RE: New specdev material about bidi

   Hello, Richard!

I have looked at sections 3.2 and 5 of the bp-i18n-specdev document and I have the following comments.

1) In section 5.2.4, we find: "They are invisible in most editors and are therefore difficult to work with, and can easily lead to orphans and overlapping ranges."
I think this is not strictly correct. Since each CC such as LRE/RLE or LRI/RLI starts a new embedding level, there cannot be overlapping ranges other than total inclusion, which in itself is not an error.

2) Ibidem, we find: "Processors that extract parts of the data, add to it, or reuse in combination with other text it may incorrectly handle the control codes."
Editorial: "it may" => "may"
This argument is not really a con for CCs, since the same can be said of markup.
The same is true for the next argument: "Search and comparison algorithms should ignore these characters, but typically don't." 
So maybe you should keep these arguments but note that although the same caveats hold for markup, implementers might support included markup better than  included CCs.

3) In 5.2.6, "guarrantee" => "guarantee"

4) In 5.4, we find "It must be possible to also set the direction for content fragments to auto. This means that the base direction will be determined by examining the content itself." and a whole discussion after it.
However section 5.5 seems to generalize the same concepts. I suggest to merge the 2 sections.

5) There is quite some overlapping between section 4.2 "Setting text direction" and section 5 "Handling text direction". I suggest considering merging them.

Shalom (Regards),  Mati

-----Original Message-----
From: [] 
Sent: Friday, July 29, 2016 9:17 PM
Subject: New specdev material about bidi


I have published an very drafty update of the spec development guidelines in a *temporary location* at

that works on bidi topics. All the changes are in section 5 of that document.  The changes so far are mostly in the background material (moved and updated from the article i used to have on my own site, and integrated into the specdev doc). The actual recommendations still need rework – i'm actually considering whether to have separate subsections for recommendations related to markup documents vs JSON.

you'll also notice that i regrouped everything again into a single section about text direction.  That seemed to fall out of adding the additional background information. (So i may need to do the same with the language information too later.)

In addition, i've been pulling together some real world examples at

one of which is the Activity Streams model we currently need to resolve, and the other is the Web Annotation model, which we worked on before.  I plan to add other models to this as time goes by, including CSV and WebVTT.  For Activity Streams, i've tried to list problems and proposed solutions.

please take a look and send any comments.

i plan to show this to the activity streams folks prior to our discussions with them coming up next week.


Received on Sunday, 31 July 2016 07:59:22 UTC