A Proposal for Bidi Isolates in Unicode

●
●
●
“text2 LRI
”, “ PDI
text4”
“text3”
“text5”
Add:
BD10. The directional isolate status is a Boolean value reflecting whether the current embedding
level was started by an FSI, LRI, or RLI (as opposed to an LRE, RLE, LRO, or RLO).
This is followed by Table 3.
Replace the first row of Table 3 (defining N) with the following:
N
Neutral, Separator or Isolate formatting code (B, S, WS, ON, FSI, LRI, RLI, PDI)
Replace the last two rows (currently defining sor and eor, which we no longer use) with the
following:
sos
eos
The text ordering type (L or R) assigned to the virtual position preceding the first
level run in an isolating run sequence.
The text ordering type (L or R) assigned to the virtual position following the last
level run in an isolating run sequence.
Section 3.2 (Bidirectional Character Types): Add the new types
The current grouping of LRE, RLE, LRO, and RLO as "Strong" and PDF as "Weak" doesn’t
make any difference to the present algorithm or even make much sense. Instead, add a fourth
group, "Formatting", consisting of LRE, RLE, LRO, RLO, PDF, LRI, RLI, FSI, and PDI.
Section 3.3.1 (The Paragraph Level), rule P2: Skip over isolates
After "In each paragraph, find the first character of type L, AL, or R", insert:
while skipping over any characters between an FSI, LRI or RLI and its matching PDI or, if it has
no matching PDI, the end of the paragraph.
Section 3.3.2 (Explicit Levels and Directions): Include isolates
Modify the first paragraph to read as follows: