Draft Minutes of UTC Meeting 188
Redmond, Washington, United States — July 28-30, 2026
Hosted by Microsoft in Redmond and virtually
UTC #188 Agenda
Revision date: 2026-08-27
Logistics: https://www.unicode.org/timesens/logistics-utc188.html
Total full members in good standing = 8 (Adobe, Airbnb, Apple, Google, Meta, Microsoft, Salesforce, Translated)
Total supporting members in good standing = 4 (MeitY, Stanford, Tamil Nadu, UCB)
Voting members in regular attendance = 4.5: Adobe, Apple, Google, Microsoft, UCB.
Quorum for UTC #188 is 2.25.
Peter Constable opened the meeting at 11:01 UTC−7 (20:01 UTC+2) .
Peter Constable presented the agenda.
4.5 voting members represented: Adobe, Apple, Google, Microsoft, UCB.
Associate members present: Ebanah, SIL.
[188-C1] Consensus: Approve minutes of UTC meeting #187, as amended in discussion.
[188-A1] Action Item for Robin Leroy, UTC: Post approved minutes of UTC #187.
Peter Constable reminded participants that most action items are now tracked by a working group.
Robin Leroy apologized for the delay in posting approved minutes of meetings #182 through #186, and stated that these action items have now been completed.
Oral report from Peter Constable.
Debbie Anderson presented document L2/26-162.
Debbie Anderson asked how many pages would be removed from the main document of ISO/IEC 10646 by moving the code charts to linked content. Michel Suignard estimated around 3300.
Ken Whistler asked whether the plan was for Amendment 1 to be synchronized with Unicode 19. Michel replied that this was the case. Peter Constable noted that there would be an option to synchronize with Unicode 20.
Peter Constable noted that the PAG had been immediately informed of the WG 2 decisions to change the name of the Seal characters and to remove the Chisoi script, and were able to make these changes in draft data for The Unicode Standard 18.0.
Markus explained that because of the timelines of CLDR processes, he had not yet been able to do the ICU and CLDR work falling out of the removal of Chisoi.
Peter Constable noted that a maintenance agency would help align the timelines, the flipside being that targeting for a specific version of The Unicode Standard would occur after an approval by the MA voting group.
Ken Whistler reaffirmed that these late removals are highly disruptive, as many data files are already prepared. Michel Suignard pointed out that the fact that WG 2 only meets once a year makes synchronization difficult.
Peter Constable explained that the creation of the AG would allow IRG (currently a subgroup of WG 2) to exist if WG 2 were dissolved.
Oral report from Peter Constable.
Peter Constable presented document L2/26-160.
Roozbeh Pournader asked about the glyph changes for the archaic shriis recommended at the WG 2 meeting. Michel Suignard stated that all changes could be found in the WG 2 recommendations. Michel Suignard clarified that the glyphs required further adjustments to match the style of the existing font, and that experts from BIS approved of the corrected glyphs.
[188-C2] Consensus: Revert approval of Chisoi script for Unicode Version 18.0 (ref. 184-C2) and set the status of the 40 code points, 16D80..16DA9, to provisionally assigned.
[188-C3] Consensus: Change the algorithmic names of 11328 Small Seal characters from “SEAL CHARACTER-XXXXX” to “SMALL SEAL CHARACTER-XXXXX”, for Unicode 18.0.
[188-A2] Action Item for Ken Whistler, UTC: Update the pipeline to reflect the name change for Seal characters in Unicode 18.0, and postponement of Chisoi characters from Unicode 18.0 to Unicode 19.0 [ref. consensus 188-C2, 188-C3].
[188-A3] Action Item for Markus Scherer, PAG: Update data files for Unicode Version 18.0, UTS #10, UTS #39, UTS #46, and UTS #58, reflecting the name for Small Seal characters, and removing 40 Chisoi characters postponed to Unicode 19.0 [ref. consensus 188-C2, 188-C3].
[188-A4] Action Item for Michel Suignard, CHARTS: Update charts for Unicode Version 18.0, removing Chisoi script (postponed to Unicode 19.0) [ref. consensus 188-C2].
[188-A5] Action Item for Ken Whistler, EDC: Update the core spec for Unicode Version 18.0 as appropriate to remove references to Chisoi script (postponed to Unicode 19.0) [ref. consensus 188-C2].
[188-A6] Action Item for Ken Whistler, EDC: Update Table 4-8 in the core spec to reflect the change of algorithmic names for Small Seal characters, for Unicode Version 18.0 [ref. Consensus 188-C3].
Peter Constable reminded participants that in 2025, the Release Management Group was preparing a recommended preliminary repertoire for 18.0 in April. Since Michel Suignard, as editor of 10646, needs a tentative set, the RMG presents a recommended preliminary repertoire for Unicode 19.0 in L2/26-210. Peter Constable noted that many characters in the pipeline are not included because of uncertainty on the workload required of working groups.
Roozbeh Pournader asked whether the properties work was the limiting factor. Peter Constable replied that this was one major concern, as most of the pipeline already had draft properties last year at this time, but that other required work was also taken into account.
Ken Whistler inquired about the novel concept of approving VSes for a provisionally assigned character.
Anushah Hossain asked how the work on properties was prioritized. Peter Constable replied that no attempt is regularly made to bundle thematically similar scripts into the same release; but difficulty of the work, time since provisional assignment, and special urgency (as for the currency symbols and the Leibniz characters) are taken into account. Robin Leroy also explained that there was a desire to prioritize scripts from India to provide as much time as possible for feedback. Michel Suignard noted that there were some inconsistencies as to which IPA characters made it in. Peter Constable replied that given the steady flow of phonetic characters, they were prioritized chronologically. Ken Whistler presented a version of the pipeline split according to the preliminary targeting for 19.0. Michel Suignard also pointed out that WG 2 N5361R shows the entirety of the provisionally assigned pipeline.
Jan Kučera asked when the next Committee Draft Amendment would be issued. Michel Suignard replied that this would happen soon, certainly before the next UTC meeting.
Peter Constable clarified that the UTC would not be taking any formal decision to define the Unicode 19.0 repertoire at this meeting, but that it would do so at the next meeting. Michel Suignard noted that the preliminary repertoire would be used for the Committee Draft of the Amendment synchronized with 19, and thus is effectively a draft of the 19 repertoire. Ken Whistler stated that the pipeline would reflect this informal status, but that there was no need for the UTC to invent a new process to go on record making this distinction formally.
Ken Lunde noted that this meeting would in fact consider proposals to target some characters for Unicode Version 19.
Roll-call adjustment: Ebanah (associate member) no longer present (4.5 voting members represented).
Lunch break 12:32–13:45 UTC−7 (21:32–22:45 UTC+2) .
Discussion. Michelle Perham, Gregg Lechner, and Maralina Caldas introduced themselves.
Gregg Lechner thanked Ken Whistler for many long and instructive conversations. Maralina Caldas noted that she had a background in linguistics, and would unlikely be able to contribute to the technical parts of the work of the UTC.
Peter Constable and Ken Whistler thanked Michelle Perham and Gregg Lechner for their support to UTC work. Ken Whistler mentioned that he would soon explain to Gregg Lechner what happens when you press the key “A” on a computer, and was not sure how many hours the explanation would take. Asmus Freytag stated that this is a very good question to ask to learn about many systems, from operating systems to fonts and rendering engines. Ken Whistler, who holds degrees in linguistics, and Asmus Freytag, who holds degrees in physics, pointed out that a computer science education is not required to participate in the technical work of the UTC.
Peter Constable explained that the UTC, being a standardization committee, is the Unicode TC that most closely interfaces with other standardization agencies, and that it is where the technical group procedures are most visible. He also clarified that, as an industry consortium, The Unicode Consortium produces industry standards, not national standards.
Asmus Freytag compared his experience in ICANN with his experience in Unicode, and noted that it is useful to be able to associate changes to standardized data with the underlying documentation. Ken Whistler cautioned that this is a daunting amount of work. Markus Scherer pointed out that Robin Leroy, building upon the work of Peter Constable, had started to build tooling to provide such paper trails.
Toral Cowieson reminded participants about the Unicode Technology Workshop in Nancy in October. Maralina Caldas provided a link to the program, https://www.unicode.org/events/utw/2026/program/.
Jan Kučera presented document L2/26-158.
Roozbeh Pournader clarified that the Arabic model discussed in the report means that characters that appear to be decomposable do not have a decomposition: characters with a dot are not encoded using a combining dot.
Peter Constable asked whether there had been difficulties in Arabic arising from this model. Roozbeh Pournader replied that Arabic has been used for a long time, in many countries and languages, which means that lots of variously-dotted variants of letters keep being discovered and encoded, and that as a result Mark Davis had once suggested that it might have been better to encode the dotted letters as composites. In the case of Koré Sébèli, the application of the dots appears to be irregular.
Peter Constable also asked about the recency of the invention of the script, and about the risk of instability if the script might still be evolving. Roozbeh Pournader replied that the criterion is that the script is used widely enough to require digital interchange. Comparing the script at hand to Adlam, Roozbeh Pournader noted that Adlam had already gone through one major revision when it was encoded, but that this was not the case here, and that there did not appear to be any signs of divergence.
Asmus Freytag also remarked that some scripts fail to be adopted. Ken Whistler noted that in such cases, the script becomes historic, and if it has been adopted widely enough, there is an ample historical corpus that warrants interchange even once it becomes historic.
Anushah Hossain stated that a major source of instability of newly invented scripts is the intent of script inventors to reform them. In the absence of such intent, script evolution usually happens on a long enough time scale not to be a problem.
Michel Suignard commented on the character names, asking why characters are proposed as LOWERCASE instead of SMALL LETTER. Roozbeh Pournader replied that not enough attention was given to the character names in the script encoding working group, and proposed asking the proposal authors to revise the document accordingly, deferring the decision to provisionally assign the characters to the next meeting.
Peter Constable clarified that provisional assignment pertains to the assignment of names and code points (but not other properties).
Discussion of the IP issues mentioned in L2/26-158. It was noted that according to L2/26-055R, the inventor of the script would be providing the font, under the CLA. However, the inventor of the numerals is not; the situation of the digits should be clarified before the UTC takes action on this script.
Discussion. The IP situation should be clarified before UTC takes action on this script.
Jan Kučera presented this section of the document, for which no action was needed.
[188-C4] Consensus: Provisionally assign 1 code point U+0CF4 KANNADA SIGN COMBINING ANUSVARA ABOVE in the Kannada block, described in L2/26-170, as amended. [Ref: 2.1 in L2/26-158]
[188-A7] Action Item for Ken Whistler, RMG: Update the Pipeline to include 1 code point U+0CF4 KANNADA SIGN COMBINING ANUSVARA ABOVE, described in L2/26-170, as amended. [Ref: 2.1 in L2/26-158]
[188-A8] Action Item for Deborah Anderson, SAH: Work with the proposal authors to suggest a subheading for NamesList.txt. [Ref: 2.1 in L2/26-158]
[188-C5] Consensus: Provisionally assign 1 code point U+0CF0 KANNADA SIGN ARDHANUSVARA in the Kannada block, as described in L2/26-169. [Ref: 2.2 in L2/26-158]
[188-A9] Action Item for Ken Whistler, RMG: Update the Pipeline to include 1 code point U+0CF0 KANNADA SIGN ARDHANUSVARA, as described in L2/26-169. [Ref: 2.2 in L2/26-158]
[188-A10] Action Item for Jan Kučera, SAH: Provide the font from SEW/26-043 to the Charts group. [Ref: 2.2 in L2/26-158]
Roll-call adjustment: SIL (associate member) no longer present (4.5 voting members represented).
Break & Q&A 15:13–15:35 UTC−7 (00:13–00:35 UTC+2) .
[188-C6] Consensus: Change the code point provisionally assigned to TULU-TIGALARI LETTER ARCHAIC II by 183-C5 from U+113DA to U+113DB. [Ref: 2.3 in L2/26-158]
[188-C7] Consensus: Provisionally assign 2 code points U+113D9..U+113DA in the Tulu-Tigalari block, as described in L2/26-168. [Ref: 2.3 in L2/26-158]
[188-A11] Action Item for Ken Whistler, RMG: Update the Pipeline to include 2 code points U+113D9..U+113DA, as described in L2/26-168, moving the existing provisional assignment of U+113DA to U+113DB. [Ref: 2.3 in L2/26-158]
[188-A12] Action Item for Jan Kučera, SAH: Provide the font from SEW/26-042 to the Charts group. [Ref: 2.3 in L2/26-158]
[188-C8] Consensus: Provisionally assign 1 code point U+11462 NEWA LETTER HEAVY YA in the Newa block, as described in L2/26-171. [Ref: 2.4 in L2/26-158]
[188-A13] Action Item for Ken Whistler, RMG: Update the Pipeline to include 1 code point U+11462 NEWA LETTER HEAVY YA, as described in L2/26-171. [Ref: 2.4 in L2/26-158]
[188-A14] Action Item for Jan Kučera, SAH: Provide the font from SEW/26-051 to the Charts group. [Ref: 2.4 in L2/26-158]
Peter Constable noted that the proposals were entered in the registry four years ago, and inquired whether the reason for the delay was the concern about the impact of the encoding model described in L2/26-158.
Roozbeh Pournader replied that this was the case, and that once assurances were obtained that the encoding model posed no major issues for fonts, shaping engines, or line breaking, the only possible issue identified with the model was that the subjoined consonant would be its own grapheme cluster.
[188-C9] Consensus: Provisionally assign 1 code point U+1107E BRAHMI MINIMIZING VIRAMA in the Brahmi block, described in L2/22-219, with code point as amended in this decision. [Ref: 2.5 in L2/26-158]
[188-A15] Action Item for Ken Whistler, RMG: Update the Pipeline to include 1 code point U+1107E BRAHMI MINIMIZING VIRAMA, described in L2/22-219, with code point as amended in this decision. [Ref: 2.5 in L2/26-158]
[188-A16] Action Item for Jan Kučera, EDC: Update the Core Specification with description and illustration of the minimizing virama in Brahmi. [Ref: 2.5 in L2/26-158]
Roozbeh Pournader noted that the names list should not merely note that they are used in Luri, but that they are used in an unusual orthography of that language.
[188-C10] Consensus: Provisionally assign 3 code points U+10D90 ARABIC LETTER ALEF WITH INVERTED V, U+10D91 ARABIC LETTER KEHEH WITH INVERTED V, U+10D92 ARABIC LETTER LAM WITH INVERTED V in a new Arabic Extended-D block at U+10D90..U+10E5F, described in L2/26-053R, and with code points as amended in this decision. [Ref: 2.6 in L2/26-158]
[188-A17] Action Item for Ken Whistler, RMG: Update the Pipeline to include 3 code points U+10D90..U+10D92 in a new Arabic Extended-D block: 10D90..10E5F, described in L2/26-053R, as amended. [Ref: 2.6 in L2/26-158]
[188-A18] Action Item for Lorna Evans, SAH: Provide a compliant font for Arabic additions for Luri. [Ref: 2.6 in L2/26-158]
Roozbeh Pournader noted that this is a sexagesimal zero which came from Babylonian through Greek U+1018A 𐆊 GREEK ZERO SIGN. Robin Leroy noted that the proposal for that character, L2/04-054R, draws a parallel to cuneiform 𒑱, later encoded as U+12471 CUNEIFORM PUNCTUATION SIGN VERTICAL COLON.
Ali Basit asked whether the character is in modern use, or only in historical use. Roozbeh Pournader replied that the character appears to be used in seminaries in Iran where traditional astronomy is taught.
[188-C11] Consensus: Provisionally assign 1 code point U+1EEE0 ARABIC ASTRONOMICAL ZERO in the Arabic Mathematical Alphabetic Symbols block, as described in L2/26-172. [Ref: 2.7 in L2/26-158]
[188-A19] Action Item for Ken Whistler, RMG: Update the Pipeline to include 1 code point U+1EEE0 ARABIC ASTRONOMICAL ZERO, as described in L2/26-172. [Ref: 2.7 in L2/26-158]
[188-A20] Action Item for Jan Kučera, SAH: Provide the font from SEW/26-066 to the Charts group. [Ref: 2.7 in L2/26-158]
[188-C12] Consensus: Provisionally assign 7 code points U+10EC8 and U+10EEF in the Arabic Extended-C block and U+1EEE1..U+1EEE5 in the Arabic Mathematical Alphabetic Symbols block, as described in L2/26-173. [Ref: 2.8 in L2/26-158]
[188-A21] Action Item for Ken Whistler, RMG: Update the Pipeline to include 7 code points U+10EC8, U+10EEF, U+1EEE1..U+1EEE5, as described in L2/26-173. [Ref: 2.8 in L2/26-158]
[188-A22] Action Item for Jan Kučera, SAH: Provide the font from SEW/26-067 to the Charts group. [Ref: 2.8 in L2/26-158]
[188-C13] Consensus: Provisionally assign 2 code points U+20F1..U+20F2 in the Combining Diacritical Marks for Symbols block, as described in L2/26-141. [Ref: 2.9 in L2/26-158]
[188-A23] Action Item for Ken Whistler, RMG: Update the Pipeline to include 2 code points U+20F1..U+20F2, as described in L2/26-141. [Ref: 2.9 in L2/26-158]
[188-C14] Consensus: Provisionally assign 2 code points U+A7DE LATIN CAPITAL LETTER I WITH OBLIQUE STROKE and U+A7DF LATIN SMALL LETTER I WITH OBLIQUE STROKE in the Latin Extended-D block, as described in L2/26-144. [Ref: 2.10 in L2/26-158]
[188-A24] Action Item for Ken Whistler, RMG: Update the Pipeline to include 2 code points U+A7DE LATIN CAPITAL LETTER I WITH OBLIQUE STROKE and U+A7DF LATIN SMALL LETTER I WITH OBLIQUE STROKE, as described in L2/26-144. [Ref: 2.10 in L2/26-158]
Roozbeh Pournader mentioned that the casing for these alphabets is not necessarily consistent between the various iterations of the alphabet; clear case pairs are handled by default casing, but correct casing for any version of the alphabet will require tailoring. Peter Constable asked whether this could be a concern when assigning properties. Roozbeh Pournader replied that this would be straightforward: the case pairs would be case pairs, and the characters with unclear casing would be lowercase-only or uppercase-only.
[188-C15] Consensus: Provisionally assign 9 code points U+1DF9B..U+1DFA3 in the Latin Extended-G block, as described in L2/26-175. [Ref: 2.11 in L2/26-158]
[188-A25] Action Item for Ken Whistler, RMG: Update the Pipeline to include 9 code points U+1DF9B..U+1DFA3, as described in L2/26-175. [Ref: 2.11 in L2/26-158]
[188-A26] Action Item for Jan Kučera, SAH: Provide the font from SEW/26-056 to the Charts group. [Ref: 2.11 in L2/26-158]
[188-A27] Action Item for Kirk Miller, Ken Whistler, EDC: Provide text for the core specification describing the casing of the Albanian Frashëri characters.
Jan Kučera presented this section of the document, for which no action was needed.
Peter Constable asked whether the proposal authors were being asked to provide a rebuttal to a position that the notation is higher-level markup. Jan Kučera replied that the authors are being asked whether the currently encoded characters are useful, and whether the new ones would solve the problems faced by the users.
Markus Scherer inquired about the impact of these characters on search and collation. Roozbeh Pournader noted that these are used in religious texts where collation is rarely a primary concern. Ken Whistler clarified that the characters are ignored in default collation.
Jan Kučera presented this section of the document, for which no action was needed.
Jan Kučera presented this section of the document, for which no action was needed.
Jan Kučera presented this section of the document, for which no action was needed.
Ned Holbrook pointed out a deficiency in the language of the referenced FAQ entry.
[188-A28] Action Item for Ned Holbrook, EDC: Update the FAQ to avoid referring to presentation forms that do not exist.
Jan Kučera presented this section of the document, for which no action was needed.
Jan Kučera presented this section of the document, for which no action was needed.
Ken Whistler noted that tracking variation sequences for provisionally assigned characters is a novel process, and suggested that the approval of the variation sequences be deferred until the characters are approved for a specific version of the standard. Asmus Freytag replied that it is convenient for tracking to retain the unity of the proposal, instead of having to remember to bring back the variation sequences once the characters are approved.
[188-C16] Consensus: Provisionally assign 54 code points U+1DC00..U+1DC35 in a new Leibnizian Ambiguous Signs block at U+1DC00..U+1DC3F, as described in L2/26-143. Further, provisionally assign 5 variation sequences for Leibnizian Ambiguous Signs, described in L2/26-143, with descriptions as amended here: 1DC09 FE00 horizontal strokes with wider space; 1DC0A FE00 horizontal strokes with wide lower space; 1DC0D FE00 horizontal strokes with wider spaces; 1DC12 FE00 horizontal strokes equally widely spaced; 1DC12 FE01 horizontal strokes equally narrowly spaced. [Ref: 3.1 in L2/26-158]
[188-A29] Action Item for Ken Whistler, RMG: Update the Pipeline to include 54 code points U+1DC00..U+1DC35 in a new Leibnizian Ambiguous Signs block: 1DC00..1DC3F and the associated 5 variation sequences, as described in L2/26-143. [Ref: 3.1 in L2/26-158]
[188-A30] Action Item for Deborah Anderson, SAH: Work with the proposal authors to submit a compliant font for Leibniz Ambiguous Signs to Unicode. [Ref: 3.1 in L2/26-158]
Ken Whistler asked why the omicron upsilon was not unified with U+0223 ȣ LATIN SMALL LETTER OU, which is Latin, but derived from Greek. Peter Constable replied that the histories of the characters are different, and that it is used as a mathematical symbol, rather than orthographically; an earlier iteration of the proposal identified it as a Greek ligature, but the script encoding working group recommended against that.
Asmus Freytag added that the character is Greek, rather than Latin. Ken Whistler noted that in its latest iteration, the character is a symbol, stripped of this Greek identity. Asmus Freytag noted that the character is more similar to Greek-derived mathematical symbols. Ken Whistler remarked that this character would be confusable with U+0223.
The script encoding working group should further consider the name and script property of the omicron upsilon symbol.
[188-C17] Consensus: Accept 1 variation sequence for MATHEMATICAL SCRIPT SMALL X, as described in L2/26-142, for Unicode Version 19.0. [Ref: 3.2 in L2/26-158]
[188-A31] Action Item for Ken Whistler, RMG: Update the Pipeline for the variation sequence for MATHEMATICAL SCRIPT SMALL X as described in L2/26-142. [Ref: 3.2 in L2/26-158]
[188-A32] Action Item for Deborah Anderson, SAH: Work with the proposal authors to submit a compliant font for the variation sequence for MATHEMATICAL SCRIPT SMALL X as described in L2/26-142. [Ref: 3.2 in L2/26-158]
Ken Whistler asked why the proportion with one stroke is not unifiable with 22A4 ⊤ DOWN TACK. Robin Leroy replied that the placement on the math axis, shown on p. 2 of the proposal, which in turns stems from the function of the proportion with one stroke as a relational operator, requires that they be disunified. Asmus Freytag added that the character is wider than the down tack.
The proposal should be amended to explain this disunification from 22A4 ⊤ DOWN TACK. Ned Holbrook noted that this would be helpful for font developers so that they do not use the same glyph for both.
[188-C18] Consensus: Provisionally assign 1 code point U+1D282 MUSICAL SYMBOL DOWN ARROWHEAD BREATH MARK in the Musical Symbols Supplement block, described in L2/26-130, with the name as amended in this decision. [Ref: 3.4 in L2/26-158]
[188-A33] Action Item for Ken Whistler, RMG: Update the Pipeline to include 1 code point U+1D282 MUSICAL SYMBOL DOWN ARROWHEAD BREATH MARK, described in L2/26-130, as amended. [Ref: 3.4 in L2/26-158]
Meeting adjourned for the day at 17:28 UTC−7 (02:28 UTC+2) .
Meeting opened at 09:33 UTC−7 (18:33 UTC+2) .
4.5 voting members represented: Adobe, Apple, Google, Microsoft, UCB.
Associate members present: SIL.
Jennifer Daniel presented document L2/26-156R.
Jan Kučera asked whether the changes to chart images shown in the document would be reflected in the charts. Jennifer Daniel replied that they would be reflected in the emoji charts. Peter Constable clarified that the monochrome code charts would not be affected.
Ali Basit asked whether the cracked face was smiling. Jennifer Daniel replied that this was the case, and that this was intentional.
Ned Holbrook presented Section 2 of L2/26-156R, and stated that the first feedback item ID20260417115325 on PRI #543 would be considered as part of the report from the properties and algorithms working group, and that ID20260419151046 was editorial.
Jennifer Daniel presented Section 3 of L2/26-156R, and noted that Dr. Alexander Robertson, who conducted the survey, would be presented at the Unicode Technology Workshop in October. Peter Constable asked which of the “pure meaning only” and “includes brand bias” points in the chart on p. 2 of the document had been corrected for bias. Jennifer Daniel replied that “pure meaning only” (green) is the one that has been corrected for bias.
Jennifer Daniel presented Section 4 of L2/26-156R. Roozbeh Pournader asked whether the expectation was still to send five to ten emoji for the year to UTC. Jennifer Daniel replied in the affirmative.
Peter Constable asked whether the Emoji Standard & Research Working Group still maintains a list of possible candidates. Jennifer Daniel replied that this is the case, and that the “larger list” currently contains around 40 candidates. Peter Constable asked whether that larger list is growing over time, or whether it is pruned. Jennifer Daniel replied that it does get pruned, and grows and shrinks over time with no longer-term growth trend.
Debbie Anderson noted that she and Jan Kučera, as leaders of the Script Encoding Working Group, had been invited to observe the process of the Emoji Standard & Research Working Group, and thanked Jennifer Daniel for that opportunity.
Ned Holbrook noted that changes to UTS #51 would be discussed as part of the properties and algorithms discussion (F.1 Section 13.1).
The UTC started discussing this topic earlier than scheduled. Peter Constable stated that he had not received a report from Fraser Gordon, the chair of the text terminal working group, but that he was aware that the working group was still active.
Roozbeh Pournader stated that when the chair is available, the group meets for about one hour a month, and that the meetings are productive. The key architecture is to assign widths to strings, and to adjust these widths where the fonts do not support that. The work includes defining tables and algorithms to determine the widths of strings, based on East_Asian_Width, Default_Ignorable_Code_Point, etc. Roozbeh Pournader noted that the working draft data files make use of Unicode Set notation, and that they do not yet support South Asian scripts.
Roozbeh Pournader suggested that a vice chair be appointed for this group.
Peter Constable provided general context on the problems that the text terminal working group is tasked with solving (see 175-C12).
Ken Whistler asked whether a derived property would be created for this algorithm. Roozbeh Pournader replied that while much of the data would be based on existing properties, additional data and exceptions would be needed, but that it might not take the form of a property, although new properties might be proposed to facilitate defining the data.
Peter Constable asked whether escape sequences for applications to interact with terminals were still being discussed. Roozbeh Pournader replied that this was the case, in particular for versioning questions, although nothing concrete had emerged.
Roozbeh Pournader presented a working draft data file in the form of a sequence of regular expressions associated with widths, the idea being that the first matching line would apply.
Fraser Gordon joined the meeting at the scheduled time and summarized the current status of the work.
The group has moved on from the high level design of the algorithm, and onto producing the actual data.
The issue of changing width as text is modified does not appear to be a problem, as terminals have long had to deal with double-width characters in East Asian character sets.
Liang Hai asked whether it would be helpful to appoint a vice chair to improve communication between the working group and the UTC. Fraser Gordon replied that this could be useful.
Markus Scherer reported on the activity of the design working group of CLDR, and noted that this group had engaged with many edge cases in the CLDR specification, which for a long time had ICU as its sole implementation. The independent implementation by ICU4X has been discovering many edge cases in the specification.
Markus Scherer noted that few people are contributing to formatting code in ICU at the moment, so the formatting libraries sometimes lag behind the specification for newer features.
Markus Scherer remarked on the influx of AI-generated bug reports and fixes, which often attempt to address issues that are out of scope, and are often of poor quality.
Markus Scherer also noted ongoing work on UnicodeSet arising from the specification in DUTS #61, and on MessageFormat v2.
Asmus Freytag reported that ICANN has just released a set of string similarity data, which loosely corresponds to what Unicode does as confusables, the difference being that ICANN works in the DNS root zone, which is limited to the Recommended scripts (in the categorization of UAX #31).
Asmus Freytag noted that it could be interesting to compare this data with our confusables data, but that the ICANN data unfortunately appear to lack terms of use.
Robin Leroy reported on Unicode-related activity in several Working Groups of ISO/IEC JTC 1/SC 22.
WG 21 (C++) adopted a defect report (https://github.com/cplusplus/papers/issues/2288) based on the updated recommendations about programming language identifiers in Unicode Version 15.1.0 and Unicode Technical Standard #55. Robin Leroy stated that he would be working with WG 14 (C) to make the same change in C, as recommended by WG 21/SG 22 (Study Group 22, C/C++ liaison).
Discussions are ongoing within WG 21/SG 16 (Unicode) on the appropriate behaviour of std::format on std::path objects on Windows in those cases where the paths are not text because they contain unpaired surrogates.
In WG 9 (Ada), work has been ongoing to define a Text type to supersede the nine types of string: fixed-size array, Bounded, and Unbounded strings of Character (Latin-1), Wide_Character (UTF-16), and Wide_Wide_Character (UTF-32). The ARG (Ada Rapporteur Group) has approved the intent of http://www.ada-auth.org/cgi-bin/cvsweb.cgi/ai22s/ai22-0161-2.html.
Peter Constable explained that an issue in wording had come up in the context of the relationship between ISO 15924/RA and ISO 639/MA: a rep from ISO CS noted that the ISO 639/MA Terms of Reference do not explicitly allow the maintenance agency to appoint a liaison representative to ISO 15924/RA who is not part of its voting group. But it was unclear to others if that needed to be stated explicitly.
Peter Constable reported that he attended meetings of TC 37/SC 4 (Language resource management) and its working groups, where much of the work was academically motivated but not obviously relevant for the tech industry.
Peter Constable noted the creation of a joint working group (ISO/IEC JTC 1/SC 42/JWG 5) between ISO/TC 37 and ISO/IEC JTC 1/SC 42. Peter Constable noted that a draft standard being developed by JWG 5 that had been presented as required by the EU AI Act was not in fact required by it.
Related to ISO 639, Peter Constable also reported that an IETF process had been initiated to appoint a new Language Subtag Reviewer according to RFC 5646 (BCP 47): Doug Ewell is the new Language Subtag Reviewer, replacing Michael Everson who served in that role for many years.
A question was raised about the relationship between ISO 639 and BCP 47. Peter Constable replied that BCP 47 makes normative reference to ISO 639 (for language subtags), ISO 15924 (for script subtags), as well as ISO 3166 and UN M.49 (for region subtags).
Break 10:57–11:12 UTC−7 (19:57–20:12 UTC+2)
Markus Scherer presented document L2/26-154.
Peter Constable provided additional context. In January, it came to the attention of the UTC that important work from C2PA on the provenance of information had resulted in a C2PA recommendation to use variation selectors to encode provenance manifests within supposedly plain text. The UTC had noted that this caused it to no longer be plain text. Peter Constable remarked that the stated motivation was Article 50, Clause 2 of the EU AI Act. The EU AI office has developed a Code of Practice for AIA enforcement, which explicitly says that metadata cannot be embedded in plain text. For “free-form text” above a minimum length, the Code of Practice requires use of watermarking techniques that utilize visible characters (e.g., SynthID-Text).
Peter Constable noted that standardized misuse of Variation Selectors would be a concern for SC 2, and that he had discussed the matter with the chair of SC 2, who would follow up with other committees under JTC 1.
Markus Scherer noted that a representative from C2PA was present at the January meeting, and that their interpretation of the standard was that their use of variation selectors was conformant, which resulted in the changes made in the draft of The Unicode Standard, Version 18.0.
Markus Scherer also noted that additional language was planned for tag characters, and generally default ignorable code points.
Roozbeh Pournader reported that some users of the C2PA specifications expressed interest in the ability to watermark plain text beyond the mandate of the EU AI Act.
Peter Constable explained that some techniques that can be used to watermark plain text can also be used as an attack vector, in particular against LLMs, so that they might be removed by implementers as a security risk; and also that it is clear in C2PA specifications that provenance metadata can be removed.
[188-C19] Consensus: Add notes to section 3.6.2 "Variation Selectors" in the Core spec to indicate that interpreting either misplaced variation selectors or unassigned variation sequences in any way not sanctioned by the Unicode Standard is not conformant. For Unicode Version 18.0. See L2/26-154 item 1.1.
[188-A34] Action Item for Asmus Freytag, EDC: Add notes to section 3.6.2 "Variation Selectors" in the Core spec to indicate that interpreting either misplaced variation selectors or unassigned variation sequences in any way not sanctioned by the Unicode Standard is not conformant. For Unicode Version 18.0. See L2/26-154 item 1.1.
Markus Scherer presented this section of the document, for which no action was needed.
Jan Kučera asked whether the recommendation was to change text that people did not read to other text that they might also not read. Ned Holbrook replied that the large amount of prose involved in the definition is unusual, but calling out up front that the file does not define the properties should help. Liang Hai stated that he was not aware of the special situation of ArabicShaping.txt, and expressed his support for the proposed change.
Peter Constable asked whether UAX #44 should mention this. Roozbeh Pournader replied that it already does.
Peter Constable asked whether the proposed text in ArabicShaping.txt should point to the relevant section of UAX #44. Markus Scherer replied that this was not necessary.
[188-C20] Consensus: Change the language and recommendations in ArabicShaping.txt to more strongly recommend that implementers use the derived files. For Unicode Version 18.0. See L2/26-154 item 2.2.
[188-A35] Action Item for Roozbeh Pournader and Robin Leroy, PAG: Change the language and recommendations in ArabicShaping.txt to more strongly recommend that implementers use the derived files. For Unicode Version 18.0 See L2/26-154 item 2.2.
[188-A36] Action Item for Roozbeh Pournader, Ken Whistler, PAG: Correct the description of ArabicShaping.txt in Table 9 of UAX #44, Unicode Character Database, for Unicode Version 18.0.
[188-C21] Consensus: Change the Indic_Syllabic_Category of U+16100 GURUNG KHEMA LETTER A from Vowel_Independent to Consonant. For Unicode Version 18.0. See L2/24-203 and L2/26-154 item 2.3.
[188-A37] Action Item for Robin Leroy, PAG: In UCD file IndicSyllabicCategory.txt, Change the Indic_Syllabic_Category of U+16100 GURUNG KHEMA LETTER A from Vowel_Independent to Consonant. For Unicode Version 18.0. See L2/24-203 and L2/26-154 item 2.3.
Robin Leroy clarified that the invariant test referenced in L2/26-154 item 2.3 checks for each script that the number of characters with InSC=Independent_Vowel in that script is not 1 (it may be 0, or it may be greater than 1).
Markus Scherer presented this section of the document, for which no action was needed.
Markus Scherer presented this section of the document, for which no action was needed.
Robin Leroy noted the particularly interesting case of the bidi class of the ambiguous signs as a good example of the difficulty of assigning properties for new characters, and of the need to determine the reasons for the property assignments of existing characters.
Peter Constable asked Andrew Glass for an update on the keyboards working group.
Andrew Glass stated that the working group had started with revising the keyboard specification in LDML https://www.unicode.org/reports/tr35/tr35-keyboards.html#Contents (resulting in “keyboard v3”). It is now sufficient to round-trip the definition of a keyboard.
The next step is to connect language communities with people familiar with keyboard v3, to aid onboarding, and to connect with platform implementers so that they can use keyboard definitions in that format.
Andrew Glass noted that there would be a tutorial (two 90-minute sessions) at the upcoming Unicode Technology Workshop in Nancy in October.
Andrew Glass stated that work had been conducted in coordination with Keyman to use it as an editor for keyboard v3 definitions.
Roozbeh Pournader asked whether Andrew Glass was expecting more implementers to come forward than Keyman. Andrew Glass replied that any platform offering a keyboard could in principle directly consume a keyboard v3 file.
Liang Hai asked whether there had been any work on an Android implementation. Andrew Glass replied that he had no specific information.
Peter Constable asked about the transforms, and their impact on the encoding of South East Asian scripts.
Andrew Glass explained that everything can be done with simple transforms, but for some visual-to-logical order transformations, it is possible to represent the transforms in a more compact form. Where the encoding uses a visual order, but the expected typing order is logical, reordering transforms can be used to transform one into the other. This enables us to make encoding decisions based on what is beneficial for the orthography and writing system as a whole, and not make the transform rules steer the encoding.
Andrew Glass noted that this technology was not available when the Myanmar script was encoded, which led users to construct the Zawgyi hack encoding.
Jan Kučera clarified that the reordering transforms were standardized before v3, and taken as-is. Jan Kučera suggested a possibility to standardize the per-script reordering rules separately from the keyboard specification.
Roozbeh Pournader asked whether there was work on the standardization of backspacing behaviour. Andrew Glass replied that there is a special kind of transform for backspacing. Whether it is supported is platform-dependent. Implementations that give access to the text buffer can provide natural backspacing.
Liang Hai asked whether the working group was planning to provide a reference implementation, in particular with respect to parsing. Andrew Glass replied that the de facto reference implementation is the Keyman implementation.
Lunch break 12:29–14:48 UTC−7 (21:29–23:48 UTC+2) .
[188-A38] Action Item for Robin Leroy, PAG: Update the text of UAX #9 as proposed in L2/26-154 item 5.1. For Unicode Version 18.0. See L2/26-154 item 5.1.
[188-C22] Consensus: Change the Bidi_Class of U+1B42 ᭂ BALINESE VOWEL SIGN PEPET, U+1B3C ᬼ BALINESE VOWEL SIGN LA LENGA, and U+1B3A ᬺ BALINESE VOWEL SIGN RA REPA from Nonspacing_Mark (NSM) to Left_To_Right (L). For Unicode Version 18.0. See L2/26-154 item 5.2.
[188-A39] Action Item for Robin Leroy, PAG: In UCD file UnicodeData.txt and derived files, change the Bidi_Class of U+1B42, U+1B3C, and U+1B3A from Nonspacing_Mark (NSM) to Left_To_Right (L). For Unicode Version 18.0. See L2/26-154 item 5.2.
[188-N1] Note: In Unicode Versions 5.0 through 17.0, the Bidi_Class values of U+1B42 ᭂ BALINESE VOWEL SIGN PEPET, U+1B3C ᬼ BALINESE VOWEL SIGN LA LENGA, and U+1B3A ᬺ BALINESE VOWEL SIGN RA REPA (all Bidi_Class=Nonspacing_Mark) were inconsistent with those of U+1B43 ᭃ BALINESE VOWEL SIGN PEPET TEDUNG, U+1B3D ᬽ BALINESE VOWEL SIGN LA LENGA TEDUNG, and U+1B3B ᬻ BALINESE VOWEL SIGN RA REPA TEDUNG (all Bidi_Class=Left_To_Right), contradicting the Property Value Stability Policy “The property values for the bidirectional properties Bidi_Class and Bidi_Mirrored preserve canonical equivalence”.
Roozbeh Pournader asked whether the applicable version for the relevant stability policy should be corrected from 4.0.0 to 18.0.0. Ken Whistler and Asmus Freytag replied that this would be unhelpful, as the intent had long been for this stability policy to apply. They suggested that a note be added to the table instead.
[188-C23] Consensus: Recommend to the TCG that the character encoding stability policy “The property values for the bidirectional properties Bidi_Class and Bidi_Mirrored preserve canonical equivalence” be updated to address the issue identified in 188-N1.
[188-A40] Action Item for Markus Scherer, PAG: Work with the TCG to update the character encoding stability policy “The property values for the bidirectional properties Bidi_Class and Bidi_Mirrored preserve canonical equivalence” to address the issue identified in 188-N1.
Markus Scherer presented this section of the document, for which no action was needed.
Roozbeh Pournader asked whether UAX #14 needed to be updated to reflect this change. Robin Leroy replied that this was not needed.
[188-C24] Consensus: Change the Line_Break assignments of U+00B7 · MIDDLE DOT from Ambiguous (AI) to Infix_Numeric (IS) and U+0387 · GREEK ANO TELEIA from Alphabetic (AL) to Infix_Numeric (IS). For Unicode Version 18.0. See L2/26-154 item 7.1.
[188-A41] Action Item for Robin Leroy, PAG: In UCD file LineBreak.txt and derived files, change the Line_Break assignments of U+00B7 · MIDDLE DOT from Ambiguous (AI) to Infix_Numeric (IS) and U+0387 · GREEK ANO TELEIA from Alphabetic (AL) to Infix_Numeric (IS). For Unicode Version 18.0. See L2/26-154 item 7.1.
[188-A42] Action Item for Robin Leroy, PAG: In UCD file LineBreakTest.txt and derived files, add test cases for Chinese usage of U+00B7. For Unicode Version 18.0. See L2/26-154 item 7.1.
[188-C25] Consensus: Change the Line_Break assignment of code points U+1F7DB and U+1F7F1..U+1F7FF from Ideographic (ID) to Alphabetic (AL). For Unicode Version 18.0. See L2/26-154 item 7.2.
[188-A43] Action Item for Robin Leroy, PAG: In LineBreak.txt, change the Line_Break assignment of code points U+1F7DB and U+1F7F1..U+1F7FF from Ideographic (ID) to Alphabetic (AL). For Unicode Version 18.0. See L2/26-154 item 7.2.
[188-A44] Action Item for Josh Hadley, PAG: In Unicode Standard Annex #29, add a note after the second paragraph of Section 3 pointing out that character count limits based on grapheme cluster breaking are not data size limits. For Unicode Version 18.0. See L2/26-154 item 8.1.
Robin Leroy pointed out that this was an example where a technical decision is informed by historiography of the standard, and that the properties and algorithms group had been considering changing UTS #51 instead until Ned Holbrook found 171-C27.
Roozbeh Pournader noted that it was not clear that the UTC #171 had considered the implications of 171-C27 on grapheme cluster segmentation. Ned Holbrook agreed and stated that the option remains to rescind that decision; however, no concrete issue was found with it so far.
Peter Constable recalled that some decisions related to emoji were meant for forward compatibility, and asked whether this was such a case. Roozbeh Pournader replied that this did not appear to be the case.
[188-A45] Action Item for Josh Hadley, PAG: In Unicode Standard Annex #29, in Section 3.1.1, Grapheme Cluster Boundary Rules, correct the note about emoji sequences as described in L2/26-154 item 8.2. For Unicode Version 18.0.
Markus Scherer presented this section of the document, for which no action was needed.
[188-C26] Consensus: In Unicode Standard Annex #44, Unicode Character Database, update the derivation of the Indic_Conjunct_Break property to use the Script_Extensions (scx) property rather than the Script (sc) property. For Unicode Version 18.0. See L2/26-154 item 8.4.
[188-A46] Action Item for Ken Whistler, PAG: In Unicode Standard Annex #44, Unicode Character Database, update the derivation of the Indic_Conjunct_Break property to use the Script_Extensions (scx) property rather than the Script (sc) property; include a note that \p{scx=𝑆} means the set of characters that have 𝑆 as one of their Script_Extensions values. For Unicode Version 18.0. See L2/26-154 item 8.4.
[188-A47] Action Item for Robin Leroy, PAG: Update UCD data files to reflect the changed derivation of Indic_Conjunct_Break. For Unicode Version 18.0. See L2/26-154 item 8.4.
[188-A48] Action Item for Asmus Freytag, PAG: Update UAX #29 to clarify wording around the use of hyphens in special text types as described in L2/26-154 item 8.5. For Unicode Version 18.0. See L2/26-154 item 8.5.
Roozbeh Pournader noted that the properties and algorithms group discussed that this definition and surrounding ones were of limited modern use, since the actual definition of graphemes clusters is now found in UAX #29.
[188-A49] Action Item for Robin Leroy, Ken Whistler, EDC: Clean up bullet points of D59 of section 3.6.2 of the Core Spec. For Unicode Version 18.0. See L2/26-154 item 8.6.
[188-C27] Consensus: In the UCA spec, remove the Variable Weighting option Shift-Trimmed. For Unicode Version 18.0. See L2/26-154 item 9.1.
[188-A50] Action Item for Markus Scherer, PAG: In the UCA spec, remove the Variable Weighting option Shift-Trimmed. For Unicode Version 18.0. See L2/26-154 item 9.1.
[188-A51] Action Item for Mark Davis, PAG: Create a CLDR ticket to deprecate the name of the option. See L2/26-154 item 9.1.
[188-A52] Action Item for Mark Davis, Robin Leroy, PAG: Fix the regular expression referenced in L2/26-154 item 10.1 in UTS #18 in a future version.
Peter Constable asked whether the data presented in this section had been incorporated into the beta. Markus Scherer replied that it had been.
Markus Scherer presented this section of the document, for which no action was needed.
Asmus Freytag noted that in the context of domain names, these characters have all long been restricted, and welcomed the fact that the finer classification of the reasons for restriction.
A question was raised as to whether there were stability constraints on these properties. Markus Scherer replied that this was not the case. Asmus Freytag added that since these assignments are meant to reflect real usage, they cannot be stable.
[188-C28] Consensus: Update the Identifier_Type of Arabic and Hebrew combining marks per L2/26-154 item 11.2. For Unicode Version 18.0.
[188-A53] Action Item for Roozbeh Pournader and Josh Hadley, PAG: Update IdentifierType.txt to reflect the new Identifier_Type values of Arabic and Hebrew combining marks per L2/26-154 item 11.2. For Unicode Version 18.0.
Peter Constable stated that there was a proposed update for UTS #39 for Unicode Version 18.0, but that this proposed update did not include this change.
Roozbeh Pournader noted that this change implies that some currently non-single-script strings would become single-script in version 18.0.
[188-C29] Consensus: In Unicode Technical Standard #39, Unicode Security Mechanisms, take Hntl into account in the Mixed-script detection algorithm. For Unicode Version 18.0. See L2/26-154 item 11.3.
[188-A54] Action Item for Josh Hadley and Roozbeh Pournader, PAG: Add Hntl to Section 5.1 of UTS #39. For Unicode Version 18.0. See L2/26-154 item 11.3.
Asmus Freytag noted that this was a departure from the IDNA2008 CONTEXTJ rule. Roozbeh Pournader replied that we have three different rules for joiners in the core specification, and the Indic ones are already more strict than CONTEXTJ. Asmus Freytag suggested that the text of UTS #39 be amended to clarify that the restrictions are stricter than the CONTEXTJ rules.
[188-C30] Consensus: Change the regular expression in Section 3.1.1.1 of UTS #39 from /$LJ $T* ZWNJ $T* $RJ/ to /$LJ $T* ZWNJ $RJ/, for Unicode Version 18.0. See L2/26-154 item 11.4.
[188-A55] Action Item for Josh Hadley and Roozbeh Pournader, PAG: Change the regular expression in Section 3.1.1.1 of UTS #39 from /$LJ $T* ZWNJ $T* $RJ/ to /$LJ $T* ZWNJ $RJ/, for Unicode Version 18.0. See L2/26-154 item 11.4.
[188-A56] Action Item for Roozbeh Pournader, EDC: Add text to the Core Specification to recommend that combining marks should appear before joiners in Arabic and similar scripts. For Unicode Version 18.0. See L2/26-154 item 11.4.
[188-A57] Action Item for Asmus Freytag, PAG: Provide text to Roozbeh Pournader and Josh Hadley for Section 3.1.1.1 of UTS #39 stating that the restrictions on ZWJ and ZWNJ are stricter than the CONTEXTJ rules, rather than merely similar.
Break 15:23–15:36 UTC−7 (00:23–00:36 UTC+2)
Markus Scherer presented this section of the document, for which no action was needed.
[188-A58] Action Item for Roozbeh Pournader, PAG: Add a clarification note to UTS #39 about casefolding and normalization in relation to confusables. For a future version of the standard. See L2/26-154 item 12.2.
Markus Scherer presented this section of the document, for which no action was needed.
Roozbeh Pournader presented this section of the document, for which no action was needed.
Roll-call adjustment: SIL (associate member) no longer present (4.5 voting members represented).
Roozbeh Pournader presented this section of the document, for which no action was needed.
Roozbeh Pournader presented this section of the document, for which no action was needed.
Roozbeh Pournader noted that Chisoi is listed, but would be removed from published 18.0 data. The data for Chisoi remains in the source file and would reappear whenever Chisoi gets encoded.
Roozbeh Pournader presented this section of the document, for which no action was needed.
[188-A59] Action Item for Ned Holbrook, ESC: Edit the text of UTS #51 to improve guidance on the usage of emoji presentation selectors and the Emoji_Presentation property based on L2/26-154 item 13.1. For Unicode Version 18.0. See L2/26-154 item 13.1.
Ned Holbrook presented L2/26-200, which implements this action item.
Roozbeh Pournader expressed a concern that the new language in Section 2.7.1 recommended to text-only implementers that they add text presentation selectors even for characters with Emoji_Presentation=No. Asmus Freytag replied that this was only a recommendation, and that there was a desire to reduce ambiguity in all cases. Ned Holbrook explained that there is no guarantee that even characters with Emoji_Presentation=No would be displayed using the text presentation in the absence of a variation selector. Robin Leroy posted U+267E ♾ PERMANENT PAPER SIGN (Emoji_Presentation=No) in the chat, with and without a text presentation selector; the one without a text presentation selector was displayed with emoji presentation on some clients.
Roozbeh Pournader pointed out that this is a recommendation for implementers who do not care about emoji, but which only applies to implementers who claim conformance to the emoji specification. Asmus Freytag replied that this is not a conformance requirement. Roozbeh Pournader clarified that he was concerned about discoverability: if any text producing process should be adding these variation selectors, this should be mentioned in the core specification. Asmus Freytag stated that adding such a reference in the core specification was editorial.
Peter Constable clarified Roozbeh’s concern: the developer of an application designed to edit legal documents (for example) should read this recommendation, even though they are not generally interested in UTS #51.
Peter Constable asked whether the recommendation in Section 2.7.1 should, or could, be interpreted as recommending amending old documents to add variation selectors when they are opened. Asmus Freytag stated that in context, this would be a far-fetched interpretation. Peter Constable noted that earlier in the day we had discussed cases of readers following unexpected interpretations of our standards. Ken Whistler expressed his agreement with Asmus Freytag, and noted that a text editor would in general not be able to choose which variation selector to apply.
Regarding the concerns about conformance for text input processes, Ken Whistler stated that conformance was irrelevant, and noted that Chapter 22 already references UTS #51.
Roozbeh Pournader noted that there are conformance requirements about not silently modifying text (see https://unicode.org/versions/Unicode17.0.0/core-spec/chapter-3/#G2155).
Markus Scherer suggested that the wording in UTS #51 be clarified to add “on input”. Ned Holbrook suggested “according to the user’s intent”, which would imply that it is done based on the user’s request. Asmus Freytag expressed support for this wording.
Ned Holbrook noted that the suggestion in L2/26-154 to document in UTS #51 that Emoji_Presentation assignments should not be changed for pre-existing characters seemed like a matter of policy for UTC to separately decide, and would not be incorporated into the text of UTS #51.
Markus Scherer presented this section of the document, for which no action was needed.
Ken Whistler noted that the issue was a consequence of a failure of UTC to account for the implications of the consequences of decisions 147-C17 and 147-C33, arising from emoji-related changes.
Markus Scherer stated that there was no new feedback on this PRI, but that he would be preparing an update of the document.
Robin Leroy apologized for mischaracterizing a noncharacter as a character, as reported in PRI #523 feedback ID20260712083815, and stated that the error would be corrected in an updated version of the document.
Murray Sargent stated that work is ongoing on the proposed update of UTR #25. Asmus Freytag stated that there was a need to somewhat formalize the leadership of the maths interest group within the PAG, and that John Wilcock had expressed willingness to take on this role. Asmus Freytag also stated that the document ought to eventually become a synchronized UTS.
Peter Constable stated that no new feedback had been received on this PRI.
Peter Constable stated that no new feedback had been received on this PRI.
Peter Constable stated that no new feedback had been received on this PRI.
Peter Constable stated that no new feedback had been received on this PRI.
Peter Constable stated that no new feedback had been received on this PRI.
Peter Constable stated that no new feedback had been received on this PRI.
Peter Constable stated that no new feedback had been received on this PRI.
Peter Constable stated that no new feedback had been received on this PRI.
Peter Constable stated that these documents would be re-issued for Unicode Version 18.0.
Ken Whistler cautioned editors that there was no need to do the boilerplating of producing proposed updates for public review for these documents.
Peter Constable stated that all feedback had been covered as part of the properties and algorithms report L2/26-154, except for the two items listed in Section 15 of that report.
Robin Leroy briefly presented the document, and noted that he would give a talk on line breaking at the Unicode Technology Workshop in Nancy in October. A PRI has been posted and implementers are encouraged to review it.
Peter Constable noted that OpenType-aware implementations are unaffected by this issue.
Roozbeh Pournader stated that the question was how much advance warning to give to implementers.
Peter Constable clarified that unless UTC takes action, it would be following alternative 2.
Robin Leroy noted that a PRI would make it possible to call attention to the more general issue, in particular to the possibility of cross-plane case mappings.
Peter Constable stated that we could still create such a public review issue without complicating our release process for 18.0.
Ken Whistler noted that besides the cross-plane mapping, this is the first non-BMP character in BidiMirroring.txt at all, so implementations might break because of that.
Ned Holbrook asked how cross-plane case mappings could arise. Roozbeh Pournader replied that once the BMP is full, uppercase counterparts of currently lowercase-only letters in the BMP would be encoded in the SMP. Peter Constable clarified that the case pair stability policy only prohibits forming a case pair out of pre-existing characters: a case pair involving a pre-existing character can be formed when encoding the other character of the pair.
The UTC went with alternative 2 by default.
Roozbeh Pournader asked whether a migration note could be added to UAX #9. Peter Constable replied that this would be an editorial change.
[188-A60] Action Item for Robin Leroy, PAG: in UAX #9, Unicode Bidi Algorithm, add a warning to Section 7, Mirroring, that Bidi_Mirroring_Glyph mappings are not size-preserving.
Apple offers to host UTC #190 on January 20–22, 2027.
UCB offers to host UTC #191 on April 13–15, 2027.
Microsoft offers to host UTC #192 on July 27–29, 2027.
Roozbeh Pournader stated that he was not certain whether our practice was to repeat the Script value (here Tamil) in the Script_Extensions value, but that the intent of the action item was clear.
Robin Leroy noted that as the Script_Extensions property is not provisional, changing its assignment for a pre-existing character requires a decision.
[188-C31] Consensus: Add Kannada, Malayalam, and Telugu to the Script_Extensions values of U+0B83 TAMIL SIGN VISARGA, for Unicode Version 18.0. [Ref: 4.1 in L2/26-158]
[188-A61] Action Item for Roozbeh Pournader, PAG: Add scx=Knda Mlym Taml Telu to U+0B83 TAMIL SIGN VISARGA, for Unicode Version 18.0. [Ref: 4.1 in L2/26-158]
[188-C32] Consensus: Change the names of the following three Leke characters which had been provisionally assigned as a part of Consensus [186-C21]: U+11BA2 LEKE SIGN SEUNG NU CUI to LEKE SIGN SEUNG NU CEU; U+11BA3 LEKE SIGN THAEU WEE KLEU to LEKE SIGN THAE WEE KLEU; and U+11BA4 LEKE SIGN DWAI THAEU WEE to LEKE SIGN DWAI THAE WEE. [Ref: 4.2 in L2/26-158]
[188-A62] Action Item for Ken Whistler, SAH: Update the Pipeline with the revised Leke character names for U+11BA2..U+11BA4. [Ref: 4.2 in L2/26-158]
[188-A63] Action Item for Robin Leroy, PAG: Update the draft data to change the three Leke character names U+11BA2..U+11BA4, as described in the SEW report. [Ref: 4.2 in L2/26-158]
[188-A64] Action Item for Rick McGowan, Ken Whistler, EDC: Consider updating the Core Specification to point out the different conventions for compound accidentals. [Ref: 5.1 in L2/26-158]
[188-A65] Action Item for Ken Whistler, Charts: Consider adding annotations for U+1D1D2 MUSICAL SYMBOL SQUARE B, as described in L2/26-178. [Ref: 5.2 in L2/26-158]
Andrew Glass asked whether these changes affect the Noto font. Liang Hai replied that the Noto font had been updated two years ago by Kushim Jiang himself in accordance with UTN #57 and the proposal.
Michel Suignard asked whether there was a desire to show these in the code charts. Ken Whistler and Liang Hai replied that this should not be done. Liang Hai added that the code charts cannot be made useful for implementers of Mongolian, and that UTN #57 is where usable guidance can be found.
Peter Constable asked whether the variation sequences assigned in StandardizedVariants.txt as modified were still meant to be used. Liang Hai replied that this was the case, with the exception of those marked as withdrawn. Asmus Freytag asked what the definition of the variation sequences was. Ken Whistler replied that they were defined in the Chinese standard. Peter Constable noted that the data file would point to UTN #57, which in turn references the Chinese standard.
Ken Whistler suggested changing “withdrawn” to “not in use in GB/T 25914”. Asmus Freytag expressed support for this wording. Ken Whistler also stated that he intends to provide additional context in the comment. Liang Hai cautioned that three different Chinese standards are involved, and that no single one of them should be directly referenced in the data file.
Markus Scherer suggested delaying the implementation of this decision to Unicode Version 19.0. Roozbeh Pournader replied that implementations are already several years ahead of us, and that we should catch up.
Ken Whistler offered to provide a draft of StandardizedVariants.txt by the next day incorporating input from UTC discussion.
[188-A66] Action Item for Ken Whistler, Charts: Consider revising annotations for U+11300 GRANTHA SIGN COMBINING ANUSVARA ABOVE and U+11302 GRANTHA SIGN ANUSVARA, as described in L2/26-164. [Ref: 5.4 in L2/26-158]
Jan Kučera presented this section of the document, for which no action was needed.
Asmus Freytag asked whether this was for modern or historical use. The document mentions Sanskrit. Ned Holbrook stated that supporting this usage would have implications for both shaping engines and fonts.
Jan Kučera presented this section of the document, for which no action was needed.
Peter Constable asked whether this proposal should be added to the notices of non-approval. Ken Whistler replied that it did not deserve this much attention.
Meeting adjourned for the day at 17:53 UTC−7 (02:53 UTC+2) .
Meeting opened at 09:32 UTC−7 (18:32 UTC+2) .
4.5 voting members represented: Adobe, Apple, Google, Microsoft, UCB.
Associate members present: SIL.
[188-A67] Action Item for Robin Leroy, PAG: Investigate the various encoding models for Indic scripts listed in the Core Specification and propose appropriate changes, as discussed in PRI #544 ID20260416070531. [Ref: 6.1 in L2/26-158]
Jan Kučera presented this section of the document, for which no action was needed.
Jan Kučera presented this section of the document, for which no action was needed.
Jan Kučera presented this section of the document, for which no action was needed.
Jan Kučera presented this section of the document, for which no action was needed.
Jan Kučera presented this section of the document, for which no action was needed.
Jan Kučera presented this section of the document, for which no action was needed.
Jan Kučera presented this section of the document, for which no action was needed.
Ken Whistler presented an early draft of L2/26-203.
Ken Whistler explained that in that draft, he had added a number of notes in the comments above the list of Mongolian variation sequences, and reworded “withdrawn” to “not in use”.
Short discussion of the wording of the notes, resulting in the version of L2/26-203 in the document register.
[188-C33] Consensus: Update the Mongolian section of StandardizedVariants.txt as discussed in the meeting, for Unicode Version 18.0. [Ref: L2/26-203 and 5.3 in L2/26-158]
[188-A68] Action Item for Ken Whistler, PAG: Make changes to StandardizedVariants.txt per 188-C33, for Unicode Version 18.0.
[188-A69] Action Item for Liang Hai, EDC: Consider updating Chapter 13 and Chapter 23 of the Core Specification regarding variation sequences for Mongolian, as described in L2/26-163. [Ref: 5.3 in L2/26-158]
Peter Constable stated that all feedback had been covered as part of L2/26-158.
Discussion of the title of UAX #60. Ken Whistler noted that such changes are particularly disruptive, and that we tend to miss references to outdated titles.
[188-C34] Consensus: UTC approves the change of title for UAX #60 from “Data for Large East Asian Scripts” to “Data for East Asian Scripts”, for Unicode Version 18.0.
[188-A70] Action Item for Michel Suignard, SAH: Change the title of UAX #60 to “Data for East Asian Scripts”, and include a reference to UAX #38, for Unicode Version 18.0.
[188-A71] Action Item for Ken Whistler, RMG: Update any references to outdated titles of UAX #60, for Unicode Version 18.0.
Michel Suignard presented document L2/26-199.
Michel Suignard explained that he is moving from using Gardiner-style syntax to zero-padded numbers as used in kEH_UniK as requested by the user community. The kEH_Desc values describe the components, but they sometimes have details missing because they are too small.
Michel Suignard also stated that he planned to cleanly make the kEH_Func and kEH_FVal multi-valued in a future version of standard.
Robin Leroy noted that most of the source value changes were explained in L2/26-060.
Michel Suignard noted that the UniHieroglyphica font is available under the SIL Open Font License from the Thesaurus Linguae Aegyptiae (TLA) project.
Michel Suignard explained that, following the ad hoc discussions during lunch over the first two days, the wording of UAX #57 would be simplified.
Peter Constable stated that no new feedback had been received on this PRI.
Peter Constable stated that all feedback had been covered as part of L2/26-158.
Break 10:27–10:33 UTC−7 (19:27–19:33 UTC+2)
Ken Lunde presented document L2/26-157.
[188-A72] Action Item for Ken Lunde, CJK: Update the Unihan database to add “I:U+82A9” as the kStrange property value of U+2B1E7, based on PRI #534 feedback [ID20260517103203] and Section 01 of document L2/26-157, for Unicode Version 18.0.
[188-A73] Action Item for Ken Lunde, CJK: Convey to the feedback submitter the CJK & Unihan Working Group comments in Section 01 of document L2/26-157. See PRI #534 feedback [ID20260517103203].
[188-C35] Consensus: Accept the informative kRSUnicode property value change for U+2BB89, based on PRI #534 feedback [ID20260517144156] and Section 02 of document L2/26-157, for Unicode Version 18.0.
[188-A74] Action Item for Ken Lunde, CJK: Update the Unihan database to change the kRSUnicode property value of U+2BB89, based on PRI #534 feedback [ID20260517144156] and Section 02 of document L2/26-157, for Unicode Version 18.0.
[188-A75] Action Item for Ken Lunde, CJK: Convey to China feedback on document IRG N2929, based on the CJK & Unihan Working Group comments in Section 02 of document L2/26-157. See PRI #534 feedback [ID20260517144156].
[188-A76] Action Item for Ken Lunde, CJK: Convey to the feedback submitter the CJK & Unihan Working Group comments in Section 02 of document L2/26-157. See PRI #534 feedback [ID20260517144156].
[188-A77] Action Item for Sharon Chen, CJK: Ask the CLDR-TC to check 164 kMandarin property value discrepancies, then report back to the UTC. See PRI #534 feedback [ID20260615125721] and Section 03 of document L2/26-157.
[188-A78] Action Item for Ken Lunde, CJK: Convey to the feedback submitter the CJK & Unihan Working Group comments in Section 03 of document L2/26-157. See PRI #534 feedback [ID20260615125721].
Markus Scherer asked whether the CJK working group maintains tooling for such consistency checks, and offered that PAG might implement such checks. Ken Lunde welcomed the offer from PAG, but cautioned that kMandarin values sometimes intentionally mismatch. Ken Lunde estimated that the number of exceptions may be in the 90s.
[188-C36] Consensus: Accept the informative kRSUnicode property value change for U+2C07B, based on PRI #534 feedback [ID20260628220827] and Section 04 of document L2/26-157, for Unicode Version 18.0.
[188-A79] Action Item for Ken Lunde, CJK: Update the Unihan database to change the kRSUnicode property value of U+2C07B, based on PRI #534 feedback [ID20260628220827] and Section 04 of document L2/26-157, for Unicode Version 18.0.
[188-A80] Action Item for Ken Lunde, CJK: Update the Unihan database to add “2026” as the kJinmeiyoKanji property value of U+52D2, based on PRI #534 feedback [ID20260629042814] and Section 05 of document L2/26-157, for Unicode Version 18.0.
[188-A81] Action Item for Ken Lunde, CJK: Update the Description of the kJinmeiyoKanji property in UAX #38, based on PRI #534 feedback [ID20260629042814] and Section 05 of document L2/26-157, for Unicode Version 18.0.
[188-A82] Action Item for Ken Lunde, CJK: Convey to the feedback submitter the CJK & Unihan Working Group comments in Section 05 of document L2/26-157. See PRI #534 feedback [ID20260629042814].
[188-A83] Action Item for Ken Lunde, CJK: Update the Unihan database to add kZVariant property values to U+7C5C, U+9832, U+2BE14, and U+2BF9F, based on PRI #534 feedback [ID20260630175343] and Section 06 of document L2/26-157, for Unicode Version 18.0.
[188-A84] Action Item for Josh Hadley, Roozbeh Pournader, PAG: Update the UTS #39 confusables.txt data file by adding two new confusable pairs, U+7C5C/U+2BF9F and U+9832/U+2BE14, based on PRI #534 feedback [ID20260630175343] and Section 06 of document L2/26-157, for a future version of the standard.
[188-A85] Action Item for Ken Lunde, CJK: Convey to the feedback submitter the CJK & Unihan Working Group comments in Section 06 of document L2/26-157. See PRI #534 feedback [ID20260630175343].
[188-N2] Note: This feedback demonstrates that feedback submitters should make an effort to more clearly specify which particular addition or change is being proposed, and should not be left to a UTC working group or the UTC to figure out.
Peter Constable asked about the discussion that led to agreement from China. This involved a discussion with Tao Yang from CESI and Eiso Chan.
[188-C37] Consensus: Accept the normative kIRG_GSource property value change for U+32CDD, based on PRI #534 feedback [ID20260706081427] and Section 07 of document L2/26-157, for Unicode Version 18.0.
[188-A86] Action Item for Ken Lunde, CJK: Update the Unihan database to change the kIRG_GSource property value of U+32CDD, based on PRI #534 feedback [ID20260706081427] and Section 07 of document L2/26-157, for Unicode Version 18.0.
[188-A87] Action Item for Ken Lunde, CJK: Change the UAX #45 records for UTC-00970, UTC-01044, UK-01462, and UK-01653, based on PRI #544 feedback [ID20260614092409] and Section 08 of document L2/26-157, for Unicode Version 18.0.
[188-A88] Action Item for Ken Lunde, CJK: Convey to the feedback submitter the CJK & Unihan Working Group comments in Section 09 of document L2/26-157. See PRI #545 feedback [ID20260426150711].
[188-C38] Consensus: Accept the T-source representative glyph changes for U+20CB9, U+21731, U+21DDE, U+22681, and U+22B44, based on PRI #548 feedback [ID20260702100216] and Section 10 of document L2/26-157, for Unicode Version 18.0.
Michel Suignard expressed his agreement with the comments from the CJK & Unihan working group: such a property would be useful, but it is too late for 18.0
[188-A89] Action Item for Ken Lunde, CJK: Convey to the feedback submitter the CJK & Unihan Working Group comments in Section 11 of document L2/26-157. See PRI #548 feedback [ID20260706210943].
Markus Scherer asked whether omitting trailing empty fields was considered; if a new variant were added, a semicolon would have to be added to each line. Ken Lunde replied that the likelihood of finding a third non-Chinese simplified variant of a radical was vanishingly small.
[188-C39] Consensus: Accept the syntax change for the RSIndex.txt data file, based on document L2/26-134 (Lunde), and as amended in Section 12 of document L2/26-157, for Unicode Version 18.0.
[188-A90] Action Item for Ken Lunde, CJK: Update the RSIndex.txt data file, based on document L2/26-134, and as amended in Section 12 of document L2/26-157, for Unicode Version 18.0.
[188-A91] Action Item for Ken Whistler, RMG: Update the Unicode Version 18.0 landing page to mention the syntax change of the RSIndex.txt data file, based on document L2/26-134, and as amended in Section 12 of document L2/26-157.
[188-C40] Consensus: Accept the 458 informative kTotalStrokes property value changes, based on document L2/26-148 (Lunde) and Section 13 of document L2/26-157, for Unicode Version 18.0.
[188-A92] Action Item for Ken Lunde, CJK: Update the Unihan database to change 458 kTotalStrokes property values, based on document L2/26-148 and Section 13 of document L2/26-157, for Unicode Version 18.0.
[188-A93] Action Item for Josh Hadley, Roozbeh Pournader, PAG: Update the UTS #39 confusables.txt data file by adding new confusable pairs, based on document L2/26-127 and Section 14 of document L2/26-157, for a future version of the standard.
[188-C41] Consensus: Provisionally assign U+16FE6 DOUBLE IDEOGRAPHIC FULL STOP in the Ideographic Symbols and Punctuation block; further, provisionally assign the following two new standardized variation sequences; based on document L2/26-133R (Chan) and Section 15 of document L2/26-157:
16FE6 FE00; corner-justified form; # DOUBLE IDEOGRAPHIC FULL STOP
16FE6 FE01; centered form; # DOUBLE IDEOGRAPHIC FULL STOP
[188-A94] Action Item for Ken Lunde, CJK: Provide a font for U+16FE6 DOUBLE IDEOGRAPHIC FULL STOP to Michel Suignard and the Charts Working Group, based on document L2/26-133R and Section 15 of document L2/26-157, for Unicode Version 19.0.
[188-A95] Action Item for Ken Lunde, CJK: Prepare a Proposed Update of UAX #50 that adds the glyphs for U+16FE6 DOUBLE IDEOGRAPHIC FULL STOP to Table 3, based on document L2/26-133R, and as amended in Section 15 of document L2/26-157, for Unicode Version 19.0.
[188-A96] Action Item for Michelle Perham, RMG: Post the Proposed Update of UAX #50 as a new PRI with 2026-10-05 as the closing date, based on document L2/26-133R and Section 15 of document L2/26-157.
[188-A97] Action Item for Ken Lunde, EDC: Prepare text for the Core Specification that describes U+16FE6 DOUBLE IDEOGRAPHIC FULL STOP, based on document L2/26-133R and Section 15 of document L2/26-157, for Unicode Version 19.0.
[188-A98] Action Item for Ken Whistler, RMG: Update the Pipeline page to add U+16FE6 DOUBLE IDEOGRAPHIC FULL STOP and two standardized variation sequences, based on document L2/26-133R and Section 15 of document L2/26-157, as tentative for Unicode Version 19.0.
[188-A99] Action Item for Markus Scherer, PAG: Add two variation sequences to StandardizedVariants.txt, based on document L2/26-133R and Section 15 of document L2/26-157, for Unicode Version 19.0.
[188-A100] Action Item for Michel Suignard, CHARTS: Add two variation sequences and their representative glyphs to the code chart for the Ideographic Symbols and Punctuation block, based on document L2/26-133R and Section 15 of document L2/26-157, for Unicode Version 19.0.
[188-C42] Consensus: Provisionally assign U+16FE5 WHITE IDEOGRAPHIC COMMA in the Ideographic Symbols and Punctuation block, based on document L2/26-136 (Kojitani) and Section 16 of document L2/26-157.
[188-A101] Action Item for Ken Lunde, CJK: Request the author of document L2/26-136 to redesign the glyphs for U+16FE5 WHITE IDEOGRAPHIC COMMA, then provide the font to Michel Suignard and the Charts Working Group, based on document L2/26-136, and as amended in Section 16 of document L2/26-157, for Unicode Version 19.0.
[188-A102] Action Item for Ken Lunde, CJK: Update UAX #50 to add the glyphs for U+16FE5 WHITE IDEOGRAPHIC COMMA to Table 3, based on document L2/26-136, and as amended in Section 16 of document L2/26-157, for Unicode Version 19.0.
[188-A103] Action Item for Ken Lunde, EDC: Prepare text for the Core Specification that describes U+16FE5 WHITE IDEOGRAPHIC COMMA, based on document L2/26-136 and Section 16 of document L2/26-157, for Unicode Version 19.0.
[188-A104] Action Item for Ken Whistler, RMG: Update the Pipeline page to add U+16FE5 WHITE IDEOGRAPHIC COMMA, based on document L2/26-136 and Section 16 of document L2/26-157, as tentative for Unicode Version 19.0.
[188-A105] Action Item for Ken Whistler, CHARTS: Add a cross reference to U+FE46 WHITE SESAME DOT to the entry for U+16FE5 WHITE IDEOGRAPHIC COMMA in NamesList.txt, based on document L2/26-136, and as amended in Section 16 of document L2/26-157, for Unicode Version 19.0.
[188-A106] Action Item for Josh Hadley, Roozbeh Pournader, PAG: Update the UTS #39 confusables.txt data file by adding U+FE46 and U+16FE5 as a new confusable pair, based on document L2/26-136, and as amended in Section 16 of document L2/26-157, for a future version of the Unicode Standard.
[188-A107] Action Item for Ken Lunde, CJK: Convey to the author of document L2/26-136 the CJK & Unihan Working Group comments in Section 16 of document L2/26-157.
Markus Scherer asked how these components should behave in collation, and whether they have radical-stroke data and the Unified_Ideograph property. Ken Lunde and Ken Whistler replied that these characters include kRSUnicode, kTotalStrokes, and Unified_Ideograph properties, and should be treated as unified ideographs. Markus Scherer noted that a note would be needed in UTS #10. Markus Scherer noted that since they have the Unified_Ideograph property, they would automatically be handled by UTS #10.
Peter Constable noted that these characters have been recommended for Amendment 1 by WG 2, it makes sense to approve them for Unicode Version 19.0, rather than provisionally assign them.
[188-C43] Consensus: Accept 494 CJK Unified Ideographs in the new CJK Unified Ideographs Components-A and CJK Unified Ideographs Components-B blocks with code point ranges U+2FA20..U+2FA96 (block: U+2FA20..U+2FA9F) and U+2FAA0..U+2FC16 (block: U+2FAA0..U+2FC1F), respectively, along with their provided normative IRG source references and informative properties, based on documents IRG N2878R3 (China & TCA) and WG 2 N5358 (China & TCA), and Section 17 of document L2/26-157, for Unicode Version 19.0.
Ken Whistler celebrated this approval for blocks that had been in the works for twenty years. Ken Lunde noted that this would set a precedent for future work on components.
[188-A108] Action Item for Ken Lunde, CJK: Update the Unihan database to add the CJK Unified Ideographs Components-A and CJK Unified Ideographs Components-B blocks, based on documents IRG N2878R3 and WG 2 N5358, and Section 17 of document L2/26-157, for Unicode Version 19.0.
[188-A109] Action Item for Ken Lunde, CJK: Prepare a Proposed Update of UAX #38 that adds the CJK Unified Ideographs Components-A and CJK Unified Ideographs Components-B blocks to Section 4.4, to document the new “GCP” and “GCW” source prefixes for the kIRG_GSource property, to document the new “HCP” source prefix for the kIRG_HSource property, and to document the new “TCP” source prefix for the kIRG_TSource property, based on documents IRG N2878R3 and WG 2 N5358, and Section 17 of document L2/26-157, for Unicode Version 19.0.
[188-A110] Action Item for Michelle Perham, RMG: Post the Proposed Update of UAX #38 as a new PRI with 2026-10-05 as the closing date, based on documents IRG N2878R3 and WG 2 N5358, and Section 17 of document L2/26-157.
[188-A111] Action Item for Ken Whistler, RMG: Update the Pipeline page to add the CJK Unified Ideographs Components-A and CJK Unified Ideographs Components-B blocks, based on documents IRG N2878R3 and WG 2 N5358, and Section 17 of document L2/26-157, for Unicode Version 19.0.
[188-A112] Action Item for Ken Whistler, RMG: Add the CJK Unified Ideographs Components-A and CJK Unified Ideographs Components-B blocks to the Alpha review notes for Unicode Version 19.0, based on documents IRG N2878R3 and WG 2 N5358, and Section 17 of document L2/26-157.
[188-A113] Action Item for Ken Whistler, EDC: Update Table 4-8 of the Core Specification to add these two new blocks and their Rules (NR2) and Prefix Strings (“CJK UNIFIED IDEOGRAPH-”), based on documents IRG N2878R3 and WG 2 N5358, and Section 17 of document L2/26-157, for Unicode Version 19.0.
[188-A114] Action Item for Ken Lunde, EDC: Update the Core Specification to add the CJK Unified Ideographs Components-A and CJK Unified Ideographs Components-B blocks, based on documents IRG N2878R3 and WG 2 N5358, and Section 17 of document L2/26-157, for Unicode Version 19.0.
[188-A115] Action Item for Ken Whistler, PAG: In Unicode Technical Standard #10 (UCA), add a note about the new CJK Unified Ideographs Components-A and CJK Unified Ideographs Components-B blocks near Table 16. Computing Implicit Weights, for Unicode Version 19.0.
Jan Kučera stated that the script encoding working group would keep the roadmap up to date based on UTC decisions, and did not require separate action items.
[188-C44] Consensus: Accept the normative kIRG_GSource property value change for U+48B4, based on documents IRG N2926 (Chan) and WG 2 N5359 (Chan), and Section 18 of document L2/26-157, for Unicode Version 18.0.
[188-C45] Consensus: Accept the new normative kIRG_SGSource property, based on documents IRG N2926 (Chan) and WG 2 N5359 (Chan), and Section 18 of document L2/26-157, for Unicode Version 19.0.
Ken Whistler noted that the addition of a new source will require changes in the chart generation software. Michel Suignard noted that a Malaysian source may be coming soon.
[188-A116] Action Item for Ken Lunde, CJK: Update the Unihan database to change the kIRG_GSource property value of U+48B4, based on documents IRG N2926 and WG 2 N5359, and Section 18 of document L2/26-157, for Unicode Version 18.0.
[188-A117] Action Item for Ken Lunde, CJK: Update the Unihan database to remove 225 kIRG_GSource property values, to add the new kIRG_SGSource property, and to add 226 new kIRG_SGSource property values, based on documents IRG N2926 and WG 2 N5359, and Section 18 of document L2/26-157, for Unicode Version 19.0.
[188-A118] Action Item for Ken Lunde, CJK: Update UAX #38 to add the new kIRG_SGSource property, based on documents IRG N2926 and WG 2 N5359, and Section 18 of document L2/26-157, for Unicode Version 19.0.
[188-A119] Action Item for John Wilcock, PAG: Add the kIRG_SGSource property to UAX #42, based on documents IRG N2926 and WG 2 N5359, and Section 18 of document L2/26-157, for Unicode Version 19.0.
[188-A120] Action Item for Ken Whistler, RMG: Add the new kIRG_SGSource property to the Alpha review notes for Unicode Version 19.0, based on documents IRG N2926 and WG 2 N5359, and Section 18 of document L2/26-157.
[188-N3] Note: The UTC endorses document L2/26-147R as a UTC contribution to the IRG that proposes to horizontally extend 45 ideographs in existing CJK Unified Ideographs blocks.
Ken Lunde noted that AG 1 now appears on the ISO website. Peter Constable noted that the Unicode Consortium, as well as the US National body, would be asked to appoint experts to AG 1. Ken Lunde stated that the call for participation had not yet been generally circulated, but was expected to be active through the end of August.
Ken Whistler requested that Ken Lunde provide updated text for Section E.2.2 of The Unicode Standard.
Peter Constable asked Ken Lunde to provide an overview of the relationship between IRG and the CJK & Working Group. Ken Lunde replied that the IRG really focuses on producing the extension blocks, through the IRG working sets. A new block typically undergoes six review cycles over three years. The CJK & Unihan working group effectively acts as a liaison between the IRG and the UTC. It also handles changes that are out of scope for the IRG. When proposals before the CJK & Unihan Working Group are in scope for the IRG, or when encoding proposals are considered, such as proposals for new punctuation marks, the relevant IRG member bodies are engaged. Ken Lunde noted that CJK & Unihan working group participants include people who are also appointed to IRG by the Chinese and Vietnamese National Bodies.
Discussion. No action needed for this section.
Peter Constable stated that all feedback had been covered as part of the CJK & Unihan Working Group report L2/26-157.
Andrew Glass presented L2/26-204, which shows draft text for Section 11.4.2, Egyptian Hieroglyph Format Controls, of The Unicode Standard, and reported on ad hoc discussions over lunch during the first two days.
Andrew Glass explained that the UTC had already agreed that the ordering of the operands of EGYPTIAN HIEROGLYPH OVERLAY MIDDLE was based on kEH_Cat values. The question raised during the ad hoc meetings was the ordering when one or more of the operands is a group. The order is then lexicographic in the kEH_Cat values of component hieroglyphs; at a second level, the code points of the format controls in the groups are used.
Andrew Glass noted that when there are different ways of decomposing a group, the appropriate decomposition should be considered an orthographic matter, and is not for The Unicode Standard to define.
Andrew Glass clarified that some system fonts have an incorrect glyph for U+1325F 𓉟 (missing the bread U+133CF 𓏏), but that this would not be an issue in the core specification.
[188-C46] Consensus: Approve the changes to kEH_Cat, kEH_Desc, kEH_HG, kEH_IFAO, and kEH_JSesh, listed in document L2/26-199, for Unicode Version 18.0.
Liang Hai presented document L2/26-155.
[188-A121] Action Item for Ken Whistler, EDC: Add Ol Onal and Tolong Siki to Table 22-3, and adjust text as suggested for Myanmar, for Unicode Version 18.0. [Ref: 3.1 in L2/26-155 and ID20260524103017 in PRI #548]
[188-A122] Action Item for Ken Whistler, EDC: Review the feedback and make adjustments when appropriate, for Unicode Version 18.0. [Ref: 3.1 in L2/26-155 and ID20260526172722 in PRI #548]
[188-A123] Action Item for Ken Whistler, EDC: Reword NR2 to clarify that Egyptian Hieroglyphs are indeed covered by it, for Unicode Version 18.0. [Ref: 3.1 in L2/26-155 and ID20260706200358 in PRI #548]
[188-A124] Action Item for Ken Whistler, EDC: Correct [Knuth78] in UAX #41 as suggested, for Unicode Version 18.0. [Ref: 3.2 in L2/26-155 and ID20260529151202 in PRI #550]
[188-A125] Action Item for Richard Ishida, EDC: Investigate the issues and take ID20260507142418 into consideration, then propose edits, for Unicode Version 19.0. [Ref: 3.3 in L2/26-155, L2/26-128, and ID20260507142418 in PRI #544]
[188-A126] Action Item for Asmus Freytag, EDC: Clarify U+2E31’s relationship with U+00B7 in Chapter 6, and also survey the usage recommendations in various script sections, for Unicode Version 18.0. [Ref: 3.4 in L2/26-155 and ID20260512134510 in PRI #544]
[188-A127] Action Item for Ken Whistler, EDC: Incorporate the suggested text in the core spec, for Unicode Version 18.0. [Ref: 3.4 in L2/26-155 and ID20260627134929 in PRI #544]
Lunch break 12:28–13:16 UTC−7 (21:28–22:16 UTC+2) .
Liang Hai presented document L2/26-159.
Liang Hai noted that there are two new participants to the group: Huang Junliang and Kushim Jiang.
Peter Constable asked whether delta charts are created when only annotations change. Michel Suignard replied that this was not the case, otherwise the delta charts would be a copy of the whole code charts. Minor stylistic glyph changes are also treated as editorial and not indicated in the delta charts.
Asmus Freytag stated that while it is desirable for newer versions of the standard to be found in preference to older versions in web searches, it is also important to be able to search older versions.
Michel Suignard stated that he expected the commitment from the charts working group described in Section 2 of L2/26-159 to be of limited duration, as after the transition of ISO/IEC 10646 to a maintenance agency, the regular code charts could be used.
Regarding Section 3.4 of L2/26-159, Peter Constable commented that in both the cases of the Omani and Saudi riyal signs, for which specific dimensions were specified, there was concern about expectations of integration into typefaces. Peter Constable reached out to both monetary authorities, and that in the case of the Omani riyal sign, discussions with typographers had led to a narrower version of the sign in the guidelines.
Peter Constable noted that all feedback had been covered at the meeting.
Short discussion among release management group participants, resulting in a revision of the report, L2/26-160R.
[188-C47] Consensus: The UTC authorizes the release of Unicode 18.0 and synchronized standards UTS #10, UTS #39, UTS #46, UTS #51 and UTS #58, incorporating the results of decisions and actions as recorded during UTC #188, with a target release date of September 16, 2026.
[188-A128] Action Item for Ken Whistler, PAG: Prepare an updated NamesList.txt for Unicode Version 18.0, synched with the final Unicode Version 18.0 repertoire, as finalized during UTC #188.
[188-A129] Action Item for Markus Scherer, PAG: Prepare final candidate data files for the UCD, and the data directories for UTS #10, UTS #39, UTS #46, and UTS #58, for the Unicode Version 18.0 release.
[188-A130] Action Item for Ned Holbrook, ESC: Prepare final candidate data files for the data directory of UTS #51, and a complete set of emoji final charts for Emoji 18.0.
[188-A131] Action Item for Peter Constable, RMG: Complete all tasks associated with the 18.0 release.
[188-A132] Action Item for Michelle Perham, UTC: Close PRIs ## 520, 523, 533 – 535, 538 – 539, 542 – 545, 547 – 554.
Debbie Anderson noted that following the WG 2 meeting, she had sent BIS a list of all proposals affecting scripts from India, and would do so again after this meeting.
The closing date of record is October 5th, 2026.
The next closing date of record is December 30th, 2025.
Short Discussion.
Anushah Hossain presented L2/26-161.
Markus Scherer stated that nothing had been registered recently, but that some scripts would soon need to be registered.
Roozbeh Pournader expressed a concern that the requirement for approval of decomposition mappings might be adding a requirement for ISO approval of data not currently in scope of ISO/IEC 10646. Alexis Hunt replied that this was not the intent: whatever is currently in scope should remain in scope, and whatever is not currently in scope should not be brought in scope by these terms of reference.
Robin Leroy stated that he intends to produce a document illustrating how the process might work out in practice.
UTC thanks the host, Microsoft.
UTC adjourned for the week at 15:00 UTC−7 (00:00 UTC+2) .
Shortly after the close of UTC #188, while finalizing the changes for 188-A59, it was noted that the revised guidance in L2/26-200 §2.7.1 unintentionally recommended emitting text variation selectors for ASCII characters.
In order to remedy that, the following decision was taken by email according to the TC Decision Process by Email in the TC Procedures:
[188-C48] Consensus: Approve UTS #51 Version 18.0 for release, incorporating post UTC #188 revisions to section 2.7.1, as shown in L2/26-210.
Members Represented
| Roll Call | Jul 28 | Jul 29 | Jul 30 |
|---|---|---|---|
| Full members | |||
| Adobe Inc. | Y | Y | Y |
| Airbnb | |||
| Apple | Y | Y | Y |
| Y | Y | Y | |
| Meta Platforms, Inc | |||
| Microsoft Corporation | Y | Y | Y |
| Salesforce.com | |||
| Translated | |||
| Supporting members | |||
| Ministry of Electronics and Information Technology (MeitY) | |||
| Stanford University - SILICON | |||
| Tamil Nadu, Tamil Virtual Academy | |||
| University of California at Berkeley - Script Encoding Initiative | Y | Y | Y |
| Associate members | |||
| Amazon.com Services Inc. | |||
| Beijing Zhong Yi Electronics Co. | |||
| Bloomberg LP | |||
| Decotype, Inc. | |||
| Ebanah | Y | ||
| Oracle Corporation | |||
| SAS Institute Inc. | |||
| SIL Global | Y | Y | Y |
| The Wikimedia Foundation | |||
| Wénlín | |||
UTC Attendance
| Name | Representing |
|---|---|
| Salim Al Mandhari | Ebanah |
| Basit Ali | Microsoft |
| Debbie Anderson | UCB |
| Maria Glenda Bellarosa | self |
| Maralina Caldas | Unicode |
| Christopher Chapman | Apple |
| Peter Constable | Microsoft |
| Toral Cowieson | Unicode |
| Sam Cundall | |
| Jennifer Daniel | |
| Kamilé Demir | Adobe |
| Lorna Evans | SIL |
| Tiffa Foster | |
| Asmus Freytag | self |
| Andrew Glass | Microsoft |
| Fraser Gordon | self |
| Fraser Gordon’s cat | self |
| Joshua Hadley | self |
| Ned Holbrook | Apple |
| Anushah Hossain | UCB |
| Alexis Hunt | self |
| Jiang Zhaoqin (Kushim) | self |
| Jan Kučera | Unicode |
| Gregg Lechner | Unicode |
| Robin Leroy | |
| Liang Hai | Unicode |
| Ken Lunde | Apple |
| Sudip Iglesias Murmu | self |
| Nova Patch | |
| Michelle Perham | Unicode |
| Roozbeh Pournader | Unicode |
| Murray Sargent | self |
| Markus Scherer | |
| Mark Shoulson | self |
| CheonHyeong Sim | self |
| Michel Suignard | Unicode |
| Misba Tabassun | |
| Tex Texin | Xencraft |
| Ken Whistler | Unicode |
| John Wilcock | Microsoft |