[Unicode]   Common Locale Data Repository : Bug Tracking Home | Site Map | Search
 
Modify

CLDR Ticket #11354(closed: fixed)

Opened 3 months ago

Last modified 7 weeks ago

root.xml and others have ⩾ instead of ≥

Reported by: shane Owned by: mark
Component: numbers Data Locale:
Phase: rc Review: shane
Weeks: Data Xpath:
Xref:

Description

From the latest survey tool dump: In the numberElements miscPatterns section:

4129	                <miscPatterns numberSystem="latn">
4130	                        <pattern type="approximately">~{0}</pattern>
4131	                        <pattern type="atLeast">⩾{0}</pattern>
4132	                        <pattern type="atMost">≤{0}</pattern>
4133	                        <pattern type="range">{0}–{1}</pattern>
4134	                </miscPatterns>

The patterns for atLeast and atMost have inconsistent symbols: they should be one of the following:

  • ≤{0} and ≥{0}
  • ⩽{0} and ⩾{0}

Attachments

Change History

comment:1 in reply to: ↑ description Changed 3 months ago by kent.karlsson14@…

Replying to shane:

4130	                        <pattern type="approximately">~{0}</pattern>

...

  • ≤{0} and ≥{0}
  • ⩽{0} and ⩾{0}

I find the second option here preferable.

In addition, ≈ (ALMOST EQUAL TO) is better for type="approximately", than to use tilde.

comment:2 Changed 3 months ago by mark

I agree in principle.

In practice, I think we should weigh that ~, ≥, ≤ are easier to type on at least some systems.

From https://www.unicode.org/cldr/charts/latest/keyboards/chars2keyboards.html (albeit a bit old)

≤ android:* osx:60/82

⩽ <none>

(Also, ~ is often used as a unary operator, and occurs in style books more than ≈.)

comment:3 Changed 3 months ago by mark

Note: on my system

⩾ is supported by Segoe UI Symbol (Windows), Noto Sans Symbols, Symbola, Apple Symbols.
≥ is supported by ... too many to list.

comment:4 Changed 3 months ago by shane

Locales are free to customize the symbols; French for example chose to use ≈. However, for root.xml, if in doubt we should stick with simpler symbols with wide support.

comment:5 Changed 3 months ago by kent.karlsson14@…

⩽ and ⩾ look better (I know, "in the eye of the beholder"...); I think they are more in European tradition. Pity that they are not picked up more in fonts (maybe due to their "late" addition to Unicode).

comment:6 follow-up: ↓ 7 Changed 3 months ago by srl

≤ ≥ are easier to type on a US Mac (one of the links above said that ≤≥ was the DIN standard. The others do look nicer, but I wouldn't know how to type them.

comment:7 in reply to: ↑ 6 Changed 3 months ago by kent.karlsson14@…

Replying to srl:

Good discussion. From Russia to Brazil... (and everywhere in-between...?) using ⩽ and ⩾.

And if the font is well-designed, they are perfectly distinguishable from < and >.

comment:8 Changed 3 months ago by shane

It sounds to me like locales "from Russia to Brazil" can choose the slanted version for their override of root.xml if they want.

I think root.xml should err on the side of the older and more widely supported characters.

comment:9 follow-up: ↓ 10 Changed 3 months ago by shane

On the other hand, an argument could be made that this is a good opportunity to evangelize the new characters.

comment:10 in reply to: ↑ 9 ; follow-up: ↓ 13 Changed 3 months ago by srl

Replying to shane:

On the other hand, an argument could be made that this is a good opportunity to evangelize the new characters.

Not usually… In general, CLDR should reflect usage, and does *not* use characters before they are widely implemented (currency etc.). However the 'new' character are some dozen years old.

It sounds to me like locales "from Russia to Brazil" can choose the slanted version for their override of root.xml if they want.

right.

Good discussion. From Russia to Brazil... (and everywhere in-between)

except Germany?

comment:11 Changed 3 months ago by kent.karlsson14@…

However the 'new' character are some dozen years old.

https://www.encyclopediaofmath.org/index.php/Mathematical_symbols says:

>,< 	greater than, smaller than	T. Harriot	1631

But does not mention either variant of the less/greater than or equals signs. Maybe I should spend a day or two at the math library, trying to find some kind of sampling.

except Germany?

Symbols in math aren't ever regulated by national standards, AFAIK. This probably refers to an engineering standard of some sort. And the authors may well have been unaware of this fine point...

comment:12 Changed 3 months ago by mark

  • Owner changed from anybody to mark
  • Phase changed from dsub to rc
  • Priority changed from assess to major
  • Status changed from new to accepted
  • Milestone changed from UNSCH to 34

comment:13 in reply to: ↑ 10 Changed 3 months ago by mark

Replying to srl:

Right, as Steven says, we look for deployment in fonts before using characters, no matter how "old" they are.

Replying to shane:

On the other hand, an argument could be made that this is a good opportunity to evangelize the new characters.

Not usually… In general, CLDR should reflect usage, and does *not* use characters before they are widely implemented (currency etc.). However the 'new' character are some dozen years old.

It sounds to me like locales "from Russia to Brazil" can choose the slanted version for their override of root.xml if they want.

right.

Good discussion. From Russia to Brazil... (and everywhere in-between)

except Germany?

comment:14 Changed 3 months ago by mark

  • Review set to mellie

comment:15 Changed 3 months ago by mark

  • Review changed from mellie to shane

Mellie, transferring review to Shane, since he filed.

comment:16 Changed 3 months ago by mark

  • Status changed from accepted to reviewing

comment:17 Changed 7 weeks ago by shane

  • Status changed from reviewing to reviewfeedback

LGTM

I did notice one place where the long characters are still used in the context of the atLeast/atMost patterns, the spec file. The example is still technically correct, so it is probably fine. However, more importantly, the example is missing the atMost and approximately patterns.

Probably not a big deal, but just pointing it out.

https://unicode.org/cldr/trac/browser/trunk/specs/ldml/tr35-numbers.html#L837

comment:18 Changed 7 weeks ago by mark

  • Status changed from reviewfeedback to reviewing

comment:19 Changed 7 weeks ago by shane

  1. The latest commit on this ticket appears to include things besides just documenting the new patterns; in particular: trunk/specs/ldml/tr35.html: Under "Revision 52", there is a new title "Part 1", as well as Parts 4 through 7.
  2. The indentation changed in weird ways, making it harder to see what changed.

comment:20 Changed 7 weeks ago by shane

  • Status changed from reviewing to reviewfeedback

comment:21 Changed 7 weeks ago by mark

  • Status changed from reviewfeedback to reviewing

That is just setting up the modifications section. It is clearer if you look at http://unicode.org/repos/cldr/trunk/specs/ldml/tr35.html#Modifications

Sorry about the indentation; that's an artifact of Dreamweaver. As long as the .html above looks ok we are good.

comment:22 Changed 7 weeks ago by shane

  • Status changed from reviewing to closed
  • Resolution set to fixed

Thanks!

View

Add a comment

Modify Ticket

Action
as closed
Next status will be 'new'
Next status will be 'closed'
Author


E-mail address and user name can be saved in the Preferences.

 
Note: See TracTickets for help on using tickets.