Skip to content

docs: the name is a translated string, in the reader's script - #25

Merged
mashkovd merged 1 commit into
mainfrom
docs/name-in-readers-script
Sep 11, 2026
Merged

docs: the name is a translated string, in the reader's script#25
mashkovd merged 1 commit into
mainfrom
docs/name-in-readers-script

Conversation

@mashkovd

Copy link
Copy Markdown
Contributor

Why

The typography constraint said the name is "identical in both languages" and therefore safe in the editorial face. It is not identical — Дмитрий Машков and Dmitrii Mashkov are different scripts, which is precisely why Instrument Serif, shipping no Cyrillic, would drop the Russian half onto a fallback family.

#24's proposal found that contradiction and resolved it by rendering Latin on both sides. The contradiction was real; the resolution should go the other way, and the reason is not typographic.

Every other string on the home page is properly bilingual. If the name is the one exception, it becomes the single element that looks machine-made — on a site whose entire argument is that it was not. A Russian reader shown a transliteration of a Russian name reads a page that was translated rather than written.

The rule

English Dmitrii Mashkov
Russian Дмитрий Машков
markup an .l.en / .l.ru pair, like any other copy
face Onest, never Instrument Serif
<title> stays Latin — one string, and a browser tab is a filing label, not prose

This goes in AGENTS.md rather than into #24 because the implementer reads this file on every run, and a comment on one pull request reaches neither it nor the next cycle. Three findings this evening have been lost that way.

It does leave Instrument Serif with little to do on the home page. That is a real design question for whoever needs the face; the answer to "our editorial font has no Cyrillic" cannot be "then write the name in the wrong script".

🤖 Generated with Claude Code

https://claude.ai/code/session_01MPpAwcFk2L1B94VT8LiKVx

The typography note said the name is "identical in both languages" and
therefore safe in the editorial face. It is not identical: Дмитрий Машков and
Dmitrii Mashkov are different scripts, which is exactly why Instrument Serif,
having no Cyrillic, would drop the Russian half onto a fallback family.

#24's proposal spotted that contradiction and resolved it by rendering Latin on
both sides. That resolution is wrong, and the reason is not typographic. Every
other string on the home page is properly bilingual; if the name is the single
exception it becomes the one element that looks machine-made, on a site whose
whole argument is that it was not.

Stated here rather than fixed in that pull request because it is a standing
rule the implementer reads on every run, not a one-off correction. The answer
to "our editorial font has no Cyrillic" cannot be "then write the name in the
wrong script".

Claude-Session: https://claude.ai/code/session_01MPpAwcFk2L1B94VT8LiKVx
@claude

claude Bot commented Sep 11, 2026

Copy link
Copy Markdown

Claude finished @mashkovd's task in 14s —— View job


Review

  • Fetch pinned diff (03044de...a5a00fc)
  • Assess change: docs-only addition to AGENTS.md (+9/-0), no code/logic/config changes
  • Submit verdict

This is a documentation-only change (a new bullet point in AGENTS.md clarifying that the name is a translated bilingual string, not an exempt proper noun). No code, config values, dependencies, or runtime behavior are touched — trivial per review criteria.

Approving with no inline findings.

@claude claude Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Docs-only change to AGENTS.md, no code/logic/config touched.

@mashkovd
mashkovd merged commit 39a3d1d into main Sep 11, 2026
4 checks passed
@mashkovd
mashkovd deleted the docs/name-in-readers-script branch September 11, 2026 03:32
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant