Re: HTML named entity issue [I18N-ACTION-1244]

Some points to consider:

[1] We should perhaps limit our request to invisible characters (like 
NNSBP, RLI, etc.)

[2] We have recently gone through a similar and arguably successful 
transition with the RLE->RLI change.  All major browsers now support use 
of RLI etc.

[3] Addition of any new feature to HTML causes the same backward 
compatibility issue they describe in their rationale to avoid adding these.

[4] It's not as easy as remembering the numbers of code points (which is 
already difficult for some of us, and requires us to take time out from 
authoring to check we're using the right thing), but many RTL scripts 
require the content author to remember a large number of code point 
values, some of which are used very infrequently.  See for example 
https://r12a.github.io/scripts/arab/arb.html#directioncontrols (which 
doesn't even list all of the ones you'd need when working in such scripts).

[5] The comments at 
https://github.com/mfreed7/html/commit/5662df8570a1c93c484943544bc77d8497b906da 
take into account the cost to the ecosystem, but dismiss the benefits as 
a mere convenience.  I'd argue that the current situation is actually a 
bit of a headache for people working with content in many scripts around 
the world.

ri



Addison Phillips wrote on 23/02/2023 16:35:
>
> All,
>
> Pursuant to my action item [1] to find and possibly re-open the issue 
> about named entities for isolating bidi controls, I traced the issue 
> and found the relevant part of the conversation roughly here:
>
> https://github.com/whatwg/html/issues/5121#issuecomment-719406247
>
> The WHATWG HTML group decided to add **no** new named entities ever 
> and the reason is that it would break existing HTML processors (not 
> just browsers). I did not re-open 5121. I think we should discuss 
> whether we should attempt to re-raise this.
>
> [1] https://www.w3.org/International/track/actions/1244
>
> Addison Phillips
>
> Chair (W3C Internationalization WG)
>
> Internationalization is not a feature.
>
> It is an architecture.
>