Jean-Marc Lasgouttes wrote:
Abdelrazak Younes <[EMAIL PROTECTED]> writes:
2) The existence of a style attribute does not affect how and where I
can select text
This is a con for me. I want to select the whole charstyle
automatically and not bother with micro selection.
I am abivalent on this one. When a character style is a noun, for
example, I like the fact that selection goes in one glob. After all,
one nice thing about selection in word is the way it selects full
words when it feels it should (I think it is one of the few automatic
things they got right).
I'm starting to think that maybe what we're calling charstyles really
*is* two different things. For things like "noun", "title", "biological
species" --- the inset paradigm works well (even according to the
questions that I asked at the beginning of this thread). This is
precisely because those things are all "one semantic entity", even if
they are composed of multiple words. I will rarely want to copy half of
a title, say, together with some surrounding text.
For things like "emph" or "strong" or "code", however, the situation is
very different. For the reasons that JMarc explains below; for the
reasons explained in my original message starting this thread; for the
reason that in this case, the range does *not* represent any semantic
entity, but rather attributes applied to *parts* of larger entities. And
therefore in these cases, it is not so common for me to want to select
exactly the same range (except perhaps to undo the style, but see my
suggested solution to that in a separate message) or to delete it as one
entity; and conversely, it is much more likely for me to want to select
parts of it, perhaps also with parts of the surrounding text.
So what I'm saying is: the charstyles-as-insets is good for "noun" and
"title" and "biological species", but absolutely not for "emph" or
"strong" or "code". For those, ranges would be *much* more suited.
Now, the question is, if in both cases the underlying thing that they
(the "charstyle" insets or ranges) actually *do* is the same thing
(changing the font properties) --- and I'm not sure that it necessarily
is --- then can we share that part of the code between
"charstyle-insets" and "charstyle ranges"? I suspect we won't be able to
answer this until we have a working implementation of charstyles as
ranges, and perhaps a few more "charstyles-insets". But it's worth
considering...
OTOH, when it comes to setting large parts of text in a charstyle (and
assuming the we have the 3-box model), then I do not think this is
still the case. Assuming an example like what Helge proposed:
\emph{Sentence 1. Sentence 2.} Sentence 3.
I expect that I should be able to select sentences 2 and 3 to cut and
paste them. The block model is not as good for long text as for short
one.
Dov
