Should diacritics examples be marked-up with their language?

[simoncozens]

On the context-of-diacritics tests in the Latin 2 section, we have, for example "espanhóis" and "źródłosłowu" in the "ó" section. The first is Portuguese and the second is Polish. The Polish ó is a o-kreska which may be designed differently to o-acute; OpenType localisation features in the font can be used to select the appropriate glyph for a text.

For testing localised features, would it make sense to mark up all of the context-of-diacritics words in spans declaring their languages so the browser can at least have a shot at giving you the right glyph? For example, <span lang="pt">espanhóis</span> ... <span lang="pl">źródłosłowu</span> and so on.

[rbmntjs]

This would be very useful, yes. Same with the Catalan punt volat. I have a really nice locl feature for that, and setting language helps a great deal.

[impallari]

How many web developers use the localization feature in their websites?
Mi guess is that the % is very near 0,0000000000000000000001%.

would it make sense to mark up all of the context-of-diacritics words

No

[simoncozens]

Sorry - I thought this was a project about testing fonts, not about testing browsers.

[davelab6]

@simoncozens I suggest forking :)

[impallari]

Testing the un-localized defaults is 99,9999 % higher priority than than testing the localized OTL replacement for websites, where localization is almost never used.
But yes, Simon & Rob, it will be nice if we are also able to test localized replacements as well.

I suggest that instead of tagging the "context-of-diacritics" tabs, we can create a new "Localization" tab and add paragraphs of text containing samples of the most usual localization cases.

Ideally in a table, in 2 columns: Default in one column and Localized in the other column (similar to the "Apostrophes" tab). I will accept a Pull Request like that.

[moyogo]

For what it’s worth, looking at some Catalan sites:

• out of 13 newspaper websites I looked at: 4 have the wrong or no language tag,11 have "ca", "ca-ES" or "ca-AD".
• out of 25 TV station websites I looked at: 11 have the wrong or no language tag, 14 have "ca", "ca-ES" or "ca-AD".
• out of 30 government websites I looked at: 8 have the wrong or no language tag, 22 have "ca", "ca-ES" or "ca-AD".

Google, Facebook and Wikipedia also tag the displayed locale properly.
My guess would be that a significant number of pageviews are tagged properly even if a majority of web designers have no clue, many of the ones that have high exposure do.

[impallari]

Thanks Moyogo! Those results are much better than I expected

[impallari]

Added Latin2 "Locale" tab to get the ball rolling. cfbf423
@simoncozens , @rbmntjs , @moyogo : Please add all the test you would like to see there, and submit pull requests