From: Patrick Andries (Patrick.Andries@xcential.com)
Date: Sat May 08 2004 - 07:31:36 CDT
Peter Jacobi a écrit :
>Patrick Andries <Patrick.Andries@xcential.com> wrote:
>[on tailored collations]
>>[PA] I suppose this would be true in principle, but how long before this
>>is implemented in the **actual tools** used by user such as MS Word or
>>MS SQL Server ?
>>I know with a bit of tailoring ($) other tools from other manufacturers
>>could fit the bill).
>When you get get it from other sources, why lamenting on the
>non-availability from Microsoft?
[PA] Not only Microsoft, I'm not sure that your average XML editor
(XMetal, for instance) will allow this for some time (search
undiscriminately a string you don't know is in Paleo-Hebrew or Square
>There is enough software which offloads collation to IBM ICU,
>where adding tailorings is very easy.
[PA] In principle yes, but this is still tailoring.
>A modest contribution to the Firebird Foundation or any
>decent programmer working on this OSS SQL database, will give
>you any collation for Firebird and Interbase.
[PA] Maybe, but I fear this is not really practical in many cases :
users may already have made a technological choice and that choice often
will not allow you to tailor your collation but there is already a
solution that would allow to do your work (unify, tag, use a
stylehsheet) and the proposed block is commercially marginal and there
is little hope tools will accommodate the new block for sometime.
>The argument 'we can't go this way, because Microsoft doesn't
>support it' is rather the wrong way around.
[PA] Well, it reflects real problems and begs the question : what do you
gain with "desunification" and introducing an additional block, this
introduction having a practical impact.
>And it's even not
>engraved in stone, that Microsoft won't support it.
[PA] This is true but this may take (some very long ?) time if the
non-availabilibity Khmer or French Canadian sorting is anything to go by.
Again, I'm not opposed to Phoenician in principle (it is intellectually
pleasing and cleaner), I just don't know what you gain with this
encoding that you would not be able to do today (right now, with no
additional cost) using what Dean Snyder proposed (XML tags and a
stylesheet for rendering) especially for the large bases where
Paleo-Hebrew is mixed with Square Hebrew. Not very clear to me (this may
have been explained in other emails, I will read them, apologies if the
pragmatic gain has been explained and I'm just appearing a bit dumb here).
This archive was generated by hypermail 2.1.5 : Sat May 08 2004 - 07:33:03 CDT