Re: [xml-dev] Caution! XML parsers behave differently with whitespace specified directly in attribute value versus whitespace specified via an ENTITY

On 04/08/2016 01:31 PM, Christophe Marchand wrote:
> The question is, why, today, as we have various tools to modify xml
> files, should we persist to use entities !

Those of us who deal principally with publishers' document and
documentation XML have inherited practices that were established in
SGML, and while they continue to work, I think they see little impetus
for change. I encounter three main uses:

1. In environments where W3C Schemas offer no real advantage over DTDs,
entities provide a convenient tool for the management of
frequently-reocurring constructs, whether they are 1:1 replacement texts
(as Liam describes), system references like images, or external document
fragments.

2. When handling documents for multiple outputs, entities allow the
relevant replacements or inclusions to be done at parse time, lessening
the amount of if-then-else work required in subsequent processing.

3. Parameter entities are used both for the switching from one set of
character entity declarations to another in a DTD, depending on the type
of output required; and for the construction of customisation layers for
centrally-managed DTDs (shims).

With more manipulable formats of definition like RNG, much of this can
be avoided, but to do so would mean retooling large chunks of business
processes, and the majority of the users I deal with tend to take the
view that if it ain't bust, don't fix it.

///Peter