Skip to content

Commit 445bdbe

Browse files
aphillipsaphillips
authored
Create notes-2024-01-08.md
1 parent 4619bd8 commit 445bdbe

File tree

1 file changed

+165
-0
lines changed

1 file changed

+165
-0
lines changed

‎meetings/2024/notes-2024-01-08.md‎

Lines changed: 165 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,165 @@
1+
# 8 January 2024 | MessageFormat Working Group Regular Teleconference
2+
3+
### Attendees
4+
- Addison Phillips - Unicode (APP) - chair
5+
- Eemeli Aro - Mozilla (EAO)
6+
- Ujjwal Sharma - Igalia (USA)
7+
- Simon Clark - Oracle (SCA)
8+
- Elango Cheran - Google (ECH)
9+
- Mark Davis - Google (MED)
10+
- Staś Małolepszy - Google (STA)
11+
- Mihai Niță - Google (MIH)
12+
- Richard Gibson - OpenJSF (RGN)
13+
- Matt Radbourne - Bloomberg (MRR)
14+
- Zibi Braniecki - Amazon (ZBI)
15+
16+
17+
Scribe: SCA
18+
19+
20+
21+
## Topic: Action Item Review
22+
23+
MED - when do we ask CLDR for feedback on MF2.0? Can ask them now to look at material as we have reached a place of relative stability again?
24+
25+
EAO - feature freeze in a week. We will then be a point of not considering further new topics.
26+
27+
MED - things have to be extremely compelling after that to be considered.
28+
29+
APP - feedback on open topics is still actionable. Anything new is unwelcome. Next Tuesday is the day after which we will no longer accept new considerations / topics
30+
31+
## INFO SHARE
32+
EAO - wrote up an NPM implementation of MF2.0 excluding bi-di, easier to leave out. Includes markup, npm install messageformat@next
33+
34+
Also Updated MF1 to MF2 upgrade tooling and XLIFF to MF2.0 conversion tooling.
35+
36+
APP - Mihai will update java implementation? Yes.
37+
38+
## Face-to-Face update and logistics
39+
40+
https://github.com/unicode-org/message-format-wg/wiki/Face%E2%80%90to%E2%80%90Face-Tomales-California
41+
42+
## Topic: Progressing to Done
43+
44+
The main blockers appear to be the following:
45+
- ~~Beautification of the syntax discussion~~
46+
- ~~What’s in a name? (Does NCName fix our woes? Go to UAX31? what?)~~
47+
- ~~Quoting~~
48+
- ~~Format-to-Parts~~
49+
- Spannables
50+
- Expression Attributes
51+
- Registry and default functions
52+
- Implementation and testing
53+
54+
55+
Schedule:
56+
1. No new LDML45 issues after 15 January.
57+
2. All LDML45 issues resolved by end of F2F. Balloting alpha spec to occur 11 February.
58+
3. Beta spec and registry by 11 March.
59+
4. Can make limited changes thereafter, for issues discovered by implementers.
60+
5. Release 10 April as part of LDML45
61+
62+
Mark: question as to when to start asking for feedback.
63+
64+
## Topic: Expression Attributes
65+
66+
67+
APP: we’ve had a design doc for a while. Need to decide for LDML45, even if we don’t have concrete details. https://github.com/unicode-org/message-format-wg/pull/450
68+
69+
EAO : design doc is merged in as proposed . PR is the implementation.
70+
71+
MIH : can live without specs. Missing what we consider functionality. Merging different concerns into single functionality is iffy
72+
73+
STA: Solution without a problem, not clear how this solves the problems that we do have. Issue #513 outlines concerns with whole proposal. Can do a better job of documenting use cases, Unsure if we need to add a whole new mechanism to the syntax. Worried will cause confusion. What is option to function and what is an attribute?
74+
75+
APP @locale is a useful thing. For this release- we say this is our syntax, but don’t specify anything else. Could also just put @ in the reserved sigil buck.
76+
77+
ECH - not sure why we can just use local variable for the intended usecases.
78+
79+
EAO - current spec of message resource syntax includes something similar, specifying metadata for a message. Can serve the same purpose. Other option to allow them but specify that they have no impact on the runtime. Perhaps a good explanation for why they are different and useful?
80+
81+
STA - @locale is underspecified. Intent was to specify locale for formatters in a subrun of text. Workaround is good enough, passing the option to the formatter should be enough. Markup could be a viable alternative. Also has the benefit of supporting all other parts of our functionality. Consider namespace for things that could make good attributes.
82+
83+
MED - coming around to cautious approach. Attr available to tools is interesting. Unclear if the metadata can be separated out into wrapper class. Something like case is relevant to any formatter. Leaning towards holding off until we have more discussion.
84+
85+
APP - agrees something are underspecified. Diversity of implementation is bad for interoperability. Can be a directive at a different level? Translation sub-pattern to help XLIFF or tools.
86+
87+
EAO - sounds like we have no conclusion, nor can get one by current deadline. How to best future proof - reserve @ prefix for option names, strong possibility that we might unreserve it later. Include direction the implementations or custom functions should not provide support for @locale
88+
89+
STA - to fully reserve possibility to add later, we need other changes to the spec.
90+
91+
APP - wants to never revisit syntax. “Here is a reserved sigil in the syntax for future functionality.”
92+
93+
EAO - do we spec a new sort of reserved thing.
94+
95+
APP - ok to have something that is reserved but is not a thing yet. Make a note in the spec that we are going to revisit.
96+
97+
STA - want to make it explicit that if we never find a use for it, we are ok with that.
98+
99+
MIH - do the same of markup? Reserve options for later functionality
100+
101+
APP - we will incorporate syntax for attrib as reserved, on expression and markup and make them currently explicitly not do anything at runtime. Reserved for future standardization. You do not error on them, you do not do anything with them.
102+
103+
## Topic: XLIFF compatibility
104+
105+
APP - flush out XLIFF compatibility more after deadline. Have we not broken XLIFF compatibility with anything we’ve done?
106+
107+
EAO - XLIFF <-> MF2 fully capable cross conversion
108+
109+
MIH - we had several options for this before, but no good solutions.
110+
111+
## Topic: Default registry and MF1 compatibility matrix
112+
113+
See PR #564
114+
115+
## Topic: Active PR review
116+
117+
Discussion of active PRs. We will merge or reject them in the call.
118+
The recommendation "discuss" is to ensure there is WG consensus before merging. The recommendation "merge with edits" is to merge once existing comments have been addressed.
119+
Discussion of active PRs. We will merge or reject them in the call.
120+
| PR | Description | Recommendation | Chair's Recommendation |
121+
|:----:|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:|:---------------------------:|:--------------------------:|
122+
| #580 | Add missing Forward Reference data model error EAO - this needs to be explicitly covered as an error. MIH - better to throw as an undefined reference? EAO - agreed should be an error. Name is just unclear. Will discuss name in PR | Discuss | Discuss |
123+
| #577 | Clarify that Duplicate Option Name is a data model error Defined in formatting.md that this is an error. Needs to be specified in syntax.md | Discuss | Discuss |
124+
| #576 | Forbid the logo as a valid expression APP - objections to shipping? Seems to be a generally good shange. STA will give it a thumb | Merge | Discuss |
125+
| #574 | Add markup as designed APP - would really like to see this get merged. Been discussed extensively. MIH - can we also have standalone markup in here. STA/EAO it is mentioned here. | Merge | Close (waiting on PR #574) |
126+
| #570 | Add :date and :time aliases APP : skip for now STA: has a concern about aliases in registry. Adds complexity when we don’t need it. APP : doesn’t care about aliases. Base functions should be natural to use. Writing messages should be as intuitive as we can make it. Shorthands will be used very frequently. EAO : understand that the XML of the registry is not part of the currently deliverable spec. Discus post LDML45? APP registry must be well enough described that implementations are possible. EAO : do have an action to add registry spec to registry.md STA: aliases are premature optimization. Repeating options for now is a reasonable path. Kill for now. New PR when time is right. | Discuss | Reject |
127+
| #566 | Improve bidi isolation requirements APP - trying to develop bidi strategy to EAO - does the group want to leave it to EAO and APP to resolve? MIH - bidi specifiers should only show between placeholders and text. Is it going to impact runtime? ZBI - MF1 bidi isolation caused significant confusion with users. MIH - both clojure and android has bidi markers. Has i18nEqual as opposed to equal to work with some of this. EAO we do currently have a bidi isolation spec that must be implemented. This is about the fine details of the algorithm , Propose comment on the issue in the next two days to be part of the bidi committee. | Discuss | Close (discuss) |
128+
| #564 | (Design) Default registry and MF1 compatibility matrix APP merge and iterate? MIH - concerns about aliases. EAO - :integer is currently in. PR open adding :date and :time MIH unhappy shoving stuff through, then getting stuck with it as “previously agree” APP - need to go through all options, and have a process to finalize. This starting point taken from existing registry.xml EAO - if anyone is against what is in the current spec, then register it in the next week, so we can have an explicit list of open discussions at the cutoff. APP - only discussions with the LDML45 tag will be considered ongoing. | Discuss | |
129+
| #562 | Add content-char as common root of simple-start-char, text-char, quoted-char and reserved-char | Discuss | |
130+
| #560 | Add a <matchSIgnature> for :number, together with :ordinal and :plural aliases On hold for now | Discuss | |
131+
| #558 | Add <when> to help select the right <match> | Discuss | |
132+
| #502 | Make option values options, defaulting to true STA - do we want this for expression attributes Ie @canCopy is functionally same at @canCopy = true Kill this for now. Options require values. | Discuss | |
133+
| #450 | Define @attributes on expressions | Discuss (with #473) | |
134+
| #438 | Add details to “Missing Selector Annotation error” section | Close (per 2023-12-11 call) | |
135+
136+
The recommendation "discuss" is to ensure there is WG consensus before merging. The recommendation "merge with edits" is to merge once existing comments have been addressed.
137+
138+
139+
## Topic: Open Issue Review
140+
https://github.com/unicode-org/message-format-wg/issues
141+
142+
Currently we have 38 open (was 46 last time).
143+
144+
5 are resolved-candidate and proposed for close.
145+
146+
7 are Agenda+ and proposed for discussion.
147+
148+
2 are Future (nor for this release)
149+
150+
| Issue | Status | Description | Chair's Recommendation |
151+
|:-----:|:--------------------------:|:-----------------------------------------------------------------------------------------------:|:-------------------------------------------:|
152+
| #579 | Agenda+ | Security considerations section APP - Any volunteers to write one… no… APP will take as action. | Discuss |
153+
| #547 | Agenda+ | Consider reserve syntax some more | Discuss |
154+
| #537 | Agenda+ | [Discussion] {{Spannables}} | Discuss |
155+
| #519 | Agenda+ | Name syntax should align with XML | Close (waiting on PR #574) |
156+
| #489 | Agenda+ | Consider adding { and } to reserved sigils to allow future syntax expansion | Discuss Close- can’t do with current syntax |
157+
| #376 | Agenda+ | Allow constraints on arguments | Reject |
158+
| #169 | Agenda+, resolve-candidate | Clarifications needed on XLIFF mapping | Close (discuss) |
159+
160+
## Topic: AOB?
161+
Make errors section in formatting.md a new doc.
162+
163+
STA - mention new PRs?
164+
165+

0 commit comments

Comments
 (0)