I will succeed because I shall develop in series of incremental steps, building up complexity as I gain experience and fame!
Since each stage shall result in a complete, playable game, there is no such thing as failing to reach the goal.
A simple Asteroids clone to test the basic mechanisms of my engine
A class-based command shooter (multiplayer only) based around massive destruction and construction of cube-voxel terrain
Since such game uses almost no resources, I can make it while keeping resource management very basic -- and let it evolve naturally, instead of my prior folly of building spherical sand castles in a vacuum.
It being multiplayer, the game serves as a test bed for my multiplayer architecture
Kid friendly, the player avatars are simplistic knit figures that unravel on frag. May be made using Quake2 Modeler, with very crude animation.
Animation to be upgraded after Blockservice tools are ready
Adult (18+) version of Blockmittens. Identical gameplay but avatars are fairies with jiggle physics, clothing damage and so on
Test bed for my modeling/animation tool.
Still very few resources.
A DooM engine build from scratch, only supporting Brutal DooM.
My job made immensely simpler (and thus, possible at all) because I don't need to support classic modes nor the bazilion of settings the source ports are burdened with. No, I will only implement features required specifically to run Brutal DooM.
Using dirty hacks to lift common source port limitations, with renderer fully optimized for modern GPUs (and by modern I mean GeForce 6xxx, Radeon 2600 XT, Raspberry Pi and other mighty midgets) - GZDoom can support Vulkan all it likes, there are fundamental limitations that necessitate on-the-fly portal-like processing of map geometry that put a really big stick in the wheels of scaling.
This clever Cheb knows a way around, though.
Two goals: allow running Extermination Day on a toaster and allow playing Brutal DooM with humongous maps that have humongous number of monsters, on any fossil of an office PC from the mid-2000s.
Necessitates implementing ACS Script from scratch. I plan pre-processing it to Pascal and compiling it into native code (so any classes defined in ACS, DECORATE and so on become Chentrah classes).
Will be the true test of my multiplayer architecture because I plan nothing like a proper client-server architecture - so I have to uncouple the GUI stuff from physics and still keep my creation Zandronum-compatible. I'll elaborate more when it takes off.
I also plan fixing some common deficiencies: for example, monsters won't keep firing into an obstacle as Arachnotrons are prone to. They will try shifting their aim up or down or even move to find a better position. All monsters will be conscious of their projectile's radius and correct their aim vertically not to hit nearest ledges. Revenant missiles won't be veering into edges dumbly. Most monsters will act smarter when fighting vertically: they won't move closer to a cliff their opponent is on and some would step over cliff edge to shoot down successfully. Pinkies will get the most dreaded power of all: pathfinding.
The real reason to make a DooM port: overcoming the limitations while keeping toolchain mostly the same.
Adds unique UDMF extensions (anchor things) allowing treating a map as a set of templates
Adds the ability to combine templates like 3d brushes, thus making 3d floors obsolete and unleashing creative control.
Adds procedural map generator able to make both template-based cities and organic terrain.
Better AI for monsters.
An overmind AI to direct monsters and make ambushes.
The development above used to generate a Diablo-II-like campaign based on various megawads by other people, with action-RPG elements added.
A real test for the procedural generator.
A real test for bringing varouus authors together because there are never enough templates.
Adds auto-generated story line, from Rogue Encampment on Barsoom, through ruined Loot Gala city on Gaea, to bloody and fiery wastelands of Nether realms.
Requires full world generation working in perfect harmony.
Combines the previous works with a destructible voxel open world
Adds the editor, a prototype of the future base management interface.
Procedural generator is R&D for future AI base management.
Generates a full uninhabited world with water, floods, lava and so on.
Adds "Dungeon Master" mode where one player can build a hellish fortress Dungeon Keeper style then invite people to test its defenses.
The necessary steps in-between.
Includes writing a novel to research my role-playing system from storytelling point of view.
The fully realized 4X game
Only uses custom assets (the amount of work required is enormous - should have a team of artists at this point or be dead from old age) with lots of custom animations. Full support for gore/mutilation. Full support for lots of clothing/armor. Full range of human NPCs.
Should already have my ultimate modeling tool to create the one supermodel to rule them all (where all genders/races/body shapes are just mutators). Non-skeletal physics-based animation (bulging muscles, jiggly beer bellies and more).
Assets for enough animal types to make the world alive: from bears to sheep.
Generator upgraded to fully realized worlds populated by humans, with society in artificial medieval stasis. Government, trade, roads, warring barons and AIs to control it all.
Introduce subtrefuge and stealth.
Introduce non-magical crafting & skills.