You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
PR #15 implements German, Italian, Brazilian Portuguese and Mandarin (Simplified Chinese with optional pinyin). This issue carries forward the proficient-speaker validation criteria from #10, #11, #12 and #13 so that implementation closure does not imply linguistic sign-off.
Automated checks and live iPhone sessions cover the application paths, real voice output and synthetic typed input. They do not establish human speech recognition, pronunciation, tone accuracy or correction quality. See the detailed results in verification/validation.md.
For each language, record the build, device, reviewer proficiency, scenarios and findings:
German: complete beginner and advanced spoken exchanges; assess pronunciation, recognition, target-language consistency and useful corrections.
Italian: complete beginner and advanced spoken exchanges; assess pronunciation, recognition, target-language consistency and useful corrections.
Brazilian Portuguese: complete beginner and advanced spoken exchanges; assess Brazilian usage and pronunciation, recognition, target-language consistency and useful corrections.
Mandarin: complete beginner and advanced spoken exchanges; assess recognition, spoken tones, correction quality, Simplified Chinese output, and pinyin accuracy/readability. Include ambiguous words, names and connected-speech tone changes; the current helper displays dictionary tones.
For each session, check that learners can reply in their chosen language while Mural continues in the learning language, and that optional meanings remain useful. File reproducible defects and link any fixes here. Record unresolved limitations before closing this validation issue.
PR #15 implements German, Italian, Brazilian Portuguese and Mandarin (Simplified Chinese with optional pinyin). This issue carries forward the proficient-speaker validation criteria from #10, #11, #12 and #13 so that implementation closure does not imply linguistic sign-off.
Automated checks and live iPhone sessions cover the application paths, real voice output and synthetic typed input. They do not establish human speech recognition, pronunciation, tone accuracy or correction quality. See the detailed results in
verification/validation.md.For each language, record the build, device, reviewer proficiency, scenarios and findings:
For each session, check that learners can reply in their chosen language while Mural continues in the learning language, and that optional meanings remain useful. File reproducible defects and link any fixes here. Record unresolved limitations before closing this validation issue.