Hi Paul,
Am Dienstag, 18. Oktober 2016, 20:38:00 schrieb Paul Hoadley:
>
> What is the state of the art for small text substitutions (basically just
> domain-specific terms) that can also be varied via the profiling machinery?
> I have an application that can run as more than one “product”, and each
> product shows domain-specific variations in labels, for example, in the
> user interface. There might be a generic term for some object “Foo”, but
> Application A will call this a “Bar”, and Application B calls it a “Baz”.
> What I want to be able to do is refer to a “Foo” throughout the source, but
> substitute “Bar” for App A’s output, and “Baz” for App B’s. What’s the best
> way to achieve this?
This sounds like profiling in combination with entities.
You could define your foo entity like this:
<phrase
os='a'>Bar</phrase><phrase
os='b'>Baz</phrase></phrase>">
The <phrase> element is quite neutral and is allowed where inline elements are
also allowed.
When you have this "foo" entity, you can refer to it throughout your document
as &foo;. Of course, you need to set the profiling attribute os accordingly to
filter it for application A or B.
We use it for different product names and numbers which applies the same
approach. It works great.
Keep in mind, this approach relies on a "side effect": it depends on phrase
being allowed in all contexts. This is not always the case. For example, you
cannot use &foo; inside attribute values. Or, if you have customized the
DocBook schema and you don't allow phrase in, let's say, a title. That won't
work as you will get validation errors.
Apart from this, I think you will be happy with this approach. ;-))
--
Gruß/Regards
Thomas Schraitle