All right ProtonDB contrbutors, this one's for you. In an update from just over two years ago, a questionnaire flow was introduced for more nuanced reports, more structured data, and more accurate ratings. The first modifications to that flow are now live, updating to the landscape of Linux gaming in 2022.

So how has this changed things? Let's dive in.

The nature of the Tinker

The goal of the split between 'Steam Play' and 'Tinker' reports was primarily a way to know how viable Proton gaming was for less-technical users. Not all tinkering asks the same level of effort - adding a launch option is relatively straightforward. Now, whether a report is one or the other will be inferred from contributors' responses. What tinker steps were taken are now presented at the top of the report for the reader to decide if it's worth trying.

Copyable Launch options

Speaking of launch options, some used to make a game run properly on Linux aren't specific to Proton. There's much variance and duplication across reports for how authors express this. Readers, on the other hand, often just want to copy the launch options to paste into Steam to try for themselves. To improve this on both sides, rather than select from a list, authors are now asked to paste the launch options they used. Reports will render this in a clear, easily-copyable code block at the top of the report.

GE, Experimental and other custom Protons

At the time of the last update to the contribution flow, custom Protons were still in their infancy and complex to use. Things have changed, and the rise of Glorious Eggroll in particular has been remarkable to watch. Valve themselves began offering an 'experimental' Proton with bleeding-edge features that players could opt in to. Both have been particularly important for new game releases. Various workarounds have been added to the contribute flow such as specifying '-ge-' or 'experimental' into the version field. You won't need to do that anymore because custom Proton builds are now their own question at the start of the flow and are more tightly integrated with processing of reports.

Welcoming Native reports

ProtonDB was created to answer how well games work on Proton. Back then, it was generally true that if a native build for a game existed, it was preferable to use that. But as Proton has improved, that's no longer strictly true. Authors have already been writing 'Proton' reports specifying that they're not using Proton at all. Let's fix that: now you can select 'Native. Not Proton' as an option to the question of 'What variant of Proton do you use?'

Playtime on Linux

Reports will now be submitted with Steam data on how many hours the author has played on Linux. This will be displayed, hopefully proudly, next to how much they've played compared to all those other silly operating systems.

Streamlined authoring flow

Some questions have been pruned since they can be reasonably inferred from Steam data. Linux playtime data for example means we don't need to self-report how many hours you've played on Proton (although Linux playtime is admittedly not a 1:1 measurement). Performance improvements to the questionnaire and form validation mean some questions have been consolidated to a single page. Authors of tinker reports will now click the 'Next' button four less times.

Encouraging non-English contributions

For users using the site in a non-English locale, you'll be asked a new question: What language are you writing the report for. Your answer to this will be used in a future release to allow visitors to prioritize reports in their native language.

Dual Verdicts

Sometimes you tinker cuz you want to, not because you have to. If an author wrote a tinker report, was it crucial to making the game playable or was it an optional tweak? Now if an author is writing a report that signals tinkering, at the 'Verdict' step they'll be asked if they tried playing without intervention. If so, they'll give a verdict on the 'out-of-the-box' click-play experience.

And more!

There's already a wall of text here, so I'll leave the rest summarized in the changelog:

• Tinker state is now inferred from responses.
• Proton version and launch options moved to first step after app selection.
• Launch options are now pasted, and flags inferred.
• Proton 'variant' options: GE, Experimental, Older Proton release.
• Proton options: Native (no proton) is offered if game known to have a linux build.
• Non-english users have a new question: What language will you write this report in? This will aid future sort/filtering by language on report listing.
• Customizations: Question is now integrated with setup step.
• Customizations: Removed native2proton option as it has not seen a change in several years. Specify via 'not listed' if you are using.
• Tinker flow: Will be asked at the Verdict step if you've tried playing without tinkering and what the verdict of that is.
• Launcher: Removed Lutris as a launcher option since it is not supported as a Proton runner. Specify a 'Lutris install script' as a customization if you are using one.
• Form validation: Fixed cases where answers were invalidated or cleared inconsistently, particularly on notes and subquestions.
• Form validation: Improved performance.
• Next/Prev buttons give focus indications (nice for tabbing through questions).
• Question for hours played removed and now derived from Steam data.
• Text inputs now extend width of the contribute page.
• Steps for graphics, stability, and other potential issues have been consolidated into a single 'technical details' step for a faster authoring flow and easier cross-reference of notes using.
• App selection: Less finicky on selection, delays calls to steam for a moment while typing.
• App selection: Game thumbs now bordered. Most important new feature.
• App selection: Regression fix: Show title of game in input field when selected.
• Multiplayer: Can now add notes for appraisals of online or offline.
• Multiplayer: Removed question for impact from anti-cheat as it is redundant with new game metadata and addition of the notes field for multiplayer appraisal.
• Multiplayer: Removed question for is multiplayer important. Now derived from steam data.
• Reports: Playtime now shows how much time the reporter played on linux vs other operating systems.
• Reports: Dual-verdict rendering if author has specified both.
• Reports: Inferred tinker steps are displayed at the top, SteamPlay/Tinker badges removed.
• Translations: Spanish updated thanks to @santisteban.

On the horizon

These changes won't be the last for the contribution flow. The 'molten core' is hot, so your feedback will help improve it especially now. If you have any suggestions, do reach out to us on Discord or IRC.

This update amounts to ProtonDB's most substantial change since the questionnaire. But between the lines, you might have surmised that an even bigger change is nigh upon us. And you'd be right. A new rating system, empowered by the new authoring flow and ready for the Steam Deck, is approaching ship-shape.

Finally, a big thank you to all the Patreon supporters who've helped cover hosting costs and encourage further development. You've been a big motivation and there's plenty more to come. If you'd like to continue to see me avoid boring day jobs for a more noble cause at an important juncture for Linux gaming, consider pitching in yourself here.

Cheers!

--migelius (@buck)