From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Tue Dec 13 2005 - 16:39:11 CST
From: "Addison Phillips" <addison.phillips@quest.com>
To: "Mike Ayers" <mayers@celequest.com>
Cc: "Unicode Mailing List" <unicode@unicode.org>
Sent: Tuesday, December 13, 2005 10:03 PM
Subject: RE: Globalized lists
>I was midway in replying to Mark's message when this one arrived...
>
> I tend to recommend against (and was going to point out that Mark's design
> neatly avoids) treating a list as:
>
> TOKEN[0] + SEPARATOR + TOKEN[1] + ... + LASTSEPARATOR + TOKEN[n]
>
> But rather to treat it as:
>
> TOKEN[0] + TOKEN[1] + ... + TOKEN[n]
>
> That is, to use a pattern string to allow the text on BOTH sides of the
> token to be varied. The localizable variables look more like:
>
> "nonePattern", "(none)"
> "oneItemPattern", "{0}"
> "twoItemPattern", "{0} and {1}" // only two items
> "startPattern", "{0}" // the other place the list might vary
> // is at the start!
> "itemPattern", "{0}, {1}" // used to accrete items to the list
> "lastPattern", "{0}, and {1}" // used at the end of the list only
> // note that this is where you punctuate
> // the entire list (enclosing in quotes
> // for example).
This is probably not enough: the relationship of items in the list is not
encoded ("and" is most often used when in fact it means "or" or even "only
one of". The cardinality of alternatives implied by a list is important,
notably when thelists are used in negative sentences.
Also nothing here correctly describes the possible transformation of each
item in the list, due to logical relationship (not, inclusive/exclusive), or
the relation of the list itself with other elements in a sentence (i.e. the
grammatical case such as the genetive marks like, in English, the « 's »
suffix or even just « ' » suffix after an item terminated by s, or
transformation of names or adjectives or nominal groups, or the grammatical
match in genre and number when a list contains adjectival or quilificative
items, or the repetition of grammatical words like
"of"/"from"/"to"/"in"/...)
A list in a natural language has then to be localized according to its
context of use, and trying to build a unique general rule to create
enumeration list builder function that can be used in all context would fail
in many cases.
Providing the context will then be useful when giving the localization data
to translators. In thiscase, it's good to useful to give to translators
complete sentences to translate, and the expected variations of the list,
with no item, one item, two items, three items and four items. And to say to
translators that all of them must be translated in a way that allow
developers to detect where and how the list items are presented.
The most difficult case will occur when individual list items are
transformed, notably by case variation or by mutation of leading or trailing
consonnants,or by elision or transformation of trailing vowels (including in
proper names, trademarks, company names... something very usual and
sometimes required by some languages because this is often the only
unambiguous way to specify the grammatical structure and semantic of the
sentence).
Let's not forget that the coordination word like "and" may also be present
as end of the final element and not only between the last two items, for
example the Latin "que" suffix,whose position is on the main name of a
nominal group, and which could then be in the middle of the last list item!
For many cases, it will be nearly impossible to create a single sentence to
present the list, so the translation could then need using other devices
such as creating a bulleted list that will better replace a complete
sentence from another source language.
Care must betaken also when list items are themselves complex, such as when
they are also lists possibily using another type of relationship ("and"
instead of "or"). The common assumption (in European languagesand in most
programming languages) that the and-relation has a greater priority than the
or-relation may be opposed to the common meaning used in other languages.
The way we can translate these lists may be even more difficult when the
operator is implied in long lists (the repetition of the operator may be
needed instead of just using the usual comma separation that normally
abbreviates such lists), and when at least one of the item is translatable
only with bulleted lists.
This archive was generated by hypermail 2.1.5 : Tue Dec 13 2005 - 19:20:56 CST