This is why I suggested specifying the goals of the redesign explicitly. Without a standard to measure against, however subjective, this sort of exercise will always devolve into a popularity contest. And those rarely end in a consensus.
For example, the redesign goals might read something like this:
The logo should:
- Show a clear connection to the Rust programming language
- Emphasize rust-analyzer’s nature as a productivity-enhancing tool
- Be easily recognizable at a glance, even at small scales
- Be readable on both light and dark backgrounds (possibly by having multiple variants)
- Have variants suitable for multiple uses:
- As a website favicon
- As a VSCode extension icon
- As the header of a rustdoc/mdbook sidebar
With these in hand, you can then describe how the current logo falls short of these goals, and also how the various design decisions you made support them. For (1), you can look for common elements in the graphic design of other things in the Rust ecosystem (Rust itself, Cargo, crates.io, TRPL book, etc.) and show how your proposal aligns with them better than the old logo. For (2), you’d do the same but with other software engineering tools. And so on for the rest.