Re: A character is in the eye of the beholder
At 11:13 28/10/96 +0100, Martin J Duerst wrote:
>>b) The same database is queried from France and Quebec. The database returns
>>its answers in caps, though the data itself is in mixed case, and contains
>>accented characters. What rules should be applied to the display of accents
>>in these cases, and how does this get determined from the language
>>preference statement of the HTTP request?
>First, the database shouldn't return its answer in caps, if the data
>itself is in mixed case.
There may be good reasons why they might want to do so. For example, they
may want to present the data as a single sorted list, which has been
canonicalized based on one case. In the case of databases of SGML data there
is a default set of rules about converting lowercase attribute and element
names into caps during parsing. If what the database is returning is a list
of SGML names it might be useful if it applies these rules prior to
selection. For HTML you might want to pass back a list of element and class
names used for example. While class is case dependent element names are not.
How do you sort such a mixed list for display? OK its the server doing it,
and should not worry the client, but the client still needs to be able to
control the form of the result sent to it. I'm not saying now is the time to
specify this, only that now is the time to start thinking about it before we
define locales in a way that won't allow for future expansion to cover
features we know would be nice today if the systems were clever enough.
Martin Bryan, The SGML Centre, Churchdown, Glos. GL3 2PU, UK
Phone/Fax: +44 1452 714029 WWW home page: http://www.u-net.com/~sgml/