## Feature proposal #8092

### API: Date class: Add human-readable PHP output of "time ago" time scale

**Related issues**

### History

#### #1 Updated by beat 2 months ago

**Related to***Feature proposal #8079: Internationalization: Better pluralization of time-periods*added

#### #2 Updated by beat 2 months ago

# Instructions / Reminder for translators:¶

CB Pluralization in translations (with one numeric substitution, e.g. "[X]" can happen as one of following (the indexed form being prefered):

- interval:
**{0} There are no apples|{1} There is one apple|]1,Inf] There are [X] apples** - indexed:
**There is one apple|There are [X] apples**

### For intervals: intervals are like in maths:¶

**[2,5]**includes 2,3,4,5,**]2,5]**includes 3,4,5,**]2,5[**includes 3 and 4.**Inf**and**-Inf**are infinites.

### For indexed:¶

- The indexed solution can also contain labels (e.g. one: There is one apple). This is purely for making the translations more clear - it does not affect the functionality.
- in English and most languages, there are only 2 choices: singular|plural
**In some languages, there are more cases**(up to 6 for arabic), that can be found at the end of CB file libraries/CBLib/CBLib/Language/CBTxt.php, e.g.:- Here a good description of the cases, separated by "|":

### The two methods can also be mixed:¶

**{0} There are no apples|one: There is one apple|more: There are count apples**

## Regarding this feature:¶

This feature introduces e.g. this translation string:

'X_YEARS' => "one year|[X] years"

2 examples: English:

0 hours

1 hour

2 hours

...

Translates in French (notice that 0 is singular) to:

0 heure

1 heure

2 heures

3 heures

...

Which has only 2 cases:

nplurals=2; plural=(n > 1);

But as first case has many cases as it's >1, the string must be containing the number:

'X_YEARS' => "[X] heure|[X] heures"

Alternatively to have nicer translations, it could be using a mix of the interval enumerated form for litteral exceptions, and indexed form for the numeric cases:

'X_YEARS' => "{0} aucune heure|{1} une heure|[X] heure|[X] heures"

But in Polish, it's a bit more complex with 3 forms: Particular cases for "1" and numbers ending in 2, 3 and 4 but not with 12 and 13

0 godzin

1 godzina

2 godziny

3 godziny

4 godziny

5 godzin

6 godzin

7 godzin

8 godzin

9 godzin

10 godzin

11 godzin

12 godzin

13 godzin

14 godzin

15 godzin

...

21 godzin

22 godziny

23 godziny

24 godziny

25 godzin

...

100 godzin

101 godzin

102 godziny

103 godziny

104 godziny

105 godzin

...

Which corresponds to this rule, found in the CB file mentionned above:

nplurals=3; plural=(n==1 ? 0 : n%10>=2 && n%10<=4 && (n%100<12 || n%100>14) ? 1 : 2);

So the polish translation would be (with "one year" translating to "jedna godzina"):

'X_YEARS' => "jedna godzina|[X] godziny|[X] godzin"

# Combining the years, months, days, hours, minutes and seconds¶

There are 3 new named strings as well:

'CB_FIRST_COMMA_OF_YEARS_COMMA_MONTHS_AND_DAYS' => ", ", 'CB_LAST_AND_OF_YEARS_COMMA_MONTHS_AND_DAYS' => " and ", 'CB_ONLY_AND_OF_YEARS_AND_MONTHS' => " and "

- Thie first 2 are f is fir the case with 3 or more date/time periods, e.g: "1 hour, 2 minutes and 3 seconds"
- The 3rd and last is for cases with 2 date/time periods, e.g. "4 months and 5 days"

Also note that for years, months and days, there are also calendar variants: E.g.:

'X_CALENDAR_YEARS' => "one calendar year|[X] calendar years"

**IMPORTANT**: note the spaces before and after the " and ", and after the comma ", " for correct formatting.

# Forms "X time ago", "in X time" and "X time"¶

Some languages have conjugation depending on the use of the time period. E.g. (serbian):

That's why there are 3 forms, identical in English, but can be different in other languages:

X_MONTHS

X_MONTHS_PAST

X_MONTHS_FUTURE