How Bezgelor is Built and Tested

What we know, what we guessed, and how we check it. Open to everyone.

Commands that start with / and carry the chat icon, such as /c bug, are typed in the game's chat box.

Community

Roadmap

What has been validated, what is being tested now and what comes next: the milestones of levels 1 to 30 and beyond, PvP in parallel, and how the states are derived from the test reports.

How we test

What has and has not been tested for release 1.0: the levelling evidence and the confirming runs against the release bar, with the bug budget, the open known issues, the status of each platform, and how to help test.

Coverage matrix: the 1 to 15 quest path

Every quest on the 1 to 15 path against every class and faction: which test runs have completed it, which found a blocker, and which have not reached it. Detailed evidence, not the release bar; it lists the known broken quests (bug budget).

Where Bezgelor stands

How much of the first release works today, zone by zone, rebuilt from the published quest tests: what is ready, what stopped and what is not done yet.

Engineering notes

Short, dated results from the project's engineering, testing and research work: what was tried, what was found, what it costs and what was decided.

How we build, test and tune instances and bosses

How instances and bosses are built, tested headless, tuned and compared with retail, and the results for every boss.

How we test levelling

How we test levelling: the headless day-one runs, their leaderboards, records and fun facts, rebuilt from the published reports.

Warrior

Every day-one run of the warrior: its latest results, its abilities and what stops it.

Test reports

Published, reproducible runs: the day-one levelling tests, combat and quest audits, soak and zone-tick benchmarks and the encounter simulator reports. All data is synthetic.

How Bezgelor works

Architecture: how Bezgelor's server fits together

A tour of the server, following a player from launching the client to fighting in a zone: the parts, how they talk, and where each one lives in the code.

Why Elixir & OTP

Why the server is written in Elixir: processes, message passing, supervision, binary pattern matching and in-memory tables, and how each shows up in Bezgelor.

How we work

How a change travels from a player's report to a published test result: written specs, isolated branches, independent review, seven layers of testing and honest labels.

Glossary

The words used on these pages, from Elixir and OTP (process, GenServer, supervisor, ETS) to WildStar and Bezgelor (realm, zone instance, opcode, telegraph).

How close to WildStar

Fidelity map: how close each part of Bezgelor is to WildStar

Every game system and zone rated on one scale (Unbuilt, Working, Complete, Checked), with the evidence and what it takes to go further.

Bezgelor and WildStar: where we differ

Where Bezgelor differs from retail WildStar, and how sure we are of each piece of data.

Ground truth: what survives of retail WildStar

What still exists of how retail WildStar behaved, what each source can settle, and how to help with footage, memory or a citation.

Ground truth: how our data compares with the surviving retail sources

Our creature, vendor, drop and quest data compared with the facts harvested from archived retail sources: confirmed, contradicted, filled.

Semantic registry

What each quest objective type and prerequisite field means, how sure we are, and how much of the real data our code handles without special cases.

On fidelity: what we can and can't reconstruct

Our view of a community essay on AI and MMO emulation, where it applies to us, and what we are doing about fidelity.