Other issues in the Registry section.

[macchiati]

I was looking at https://github.com/unicode-org/message-format-wg/pull/659/, but found a number of issues in the Registry section before the Default Registry. Capturing them here.

In the Conformance and Use:

The term 'registry' is being used ambiguously. That is, the distinction is not clearly enough drawn between the default registry, and the machine-readable format that is used by the default registry, but can also be used as implementations to provide machine-readable documentation/specification of their own registry, with additions to the default registry.

For example, "The MessageFormat 2.0 Registry was created to describe the core set of formatting and selection functions," seems really to be describing the default registry, not the registry format. So it should say "The MessageFormat 2.0 Default Registry"

In the next paragraph "The registry provides a machine-readable description" seems to be talking about the use as a format. It should have its own term, like 'registry format' (but doesn't have to be exactly that term).

I suggest that we don't have any 'naked' mentions of the word 'registry' in the spec, but rather qualify it to resolve the ambiguity.

Read-only options (the readonly attribute) can be displayed to translators in CAT tools, but may not be edited.

I think this is a mistake. I think the default should be read-only, and with only the (few) options that translators can modify being marked as "mutable" or some equivalent term.

validate their input with a regular expression (the validationRule attribute) Add: to some extent. Or some other qualifying phrase to indicate that they might be stricter than the regex.

The example needs to be fixed, eg to drop bogus lines like <!-- Since this applies to both cardinal and ordinal, all plural options are valid. --> and <match locales="en" values="one two few other" validationRule="anyNumber"/>

A localization engineer can then extend the registry by defining the following customRegistry.xml file.

We should add a note recommending that if the registry is available externally a namespace should be used to avoid collisions.

{$color :adjective article=indefinite accord=$object} BTW, this is probably a bad example. Some languages insert a word in front of 'color' to indicate definiteness, but others modify the noun.

[aphillips]

@macchiati Please use separate issues for separate issues.

---

Regarding the use of the term registry and the term default registry, I agree. We switched to a prose registry fairly late in the process and did not change the intro or finish off the XML part of the spec. We should make that section experimental or remove it for now.

During the post-45 period, developing the registry and registry process (whose design I just started) will be one key work item. Another will be the data model.

[macchiati]

I would recommend removing it entirely. It is so experimental it has no business being in the spec, even as a tech preview. You could instead add a Note that work will be continuing on defining machine-readable file format for registries during the tech preview period.

On Wed, Feb 21, 2024 at 4:10 PM Addison Phillips @.***> wrote:

@macchiati https://github.com/macchiati Please use separate issues for separate issues.

Regarding the use of the term registry and the term default registry, I agree. We switched to a prose registry fairly late in the process and did not change the intro or finish off the XML part of the spec. We should make that section experimental or remove it for now.

During the post-45 period, developing the registry and registry process (whose design I just started) will be one key work item. Another will be the data model.

— Reply to this email directly, view it on GitHub https://github.com/unicode-org/message-format-wg/issues/677#issuecomment-1958427825, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACJLEMCURKNIH7LVXHVXEBTYU2EJBAVCNFSM6AAAAABDT6IZPWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNJYGQZDOOBSGU . You are receiving this because you were mentioned.Message ID: @.***>

[macchiati]

Would it be appropriate for me to create a PR for doing that?

On Wed, Feb 21, 2024 at 6:18 PM Mark Davis Ⓤ @.***> wrote:

I would recommend removing it entirely. It is so experimental it has no business being in the spec, even as a tech preview. You could instead add a Note that work will be continuing on defining machine-readable file format for registries during the tech preview period.

On Wed, Feb 21, 2024 at 4:10 PM Addison Phillips @.***> wrote:

@macchiati https://github.com/macchiati Please use separate issues for separate issues.

Regarding the use of the term registry and the term default registry, I agree. We switched to a prose registry fairly late in the process and did not change the intro or finish off the XML part of the spec. We should make that section experimental or remove it for now.

During the post-45 period, developing the registry and registry process (whose design I just started) will be one key work item. Another will be the data model.

— Reply to this email directly, view it on GitHub https://github.com/unicode-org/message-format-wg/issues/677#issuecomment-1958427825, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACJLEMCURKNIH7LVXHVXEBTYU2EJBAVCNFSM6AAAAABDT6IZPWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNJYGQZDOOBSGU . You are receiving this because you were mentioned.Message ID: @.***>

[aphillips]

@macchiati

Would it be appropriate for me to create a PR for doing that?

No. I'm making a separate issue and PR now.

[aphillips]

I think all of the issues here have been addressed. If not, a separate issue for each is probably warranted. resolve-candidate so we consider for close in the 2024-11-11 call