Ancient Spaces Blog

It's been a lot of fun and a lot of hard work, but the editor and client are now working well, and as such, I thought I'd post up some interesting pics of some cool particles and a scene with some people. Enjoy :) Glen

We needed an avatar for the project so I decided to take on the task to create one. My first idea was to create one from scratch. I created a half modeled human at the end of June, which I was planning to use. This would require me to finish modeling the head and texturing the beast. I would also need to change the shape of the model for it didn't look quite right. Another solution to our avatar issue was to modify our beloved ninja model, which I ended up doing.

After several meshing changes (like adding a 'skirt') and a texture upgrade we have our first avatar from the original ninja model. One more thing I hope to do is to change the models face texture, possibly put one of the Ancient Spaces team face on the model.

Oh yeah, I need to add eyes also. Just noticed that.

Note that he is an Egyptian avatar in Machu Picchu.

Serialization has finally come together! Further work on the data model refactoring can wait for a future release. What we have now works reasonably well and meets the initial projected requirements for the project. It was a tremendous relief to see it just working! Whoot!

The Editor:
Having users be able to call up HTML content from selected objects allows them to quickly access information on the objects they are currently editing and manipulating.

The editor is a wxWidgets application and therefore has access to the wxHtmlWindow. This allows the loading and viewing of HTML pages, in a window inside the project. If this were to be implemented, I would suggest the pages exist on the local drive, and not be accessed via http protocol. The wxHtmlWindow is not a full blown web browser, so viewing custom HTML pages would guarantee it would work.

However, the editor is a windowed application and it would be very easy for a user to switch between and a web browser. There for a button that send a URL to a web browser could easily be preferable. This would allow linking to actual online information, such as the wiki, using an interface which the user is much more familiar with (IE, Firefox, etc.).

Either way, adding web pages, URLs, and web browsers, add some implications for the game engine. Information should be specified in a objects script (so every instance of Totem Poll A provides the same information), and this level of scripting is still a ways away.

The Game Client:
In order for the game client will required two important features to allow object specific content. First, there must be some form of interaction with the object besides basic collision. Second, there must be the required scripting API called for something to happen upon this interaction. Once these two components exits a whole world of possibilities open up.

A Well Organized Web Page:
By far the most piratical approach to linking online content with the current version of Ancient Spaces would be for an academic to provide web content in a well organized and highly graphical manner which reflects their respective Ancient Spaces content. This could include screen shots from the game or editor. Students would then be easily able to locate information on the site associated with a specific object. The information on the site can also give guides to where to find the object in the 3D world.

Loading the Client from other Sources:
It may be possible for command line arguments to be passed to future versions of the client, so that the user starts at a specific location, viewing a specific object. For example a user could open a prepared version of the client and be jumped right to a specific totem pole. However, much of the strength of the game client is that is immerses the user in an entire world, and a “jump to object” feature will cause the location and positioning of an object to loose significance.

Simon,
Second Additional Assistant Interface Director

I have been working on the client side of things for a while now, and I finally have some pretty pictures to put up. Currently I am working on making collision detection work, I am using the OgreOpcode library (a wrapper around OpCode which is used with the well known ODE) to try and program in some collision detection for the client. A lot more work will be needed to make collsion actions have the proper effects, but this a certainly a great start

I have set up a visualization for the collision triangles so you can kind of see what is going on and I am only using one object here (an early version of Matt's Machu Pichu, as it is quite a large model and very suitable for this test).
Note: There are no lights to make it pretty, this is JUST A TEST

!Also, please note these images are from the debug version which is why the frame rate shown is so low, the last image shows the release without drawing the collision stuff and the framerate there is quite high (200% higher than DEBUG). Considering the large model being broken down for collision detection it performs very well :)

A quick shot of the GUI while in play

A shot of the Release version

The title tells the whole story and this happy moment.

Yes, Machu Picchu is done... for now... For which, both the Three-Temple Group and the Intiwatana Group are done and exported. That means no more rocks, creases and walls. Check out the wiki for more information.

Uhhh... I've tried to import the Temple of Jerusalem model into 3DS. It worked, BUT, it took hell of a long time... And the most peculiar problem is that, I cannot export it as a 3DS file. The error message indicates the model has over 64K faces.

I tried to re-export as a OBJ file. This time it worked! the file size is only half of its original, but it will load in Blender.

Here's a render of it in MAX.

Well the Nisga village is just about done. I made this nice render of all the structures I've made all placed beside each other to give a sense of scale and context. All that's left now is exporting the rest of these to the editor (some are already exported, like the trees in the background and the construction materials in the front).

Yay, progress!

For the last two weeks I have been working on completing Machu Picchu. The outline model was competed several weeks ago and was approved by Marvin, our academic for Machu Picchu. I posted about it on the blog so you can check back for further details. Since then I have been adding detail to the model and creating/applying textures to it. Today is also our expected due date for Machu Picchu so I'm quite excited that it is ALMOST fully complete. I've supplied some screen shots of the model in action in the editor, enjoy!

The work on serialization uncovered some issues with our previous design which needed to be refactored. So very painfully, Simon and I have been struggling with refactoring the engine so that there is no transfer of ownership with any Scene Elements. That turned out to have major ramifications on the way various other parts of the system worked with elements, so pages and pages of code had to be ripped up and basically rewritten. Although we encountered a few new problems and probably introduced a few new bugs, we crushed a whole slew of bugs during this process. A particularly conspicous one was that there remained residual light when a light element was deleted. It turned out to be another one of those subtle one-liners which take forever to find and no time to fix.

Following this, Simon recommended a one-line addition to the code which implements my first new functionality in a solid week: File -> New. :D

With the initial projects we started on (Machu Pichu and Nisga) coming to a finish, I've been working on tweaking some of the houses that I've worked on along with modelling more of the plants. I've added some more details to some of the structures.

For example, the Ganee's now has nicely curved strings for the fish to hang on, as shown below:

Also, here's a new and improved smoke/summer house, with details like a slot to bar the door and ties that hold some of the logs together:

And here's an example of a wild crabapple (Malus Fusca) branch in its fruit-bearing stage:

Since part of the emphasis for the exploration of the Nisga scenes created in ancient spaces will be on the plants and trees that were present, the plants are all explicitly modeled, rather than being billboards as is typical in most games. The idea is that someone will be able to tell the different plant types from one another and identify them for one another.

The VCE team working on Machu Picchu will be switching roles by the end of this week as the two Machu Picchu sites are projected to be complete. So far, the Three-Temple Group are complete and the Intiwatana Group is close to finish. For the rest of the mountain that Matt's working on, most of the textures have been mapped.

I have been working on the textures for Intiwatana landscape several days now. Most of the textures for the walls are done. However, there's a lot of edges/ledges need to be connected. In addition, many faces cannot be seen in photos so I am going to use another textures to map on.

Making the textures and mapping them will take the majority of the time. Marvin should be coming in daily to check on our progress and make useful suggestions. Hopefully his suggestions will speed up our progress. When renders are available, check them on the wiki.

So the last few work days seem to have resulted in a lot of cool thing getting accomplished.

Cool thing number A: Orthographic View
Orthographic tool is a very useful tool for arranging items in 3D space. In orthographic view instead of viewing objects in perspective, you see a flat projection on a plane. The effect of this is that you see everything with absolutely no depth perception. The strange thing about OGRE's orthographic view mode, is that the orthographic projection is projected onto the near clipping plane, which means in orthographic view mode you change your zoom by changing you clipping distance.

Cool thing number B: Particles
Particle Systems let you do lots of fun things such as rain, fire, smoke, and water fountains. Check out Glen's post for a picture of a smokey longhouse.

Cool thing number C: Global Scene Properties
Global scene properties are properties that apply to the entire scene instead of a single object. Properties such as ambient lighting or terrain information are included in the global properties. As of right now only the ambient light color, and background color are changeable, but soon skybox and terrain options will be included.

That is all for now,
-Simon
Orthographic Grandmaster