Compiler

[Patrice Terrier]

I thought I would be able to stir them up and talk them into launching a campaign to re-implement a professional grade 64-bit "powerbasic" within a very short time frame based on a team of highly skilled BASIC compiler developers who know the ins and outs of their profession. And I feel I'm failing to do that as noone really seems to be ready or willing to assume the accompanying responsibility or work as a team.

My eyes bleed as I'm seeing people and businesses creep in all directions like cockroaches from the rotting corpse of what used to be an industry standard program development tool. It was TurboBasic that aroused my interest in BASIC and general compiler development long, long time ago. I simply cannot allow myself to go not being absolutely sure I've done all I could to prevent its oblivion. Call me a romanticist, or idealist, or childish, or foolish, whatever.

Noone of the indie BASIC developers still alive and taken separately is able to deliver, within the next 9 months before it becomes obvious Drake Software's promises have been nothing but one big vaporware bubble, a competitive product that took Bob Zale and his crew decades to develop. We could make it happen as a team. But alas, we never will.

_______________________________________________

You know, Patrice, my heart is with OpenGL and 3D graphics, and among the three most exciting pet projects I had in the last decade, two were based on OpenGL -- ObjReader+Objector taken together included. Now that I'm seeing the compiler component failing to level the balance, I'm coming home.

No, I let the ObjReader fixes rest till I clear up the matters with BASIC just to be able to concentrate on one goal at a time, which is my usual behavioral pattern under similar conditions. (I'm easy to serialize but very difficult to parallelize )

I think I will be able to come up with the texture load fixes and texture matrix manipulation scheme some time this week.

The time to create a PB's 64-bit has passed, because those realy needing it have been forced to moved on, for long ...

We have to feed our existing pet to let it get better and better, and turning OBJ file into (.OBC or .COB) compiled binary is high on my list, to speed up the loading of HP models.

All the existing mobj global parameters bust be turned into a structure type, that could be easily loaded/saved into binary format.
I could do that work if that doesn't interfer with your own plans?

I could also try to create a special decimated shadow mesh for test purpose if you need one, that could be handled with a new #shadow meta.


Added
I have attached a Zonda decimated shadow silhouette, for test purpose


UPD
I made a few changes to the code, to remove unused Mobj_xxxx procedures, that are a survival of the time Mobj was a DLL.
I made also a small change to the reader skin component rotate.png.
My friend, let me know when you are back, and ready to get these small changes.

[Michael Lobko-Lobanovsky]

Patrice,

Thanks for the mesh! It is still gorgeous but wa-a-a-a-ay too good for the model's invisible LOD to be used for rendering as the stencil in a true stencil shadow pass.

Take this mesh as the original and try to decimate it further down probably as much as you decimated its high poly parent. I think the resultant couple thousand polies will be enough to cast a veritable shadow on the surrounding geometry.

Does your sending me this mesh mean you don't want to experiment any further with your quasi alpha shadow's displacement relative to the model and would rather try true stencil shadowing solutions?

"We have to feed our existing pet to let it get better and better, and 1) turning OBJ file into (.OBC or .COB) compiled binary is high on my list, to speed up the loading of HP models.

2) All the existing mobj global parameters bust be turned into a structure type, that could be easily loaded/saved into binary format.

3) I could also try to create a special decimated shadow mesh for test purpose if you need one, that could be handled with a new #shadow meta.

I could do that work if that doesn't interfer with your own plans?"

(MLL: reformatted by me)

By all means, your every effort will be very much appreciated, my friend!

Now let's leave this thread be. Please open a new W.I.P. thread called something like "Pushing the Limits" or "Beyond the Horizon" LOL and post your extra decimated very low-LOD Zonda mesh and mods there. We'll maintain this new thread for our communication regarding ObjReader's future.

I'll merge the mods with my current sources and will respond with my ObjReaderFBO setup I'm planning to use as a sandbox to develop the projected ObjReader v3.0. I suggest we use it from now on for our collective experimentation.

Let's leave ObjReader v2.5 as ObjReader Stable till we're ready to come up with the new, better and more advanced solution!