Re: [Ltru] Re: Inversions and other problems

["Mark Davis" <mark.davis at icu-project.org>]

On 10/17/06, Addison Phillips <addison at yahoo-inc.com> wrote:

Mark Davis wrote:
> As I said before, I can live with inversion IF it is fixed. That means
> an extension of Addison's proposed rules.
>
> 1. Follow the source standard.
>
> 2. Split multiple names into multiple description fields.
> I believe Addison meant: X, Y, Y =>
> Description: X
> Description: Y
> Description: Z

Yes, exactly so.

>
> 3. Omit or edit problematic items, subject to consensus.
>
> plus

Well... my goal was to avoid having to spell all of this out. We've got
enough rules, so I think some hand-waving is in order. While I note the
problem projects like CLDR might have producing localizations, the use
of an actual Turing machine (aka Doug Ewell) to create the entries
should eliminate a lot of woes. Furthermore, I tend to think that the
*detailed* rules you propose should go into 4645bis (i.e. "how did the
entries get this way"), while the simpler version I proposed can go into
4646bis (perhaps with a pointer to "pre-existing customary practice" in
4645bis).

That's ok, as long as they are sufficiently spelled out such that they prevent mistakes.

>
> 4. Ensure that there are no duplicates (where inversion and/or removal
> of parenComments results in a duplicate within the same subtag, or where
> names are the same across different subtags of the same type)

I wasn't specific enough in my little bit of text above. "Duplicates"
(including inverted duplicates) are members of the class "problematic".

ok

>
> 5. The only information in parentheses is a comment, eg a clarification
> of the description.

This could be guidance added to rule 3.

ok

>
> 6. The final result being that there are NO description items with
> commas outside of parenComments, except for the case that there is a
> single comma indicating inversion.
>

Except that you go on to demonstrate below that sometimes inversion is
illusory.

No. What I mean is that any cases of false inversion need to be fixed. There are not that many of them.

> If these are the rules, then I'm satisfied.
>
> ==
>
> Ideally, we'd have two additional rules:
>
> 7. The most commonly known name is in the first Description.
> (This may already be the case -- I don't know.)

I don't think this is appropriate. The level of judgment required is too
great for the ietf-languages list to manage.

I'm not thinking of a requirement, more of a weak should.

>
> 8. A description is unique even if the parenComment is removed.
>

We need a "MUST" here. I'd propose:

--
For a given record type, each Description field MUST be unique.

I'd like slightly string.

For a given record type, each Description field (after removing any parantheticals and inverting any comma-separated strings) MUST be unique.

--

Addison

--
Addison Phillips
Globalization Architect -- Yahoo! Inc.

Internationalization is an architecture.
It is not a feature.

_______________________________________________
Ltru mailing list
Ltru at ietf.org
https://www1.ietf.org/mailman/listinfo/ltru