[
Lists Home |
Date Index |
Thread Index
]
Ben Trafford wrote:
>Without -something- in the XML to indicate author preference for link
>behavior, how can they know what the end result will be? How can they
>author the content?
Compare the following very similar question, and you'll see that you already
know the answer: without something in the XML to indicate author preference
for fonts, bolding, and page margins, how can they know what the result will
be? How can they author the content?
>If Acme Docutech has a common stylesheet, the authors are going to
>know what ends up bolded, what ends up italicized, and what the tags
>are for that, whether they be <strong>, or <span class="strong">,
>or whatever.
And if a common stylesheet adds linking behavior, they'll know what the end
result will be as well. Hyperlinking is presentation, and presentation
varies from output medium to output medium, so we store semantic metadata
("subtitle" instead of "14 pt bold helvetica", "endnoteRef" instead of "a
href", "footnoteRef" instead of "javascript:popup()", etc.) and then convert
the semantic metadata to display metadata in the stylesheets. Then, when new
media come along, we can use the same content on them by simply writing new
stylesheets.
Element and resource relationships should be stored using semantic
descriptions for the same reason that the elements themselves should be
stored using semantic descriptions.
Bob DuCharme www.snee.com/bob <bob@
snee.com> weblog on linking-related topics:
http://www.oreillynet.com/pub/au/1191
|
