Should U+3248 ... U+324F be wide characters?

Philippe Verdy via Unicode unicode at
Fri Aug 18 15:48:26 CDT 2017

I don't think that emojis are necessarily "square", they could be larger
(e.g. a train or a snake or an horizontal railway, or a group of several
peoples, or a cloud) or narrower (e.g. a candle).

Rendering them as square will make sense only in contexts where this makes
sense ** if possible** : monospaced fonts. But there are cases where a
single character cell would not be enough and multiple cells would be
needed (notably in text terminals, but as well in sinographic contexts
uwing multiple em-squares in a row).

The classification of widths in CJK if there to help determine how many
cells will be needed in two cases: narrow rectangular cells used in text
terminals, or square cells in classic sinographic typesetting (which is
still not mandatory because variable-width rendering is also possible, even
if it is less common, using more specific fonts for such artistic use or to
correctly render handwritten calligraphy). This classification of widths
makes no sense in Latin where it variable-width is still prefered and more

So there will be both variants for variable-width and "monospaced"
(cell-based) rendering of emojis, like they both exist for CJK and Latin:
Latin letters has a "narrow" width in sinographic square contexts only to
allow two letters side-by-side per square instead of centering them with
wide gaps or rendering them in widdened variants. Most Asian emojis from
CJK charsets will render in a single square cell, but others may still need
two square cells for better rendering (without having to use variable width
that would break the grid layout).

When rendering Latin words in CJK contexts, the alignment to the grid may
also be made only on spans of Latin letters (one or more words), by
recentering it in a row of as many cells that could fit: it would be even
more useful for Arabic sequences. This technic however would not fit very
well in classic "text terminals" where half-width Latin, Hebrew and Arabic
will still be preferable (or full-width for some Arabic letters with some
extenders, or some long Arabic ligatures).

2017-08-18 14:21 GMT+02:00 Andre Schappo <A.Schappo at>:

> On 18 Aug 2017, at 00:50, Philippe Verdy via Unicode <unicode at>
> wrote:
> 2017-08-17 18:46 GMT+02:00 Asmus Freytag (c) via Unicode <
> unicode at>:
>> On 8/17/2017 7:47 AM, Philippe Verdy wrote:
>> 2017-08-17 16:24 GMT+02:00 Mike FABIAN via Unicode <unicode at>:
>>> Asmus Freytag via Unicode <unicode at> さんはかきました:
>>> Most emoji now have "W", for example:
>>> That seems correct because emoji behave more like Ideographs.
>>> Isn’t this the same for “CIRCLED NUMBER TEN ON BLACK SQUARE”?
>>> This seems to me also more like an Ideograph.
>> Not really. They have existed since extremely long without being bound to
>> ideographs or sinographic requirements on metrics. Notably their baseline
>> and vertical extension do not follow the sinographic em-square layout
>> convention (except when they are rendered with CJK fonts, or were encoded
>> in documents with legacy CJK encodings, also rendered with suitable CJK
>> fonts being then prefered to Latin fonts which won't use the large
>> siongraphic metrics).
>> If they were like emojis, they would actually be larger : I think it is a
>> case for definining a Emoji-variant for them (where they could also be
>> colored or have some 3D-like look)
>> There's an emoji variant for the standard digits.
> Do you speak about circled numbers ? I don't think so.
> I (and Mike as well to which I was replying) was speaking about a good
> case for defining emoji variant of these circled (or squared) numbers (Mike
> spoke about circled number 10, which is not encoded as an emoji and not
> even as an ideograph, and that he proposed to give a wide width property
> like ideographs).
> Are not CJK ideographs both (W)ide and (S)quare? Does (W)ide imply or
> define that the ideograph should also be (S)quare?
> It seems to me that there are many characters that are both (W)ide and
> (S)quare eg emoji
> André Schappo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the Unicode mailing list