Skip to content

Commit c126fed

Browse files
authored
Add experimental/README.md (#518)
Explain the purpose of experimental workloads and the simplified review process.
1 parent ba41f91 commit c126fed

File tree

1 file changed

+31
-0
lines changed

1 file changed

+31
-0
lines changed

experimental/README.md

Lines changed: 31 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,31 @@
1+
# Experimental Workloads
2+
3+
Experimental Speedometer workloads can be used for workloads that do not
4+
directly meet the criteria for official workloads:
5+
6+
- Incubation of new workloads
7+
- Edge-cases that don't work well in the official setup
8+
- Testing features that aren't fully shipped in all browsers
9+
10+
The main purpose is to have a space for low-friction contributions to
11+
experiment with new ideas. There is no explicit requirement to promote
12+
experimental workloads to the official selection.
13+
14+
## Review Process
15+
16+
Contributions to experimental workloads can be merged with the approval from a
17+
single vendor.
18+
19+
This is the opposite from official workloads that are scored, where we require
20+
approval from all vendors.
21+
22+
## Requirements
23+
24+
- Stored in the `/experimental` folder
25+
- Disabled by default and thus do not contribute to the official score
26+
- Use the `experimental` tag in the Suites declaration
27+
28+
## Hosting
29+
30+
Experimental workloads will not be officially hosted on Speedometer releases on
31+
<https://browserbench.org>, unlike the official workloads.

0 commit comments

Comments
 (0)