Sparked by this comment, I'd like to discuss whether explanations should use pseudocode or "real" code.

Reasons against pseudocode:

• It isn't standardized. This leads to inconsistencies.
• It's typically less precise than real code; it may be harder to understand, being more informal.
• You can't execute and test it.
• There is no "pseudocode" syntax highlighting in GFM; you'd have to pick the syntax highlighting of a "real" programming language and hope it works decently well with your pseudocode.

Reasons against "real" code:

• It may obstruct the algorithm by introducing language boilerplate; it is less flexible than pseudocode.

Ultimatively, I'd say I'm fine with both pseudocode and "real" code. If "real" code doesn't have significantly more boilerplate than pseudocode, I'd lean towards real code. If the boilerplate makes the algorithm much harder to understand, I'd prefer textual or pseudocode descriptions.