What's New in Rust's "Devblog 40" Update

Devblog 40 went up at the very end of December 2014 and contained no patch notes at all. Instead it was a year-in-review post-mortem, opening on the idea that game development isn't a science but a creative process, and that it's important to highlight what you're doing wrong as well as what you're doing right - because making mistakes is healthy, but not learning from them isn't. Facepunch then graded themselves.

Devblogs

The devblogs had been running for about 40 weeks at this point. Before them, the site only got posts when an update shipped. The idea came from a talk Robin Walker gave at SDC, arguing that doing all this development work and then presenting it as a bare list of changes short-changes yourself. Writing it up properly presents the work to the community and to the development team - a kind of open review of time spent.

The honest admission was that the blogs got away from them. The team ended up working harder to impress devblog readers than actual players, which isn't sustainable when some weeks are purely bug fixes or performance work that doesn't read well. The correction: rein the blogs in, be more even-handed, only post things that are actually in the game, and only include concept art when there's a real reason rather than to fill space.

Cheats and Easy Anti-Cheat

Cheating was a huge problem in the old version - fairly described as having killed Rust for a lot of people. Facepunch wrote simple checks and banned a large number of players, VAC banned more, but the team ended up spending more time writing anti-cheat and trapping cheaters than developing the game.

The answer was outsourcing, and for the newer version they chose Easy Anti-Cheat. The reasoning: it was the least obtrusive solution they'd found, the people behind it clearly loved the work, and the company was small enough to actually care about a partner like Facepunch. EAC had done a lot for them, including producing OSX and Linux versions of their client. There had been short-term problems with EAC infrastructure dropping clients from servers, but they were always available to fix them. The stated goal was realistic rather than absolute - you can't remove cheating from a PC game entirely, only make it difficult and unattractive.

Early Access

The Early Access launch overwhelmed the studio. Selling an enormous number of copies and pulling in a huge player base was great, but not sustainable - Garry said he'd have preferred a slow burn into being a popular game in a year over the huge initial spike. The money was welcome but not needed for development, and once a million copies had sold a lot of people expected Facepunch to hire 50 people, finish the game in two months, and then fire them all. That, he argued, misunderstands what Early Access is - it isn't Kickstarter - and misunderstands why they're game developers in the first place: the fun of developing a game, not the fun of writing pay cheques.

Given a second run at it, he'd have sold Rust exclusively from their own website for a year before going on the Steam Store. The catch is that they'd then be handling billing, chargebacks and credit card fraud themselves, so the fantasy version has Valve handling billing while the game stays discoverable only through their site.

The reboot

On scrapping the old codebase, the verdict was nuanced. The reboot itself was inevitable, but the focus should have been on making it a cleaner 1:1 copy of the legacy version first. Where they'd failed for a long time was getting sidetracked 'developing' the game before the foundations were laid - they should have matched legacy's stability, gameplay and performance before branching out into anything new. Garry framed this as a lesson learned, and pointed at the improvement over the preceding couple of months as evidence.

Grade: C-

The self-assessment was blunt: progress was made, but it was superficial. The game looked better than it ever had, but what was gained in graphics had been lost in gameplay - and that was called unforgivable.

Plans for 2015

The coming months were about clawing that gameplay back by re-introducing legacy features: radtowns, blueprints, workbenches, roads and researching. Bringing back the legacy map was on the list, as was making user-created maps possible. There had even been talk of bringing zombies back, with no promises attached. Only once players were happy with gameplay, stability and performance would the team start exploring the mind map introduced a couple of weeks earlier.