This Torque seems nice, I guess. The repo is very weirdly structured for me, very hard to find the actual code. Seems they don't use a lot of dependencies, which is good i guess. Seems to implement their own window managers etc. for each platform, seems crazy to me but gets you the most control :)
Yeah it looks nice and all, but why exactly should we start porting features from it to Tundra? It seems to have a very different use, as in having a flexible gui editor to create projects in "click here and here to make a game" kind of way. Is this networked at all or a "make single player games" framework? I dont even know exactly what it is, just by reading the readme from github it still leaves me puzzled. Can you be more specific what would you like to see in Tundra from this? Can you point to code/sceenshots to get some kind of idea? I doubt there is any "porting" to be done, its a completely different beast without all of our dependencies/libraries/frameworks. Features/ideas can be copied (if they apply to the Tundra "idelogy") but they will probably need to be rewritten from scratch. I think we can however agree rex project should focus on documentation and examples (both c++ and javascript). Content creation tools can also be on the table, but as we have talked numerous times (in the issue tracker) Peter it might not be in the Tundra SDK scope to have fancy editors and ship certain kinds of 3D assets with it. Maybe we need to adjust that "purity goal", I don't know. I would like to see the assets at least in a different repo if this happens and move /bin/scenes and parts of /bin/media there too. I think many of the end user problems with Tundra come from that we are a very programmer oriented thing. You can't do much with Tundra if you don't know how to code or are willing to look at some headers. This can partly be solved with documentation, but at the end of the day if you want to make things move and pop (like a game), you need to write some code either with javascript and/or c++ code. Tundra is a more of a developer SDK with some examples to get you started, not a full "click a button on a editor to make a game" kind of thing, I guess that is a quite nice dream to have. We could make this kind of editor, maybe utilizing ogitor or just build our own, but we would need to move our focus out from Tundra "core". I suspect that this Torque was not made in a couple of months but there is some serious money and effort behind it before made open source. Btw. We have it on our todo list to make most of these things to Meshmoon. Nice(r) editors, easy "primitive" building, scripting templates and drag and drop content libraries. We need to make usage and building easier for our end users, I just think it might be more in Meshmoons scope that Tundra cores, hard to say. I know for a fact that if I implement these in Meshmoon I can make it just the way we like. If I'm doing it it Tundra core I need to ask for acceptance, deal with merge politics, go with cores UI style and in many cases dump down or limit the implementation so that it is generic enough and usable everywhere. *TL;DR +1 for documentation and modeling/scripting examples :)* Best regards, Jonne Nauha Adminotech developer On Wed, Jan 9, 2013 at 9:24 PM, Peter C. <[email protected]> wrote: > They open sourced the whole of Torque3D: > https://github.com/GarageGames/Torque3D > > It only runs on Windows for now, but they are pushing for Mac and Linux > support. They are also planning on open sourcing Torque2D. I think that > for inter-op it would be best to start by porting functionality Torque > has that Tundra does not to Tundra, and then figuring out ways to > convert content/scripts from Torque to Tundra. I was looking at Torque, > and the source code it self is really heavily interconnected; it would > take a complete overhaul to move it to a modern entity-component system. > I personally would rather just implement the missing functionality in > Tundra and make it easy to port from Torque to Tundra. > > On 01/09/2013 01:42 PM, Toni Alatalo wrote: > > On Wed, 2013-01-09 at 13:30 -0500, Peter C. wrote: > >> In regards to the game dev aspect: I'm sorry if that came off the wrong > >> way, it was just the google doc read a little more like "this is an > >> academic project, the focus is on professional use and not games, end of > >> story". Sorry if I misinterpreted that. Also, in regards to Ogitor, I > > Right, well it's me who has to be sorry for writing it too unclearly > > there I'm afraid. > > > > Actually the strategy we hope to drive now in the research cooperation > > with the university here etc. is that out of the academic work we would > > get concrete improvements to the tech which would really help e.g. game > > creation. > > > >> was thinking more build some sort of plug-in that would allow for using > >> Ogitor functionality in Tundra as an in-scene editor. Also, for torque, > > Right, that's what I was also wondering early on -- especially because > > also Ogitor uses Qt like Tundra does, might be possible to somehow hook > > Ogitor to be a Tundra plugin or something even (but not totally > > straightfoward, because Ogitor uses the Ogre API directly, whereas to > > integrate in Tundra nicely it'd need to be ported to the Tundra API > > which wraps Ogre). > > > >> I have to do some more research in to it, however the main two things > >> that come to mind are possibly implementing the ability to easily > >> transition from Torque to Tundra. The main problem I see with Torque is > >> it is based on the old multiple inheritance methods, and is a pain to > >> work with when you start needing custom functionality. There are > > Right - the entity-component model in Tundra has certainly proven to be > > a nice strong point, has served great for extensibility in practice, is > > easy to understand I think etc. > > > >> easy inter-op. Ideally it would be a collaboration between the RealXtend > >> association and GarageGames to create the best of both worlds, but I > >> doubt that would happen. If nothing else, my main concern would be the > >> need to implement similar DCC as torque's editor in to RealXtend, along > > Interesting - it might be a far call, but it is also good to brainstorm > > with an open mind sometimes. > > > > BTW didn't they open source something recently, was it torque3d runtime > > or what? > > > >> I think that the primary things that RealXtend could benefit from would > >> be better documentation of how to develop for it beyond basic scripting > > Agreed. As you noted, Meshmoon docs helped already, but certainly much > > is still urgently needed. > > > > I think we need tech dev too but certainly creating those docs must be > > somehow organized finally. > > > >> out to places like Ogre, and having an Ogre wiki page and link to > >> realxtend in the Ogre wiki for 3rd party projects using Ogre, would > > *nod* -- I think the time is certainly ripe to submit Tundra as a > > candidate featured project for the Ogre site. Feel free to do it, anyone > > basically, but I think I must if no one else beats me to it (should not > > be hard..) > > > > Thanks again for the good points and welcome reminders (I wasn't > > recalling Ogitor and Torque too clearly at all, have been quite occupied > > with many other things recently too). > > > > ~Toni > > > >> On 01/09/2013 01:05 PM, Toni Alatalo wrote: > >>> On Wed, 2013-01-09 at 10:22 -0500, Peter C wrote: > >>>> Just a few suggestions / things to note. > >>> Thanks - quick comments: > >>>> 1. Even if the primary intent is not games, it would be wise to > >>>> develop with the same performance mindset of a game engine. Long term > >>>> viability as a platform requires keeping up with tech. > >>> We certainly keep that up -- both Ludocraft and Playsign are games > >>> companies, the requirements from us (e.g. Ludo's old creation tool doc, > >>> the mobile tundra plan etc) are largely from games dev perspective). > >>> > >>> Sometimes it is more difficult for us to get a good understanding of > the > >>> *other* requirements, other apps besides games :) > >>> > >>> If you are referring to the remark in the google doc, it tries to say > >>> that reX is essentially about *networked multiplayer games* (or other > >>> multiuser apps) out of the box, is inherently networked, whereas e.g. > >>> Unity3D was originally for single player games (though has many mature > >>> ways to do networking nowadays with 3rd party addons and has basics > >>> builtin too). > >>> > >>>> 2. It might be wise to try porting ogitor to run inside realXtend. If > >>>> nothing else make an exporter from ogitor to tundra. It may also be > >>>> wise to look at how torque has their dcc set up. > >>> Yes we looked at Ogitor back in the early days when considering > options, > >>> I was repeatedly showing it to the guys in sprint meetings etc .. and > >>> read some of their code when considering editing things in Naali/Tundra > >>> etc. > >>> > >>> Both Tundra and Ogitor support the old simple Ogre dotscene format so > >>> they might be interoperable already, i.e. you might be able to import > >>> Ogitor authored scenes to Tundra. When I tried dotscene things in > Ogitor > >>> 0.3 some 2-3(?) years ago, though, got only some crash then, dunno > about > >>> the status of dotscene vs. their own format there -- certainly worth a > >>> new look, it always has seemed like a good project. > >>>> 3. It may be a good idea to look at porting torques functionality to > >>>> tundra. I was recently doing research on torque, and they have a > >>>> massive community, but their architecture is really bad. Don't count > >>>> out game developers as a possible audience. If you were to outreach to > >>>> game developers, and they were to get involved in improving the game > >>>> engine side of things, it would benefit the academic side too. Games > >>>> tend to be one of the biggest driving forces of technology. > >>> Yes, I actually tested it for some reason last year a little and > thought > >>> there were nice things, certainly also worth more study. > >>> > >>> Scripting / Logic authoring has been my pet peeve for long, and still > >>> kind of sold with the Scratch design for very simple things at least > >>> (for kids and non-programmers), and both Google Blockly and Waterbear > >>> would seem to allow it nicely for Tundra-JS. > >>> > >>> Do you have some things in specific in mind about Torque, does it cover > >>> everything a bit like Unity, dealing with assets etc? I'm recalling the > >>> game maker things with events and states and something for logic etc, > >>> hopefully have time for a study soon enough (or someone else has). > >>> > >>> ~Toni > >>>> Cheers, > >>>> Peter > >>>> > >>>> [email protected] wrote: > >>>> Hi, > >>>> > >>>> we had a semiofficial(*) realXtend association board meeting > >>>> yesterday, mostly to discuss and organize further planning on > >>>> development roadmap for the new year. > >>>> > >>>> My full notes are on-line, main point summarized here: We > >>>> decided to plan work on two fronts, creation tools and > >>>> pipelines coming as a new primary focus. The other area is the > >>>> tech platform & engines topic which was already worked on a > >>>> lot last already (the realXtend roadmap doc from last spring > >>>> discusses the three areas there, i.e. current Tundra, browser > >>>> based clients and the Mobile Tundra unified light client > >>>> idea). > >>>> > >>>> For the creation tools and pipeline we agreed to gather > >>>> wishes, requirements and development proposals and meet again > >>>> on Thursday next week (17th) to put together a plan. > >>>> Ludocraft made one report on this already ages ago, they’ll > >>>> check if parts of it are still valid. Francois will talk with > >>>> Matteo and Francois from Spinningwire and ENER labs. > >>>> Adminotech has some concrete needs, I think largely coming > >>>> from VW use in education. I think Evocons at least can tell > >>>> what they need in their work with the building industry. > >>>> > >>>> You, anyone, can also use this chance to inform the planning: > >>>> what would you need to be able to create applications, worlds > >>>> or whatever with realXtend better, or is that even a > >>>> bottleneck for you now? Even vague ideas are welcome but the > >>>> more concrete a plan the better of course. > >>>> > >>>> Some things discussed in the meeting: more example assets for > >>>> e.g. use of different materials / options of the SuperShader, > >>>> creating a new shader library. Better scene/ec editor with > >>>> grouping etc. A question: is tighter Blender integration, for > >>>> example with live material preview with a Tundra window as > >>>> demonstrated by blender2ogre, a good way to author or is > >>>> something else better? > >>>> > >>>> The full notes with some additional points are in > >>>> > https://docs.google.com/document/d/1IqS7Z9WUy_7jt753oSnt3HE0ISQXhT4zstP71A_6FKY/edit(not > too structured, sorry). > >>>> > >>>> I think we can use this mailing list / google group to gather > >>>> ideas and discuss, but am also interested in more structured > >>>> ways. For example getsatisfaction.com has seemed nice for > >>>> working on feature requests, I think I saw Kitely using that > >>>> long ago and tested creating a realXtend account there too, > >>>> but I don't have any real experience on using that or any > >>>> other similar service. Github issues serve well for actual > >>>> todo items and feature wishes too but I don’t think it suits > >>>> this kind of requirements elicitation. Am open for > >>>> suggestions, either here or privately. > >>>> > >>>> Finally, I’d like to explain a bit the rationale for the focus > >>>> on creation tools as how the common interest focused there > >>>> surprised me. I have earlier thought that there is a big > >>>> divide between a)professional creators and b) supporting easy > >>>> end user content creation. Basic realXtend offering, e.g. the > >>>> Tundra SDK and the little WebGL and Flash clients, target > >>>> professional creators -- people who are comfortable with > >>>> normal 3d modeling and programming etc. More Second Life or > >>>> Facebook style end user creation are implemented in custom > >>>> applications, for example the TOY content tools which are a > >>>> now a part of the Meshmoon offering, Cyberslide where you can > >>>> just create a scene from your Powerpoint slides, or > >>>> Ludocraft’s sandbox. But yesterday the common understanding > >>>> was that there are many things that we could do to help both > >>>> professional creators and services with user created content. > >>>> Ease of creation is of utmost importance in professional use > >>>> as well as it of course affects both the quality and > >>>> especially the cost duration of projects. Also we figured that > >>>> work on creation tools is relevant in any case, no matter > >>>> whether we end up using Ogre, some other native engine, or > >>>> WebGL more in the future. > >>>> > >>>> so here’s a starting point for the year! > >>>> ~Toni > >>>> > >>>> (*) not everyone in the board could participate yesterday, so > >>>> we postponed some administrative bureaucracy for a later > >>>> meeting and focused on the dev planning work > >>>> > > > > -- > http://groups.google.com/group/realxtend > http://www.realxtend.org > -- http://groups.google.com/group/realxtend http://www.realxtend.org
