A game can be kept alive by becoming less like the game that shipped.
Preservation isn’t supposed to work that way.
The usual picture is tidier. The original disc is found. The original files are copied. The original machine is made to run them again, and the work survives because somebody refused to alter it.
But a game is rarely only its files.
It’s an executable, an engine, an activation path, a server, an agreement with the machine underneath.
When one of those stops answering, a community patch may be the only version anyone can still play.
Keep the executable. Keep the old assets. Keep replacing the parts that no longer answer.
That isn’t quite preservation.
It isn’t quite replacement either.
It’s a fork with better weather.
The dependency comes first.
A mod works inside somebody else’s engine and somebody else’s economy.
Bohemia Interactive’s official tools let Dean Hall build the DayZ mod on ARMA 2.
Valve’s Half-Life SDK made Counter-Strike possible.
Those are invitations with boundaries. Not transfers of ownership.
The unofficial route has a different fragility.
Extraction and injection tools can let a mod touch a game with no editor at all.
They exist by tolerance rather than by design.
An official editor says the door has been left open.
A reverse-engineered tool says the lock has been understood.
A takedown notice says the building still has an owner.
That is the first cost of freedom.
A modder can repair what a publisher stopped repairing, add what it never cared to add, and turn an old system toward a new audience. That’s real.
The modder also works through rights that stay somebody else’s.
An end-user agreement can forbid selling the work, or soliciting donations for it.
An SDK sets one limit. A storefront agreement sets another. The power to delete a file from its host sets the last one.
A workshop upload can go live within minutes of an automated check.
It can be gone the same afternoon if the rights holder decides so.
The mod travels freely right up until the road’s owner notices it.
That permission also decides what “standalone” means.
SureAI’s Enderal replaced Skyrim’s world, quests and voice acting almost entirely, over roughly five years.
It still required the Skyrim executable and engine.
A total conversion can replace the furniture, the inhabitants and the weather.
It remains a tenant of the original foundation.
Only a renegotiation changes that.
Dean Hall joined Bohemia to make DayZ a standalone title.
Garry’s Mod reached commercial release through an independent deal that split the proceeds with Valve.
Those aren’t milestones on some natural path from mod to game.
They’re the moments a rights holder agrees that somebody else’s unpaid proof of value can be converted into licensed work.
The proof usually stays unpaid, and that economy has a strength preservation needs.
It leaves people willing to learn obsolete tools, document unglamorous failures, and keep an abandoned dependency legible long after its commercial value became hard to see.
DayZ pushed a three-year-old ARMA 2 back up the online PC sales charts.
The base game got a second life because someone built a different game inside it.
It also has a weakness preservation can’t ignore.
Whoever makes an old game practical again may change its balance, its interface, its network assumptions, its art, or its rules of access.
A community patch can preserve the ability to play while making the original experience harder to encounter.
Continued life isn’t neutral.
It arrives with authorship attached.
Preservation begins from the opposite instinct.
Dump. Verify. Store.
Dumping makes a byte-for-byte copy of a cartridge, disc or ROM chip, and a known-good checksum tells a clean dump from a damaged but convincing one.
A patch can’t substitute for that record.
A repaired version can’t establish what the unrepaired release contained.
Then comes the distinction a patch makes hardest.
A compiled binary can be recovered and, with luck, run.
Source code lets a game be understood, modified, ported, or rebuilt.
The NES Days of Thunder was rebuilt from source floppies Chris Oberth had left in a basement for roughly thirty years.
StarCraft’s remaster began with no original code and no original art assets from the game it was remastering.
One archive had everything.
One had a product and a memory.
The community patch sits uneasily between those outcomes.
It can reveal an executable well enough to amend it, without restoring the source that made the original intelligible.
It can make a binary more useful while making its undocumented assumptions less visible.
It can be the nearest available thing to maintenance, which is not the same thing as evidence.
That is the cost of preservation.
An archive can retain a shipped build, identify it against a checksum, and shelve it beside the tools required to run it.
It can also preserve an unplayable object with such fidelity that nobody meets the work except as a technical exhibit.
A perfect client is thin consolation when the missing part of the game was on a server.
The Crew was delisted, then shut down.
With no local fallback, every owner’s client went inert.
Kingdom Hearts χ took the other road.
Its developer stripped the server dependencies before shutdown, turning a server-run mobile game into something recoverable from a local copy.
Both outcomes are alterations.
One kept play by changing the game before it died.
The other lost play because no change arrived in time.
The law formalizes the difference between retaining a copy and circulating a repair. It doesn’t do so evenly, and jurisdiction decides.
Rules that let an institution preserve a work rarely let it hand the work to anyone at a distance.
Rules written for a client and a disc have little to say about a city that lived on a server.
The archive is allowed to save the game.
It’s much less often allowed to run it for you.
A patch and an archive are not rival claims on the same object.
The patch keeps a game reachable and gives up the clean boundary of the shipped work.
The archive keeps the evidence and gives up the improvisation that makes an obsolete game playable this afternoon.
One insists the work stays itself.
One insists it stay possible.
Both are honest. Neither is sufficient alone, and the corpus of surviving games is built almost entirely out of the two of them disagreeing.
Which makes authenticity the wrong test to be applying.
What must remain recoverable when alteration becomes the condition of survival?