the later case, expanded instructions on making the transformation would be useful. Like I said, I may just be hoping.

The Database - I do not even know if this one is relevant but will the database become more normalized or was the way it was implemented the best method for this MMO (and MMOs in general). I would have thought a more normalized transactional database for saving/storage would be used with a more transient in-memory database to handle transactions between saves. I have not looked at the database for a while but some data redesign may help integrate some of the tools mentioned in this forum, such as a quest tool or even help support the mob engines, etc...

Master Server - I did not read too far into the forums yet but since we now have to implement our own master servers, including this as part of a package to install may be a good thing, or at least general instructions on how to implement this to get everything working. Ignore this if it has already been addressed and I missed it.

Just some thoughts. As my team and I start to come back to our project, we would like to keep using Multiverse and from the looks of it, there may be some forward progress that will make that possible.

I am working on instructions on how to set up a master server..I know a few people have already made their own (including myself)
for the database, it could use some work right now very little is saved between server resets. For our server I made a custom data dump utility that dumps the quest statuses to a file and reloads it when the server restarts but it is more of a temporary fix than a permanent solution.

I know terrain is also a common problem point so it will probably be a high priority.

Next week, i will make a video tutorial with a full install on FreeBSD 9 (OS +game + master). I have all steps i made thanks to delurin and shane. Will be great to have this tutorial multilanguage.

Thank you delurin/neckro, that will be great.

Delurin,
I will try to take notes on the database as we explore our project, if I come up with anything better, I will pass it on.