Improve world-readiness of initials algorithm #1106
Conversation
This will enable easy unit testing, re-use, etc. going forward.
Add some special-case handling for: 1) Chinese and Korean names for which the rules for latin-alphabet names produce non-sensical results. 2) Arabic names for which the latin-rules produce potentially offensive results due to ligature contraction. Note that both of these world-readiness fixes are guaranteed to keep the existing contract that the initials function returns at most 2 characters.
|
@c-w, |
|
Putting this onto the radar of @cliffkoh. |
|
@c-w Thanks for the PR. Just wondering, did you get a chance to test if 2 asian characters (Chinese or Korean) would fit within the Persona textboy? We could possibly augment the Persona example page to include such an example which would also provide a convenient way to validate this scenario... |
| }); | ||
|
|
||
| it('calculates an expected initials for Chinese names', () => { | ||
| let result = getInitials('桂英', false); |
There was a problem hiding this comment.
Very common for Chinese to have 3 characters, sometimes even 4. It would be great to have one such test, just to ensure for instance that no more than 2 characters are returned.
There was a problem hiding this comment.
Good catch. Will add another test.
We need to ensure that even for names with three or more characters, the generated initials don't contain more than two characters.
|
Here's a render of the persona component with non-Latin names:
|
|
When I run |
|
@cliffkoh: Addressed all the comments. Could you please take another look? |
| { | ||
| "packageName": "office-ui-fabric-react", | ||
| "comment": "Improve abbreviation of non-Latin names in Persona component", | ||
| "type": "minor" |
There was a problem hiding this comment.
strictly speaking, a patch change because the API surface did not change
Pull request checklist
Description of changes
This change improves the world-readiness of the initials algorithm.
Specifically, it adds some special-case handling for:
Chinese and Korean names for which the rules for latin-alphabet names
produce non-sensical results.
Arabic names for which the latin-rules produce potentially offensive
results due to ligature contraction.
Note that both of these world-readiness fixes are guaranteed to keep the existing contract that the initials function returns at most 2 characters.
Focus areas to test
Persona component, especially with non-Latin scripts.