EVE's move to Python 3 could unlock more than performance improvements. What new features should CCP build?
I've been thinking about CCP's announcement that EVE is finally moving from Python 2.7 to Python 3.
At first, I assumed this was mainly a technical upgrade. Better performance, easier maintenance, fewer headaches for developers. All good things, obviously, but nothing particularly exciting from a player's perspective.
Then I started wondering about something else.
>What if the biggest benefit isn't making the EVE we already have run better, but making entirely new gameplay features more practical to develop?
CCP is migrating approximately 2.4 million lines of Python code, some of which dates back to the early years of the game. That's a massive undertaking, especially for a game that has been running continuously for over two decades.
Python 3 brings access to a much larger ecosystem of modern libraries, profiling tools, optimization techniques, and AI frameworks.
It doesn't automatically unlock new gameplay mechanics, of course. Most of the ideas below were technically possible under Python 2, and some could already be built using external services.
But modernizing the underlying technology could make certain ideas easier to develop, maintain, and expand.
Here are a few features I'd love CCP to consider.
Six possibilities for New Eden
- Adaptive NPC factions
- A universe that responds to player activity
- Corporation intelligence and fleet planning
- Industry and logistics planning
- NPC encounters with tactical objectives
- Fleet battle and simulation improvements
These are feature proposals. Python 3 could ease development; it does not make them automatic.
1. NPC factions that actually adapt to players
Imagine Guristas, Sansha, or Blood Raiders behaving less like predictable mission NPCs and more like actual factions operating in New Eden.
Suppose players repeatedly farm a particular system.
Instead of spawning the same NPCs indefinitely, the faction gradually changes its behavior.
Maybe it starts deploying different ship compositions, reinforces nearby systems, sends patrols along commonly used routes, or launches counterattacks against players disrupting its operations.
You could even have regional pirate campaigns where player activity influences what happens next.
For example, if players destroy enough pirate supply convoys, the faction's presence temporarily weakens. Ignore those convoys, and the pirates become better equipped.
That could create a much more dynamic PvE environment, particularly in low-sec and null-sec.
Where Python 3 could help: Modern decision-making frameworks, data-processing libraries, and simulation tools could make it easier to build, test, and tune these systems. The actual NPC logic would still need to be designed by CCP.
2. A living universe that responds to player activity
This is probably my favorite idea.
EVE already has an incredibly interesting player-driven universe, but many environmental and PvE systems still operate according to relatively fixed rules.
What if the universe reacted more noticeably to what players were doing?
Imagine a constellation experiencing increased pirate activity because trade routes have been neglected, or resource shortages triggering temporary industrial opportunities.
Maybe large-scale mining activity affects local NPC operations. Or repeated attacks on a faction's infrastructure lead to stronger defensive responses.
The important part is that players could influence these developments through their actions.
I'm not suggesting CCP should simulate every NPC or economic transaction individually. That would be unnecessarily expensive.
A background simulation could periodically evaluate regional activity and generate events based on the results.
Where Python 3 could help: Modern data-analysis and simulation libraries could support more sophisticated background calculations. These could potentially run outside the time-critical solar-system simulation, reducing their impact on combat performance.
This seems like an interesting opportunity to make New Eden feel more alive without turning EVE into a completely different game.
3. Better in-game intelligence and fleet planning
EVE players already build impressive external tools for intelligence gathering, logistics, market analysis, and fleet coordination.
But what if some of those capabilities were available directly in the game?
Imagine a corporation intelligence interface that helps summarize recent activity across your region using information your corporation is actually entitled to access.
It could identify changes in local activity, summarize corporation-owned scouting reports, display trends from historical kill reports, or help fleet commanders review previous engagements.
For fleet operations, imagine better tools for organizing doctrines, checking readiness, and estimating logistical requirements before an operation.
Not something that automatically plays the game for you. Just better access to information and planning tools.
Where Python 3 could help: Modern analytics frameworks and improved integration with data-processing services could reduce development complexity.
And because EVE depends so heavily on intelligence and information advantages, CCP would have to be careful not to expose hidden information or remove the importance of scouting.
4. An actual industry and logistics planning system
Anyone who has spent serious time managing industry knows how quickly things become complicated.
Blueprints, materials, production chains, hauling, market prices, structure bonuses, and corporation assets.
A lot of us end up using spreadsheets or third-party applications because the in-game tools don't cover everything.
Imagine having a proper production planning interface.
You select what you want to manufacture, and the game helps calculate required materials, identifies shortages using your accessible assets, and estimates production requirements.
It could also support corporation-wide production planning, task assignments, and logistics coordination.
I'm not suggesting automated trading or hauling. Players should still make those decisions and take the risks.
But spending less time manually organizing information could make industry much more accessible.
Where Python 3 could help: Modern optimization and data-processing libraries would be useful for planning calculations, especially when dealing with large production chains.
This is one of the ideas that could probably be implemented without major changes to EVE's core combat simulation.
5. More interesting NPC fleets and large-scale encounters
Another possibility is improving the complexity of encounters themselves.
Imagine an NPC fleet that has a tactical objective rather than simply trying to destroy your ship.
Maybe an enemy logistics vessel tries to escape while escorts delay you. Another group attempts to reinforce a structure. A third responds differently depending on whether players bring battleships, capitals, or smaller ships.
You could have encounters with several possible outcomes instead of simply clearing waves of NPCs.
Some of this already exists in EVE to varying degrees, but I think there's room to take it much further.
Where Python 3 could help: Better profiling tools and modern simulation libraries could help CCP develop and evaluate more complex NPC decision-making. CPU-intensive calculations might also be offloaded to suitable services or workers where the gameplay architecture allows it.
Again, this would require actual gameplay development. Python 3 alone doesn't give NPCs better intelligence.
6. More ambitious fleet battles and server-side simulation
I couldn't write about this migration without mentioning TiDi.
EVE's reliance on Stackless Python allows it to manage enormous numbers of concurrent activities, but concurrency isn't the same as parallel execution across multiple CPU cores.
Modern Python provides better performance tooling, and newer versions have optional free-threaded builds.
That doesn't mean CCP can simply enable multithreading and eliminate TiDi. Keeping thousands of ships synchronized in the same simulation is a much harder problem.
But perhaps this migration creates an opportunity to investigate which calculations could safely be separated from the main simulation.
Could independent background workloads be processed elsewhere? Could certain expensive operations be optimized or moved to specialized components?
Even if the basic architecture remains unchanged, better profiling and optimization might eventually allow CCP to squeeze more performance out of existing systems.
For players, that could mean more responsive modules during heavy fights or a higher load threshold before severe TiDi begins.
Probably one of the most difficult ideas on this list, but potentially one of the most valuable.
The bigger opportunity
What interests me most is that Python 3 isn't just about a faster interpreter.
It's about removing years of accumulated technical limitations.
New libraries, better development tools, modern testing frameworks, and easier integration with other technologies could make features more affordable to develop and maintain.
And perhaps that's where the real benefit lies.
Not in suddenly doubling performance or making everything run perfectly, but in allowing CCP to explore ideas that previously weren't worth the engineering effort.
Of course, none of these features are guaranteed, and Python 3 isn't a prerequisite for most of them. Game design, development resources, and server architecture still matter far more than the programming language alone.
Still, I think this migration could be an opportunity to look beyond maintaining the existing game.
If CCP is modernizing the foundations of EVE, why not use that opportunity to make New Eden more dynamic, intelligent, and interesting?
What would you want CCP to build?
If you could suggest one feature that takes advantage of a modernized Python ecosystem, what would it be?
Personally, I'd love to see a living universe where NPC factions adapt to player activity and where PvE has consequences beyond individual missions.
But I'm equally interested in what could be done with fleet battles, industry, and corporation tools.
I'd especially like to hear from developers familiar with Python or EVE's architecture about which of these ideas are technically realistic, and which ones would require a complete rethink of existing systems.
Sources
CCP's Python 3 announcement:
The move to Python 3 begins — CCP
Python documentation on free-threading:
Python support for free threading — official documentation
Any accompanying AI-generated images are concept illustrations, not announced or upcoming EVE features.