On 01/15/2010 08:02 AM, Steve Litt wrote:
On Friday 15 January 2010 07:34:02 Philiрp Rеichmuth wrote:
Am Fri, 15 Jan 2010 10:28:10 +0000 (UTC) schrieb Guenter Milde:
If ever we want to switch to a XML like math format, it should use
Unicode where available. For an incremental transition, we could use
Unicode chars in LaTeX format right now.
I think this is a fundamentally bad idea. I use LyX with XeTeX a lot these
days and it's definitely nice to be able to distinguish between Unicode and
the LaTeX way of getting certain characters.
If one does a big switch like XML, that is precisely the moment to think
about such conceptual changes as well. Otherwise, trying to induce some
kind of incremental mode of transition is more likely to confuse things and
introduce intermediate instabilities than anything else.
Philipp
Possibly in line with what Philipp is saying, indexing has turned into a holy
terror where you're forced to include a whole bunch of ERT as workarounds
where formerly it just worked, and it seems to me like the explanation is
always "we have to make it work with Unicode."
The reason for the indexing transition had little to do with Unicode. It
was that what goes into the argument is LaTeX, and we want people to be
able to enter it the way they enter other LaTeX in LyX. Think e.g. about
entering math in an index entry. The old way, you had to do $2^x$ and
the like. Now you use math insets. As a side-effect, you can also enter
Unicode more easily, and this matters rather a lot to many of our users.
Indeed, one might put this by saying: The old way, it was ALL ERT. Now
only some of it is, and none of it is unless you are using more advanced
features.
We all agree that the way things like sort order, "see", and so forth
are handled could use improvement. The way forward is to create a kind
of insets that basically contains other text insets. Or a dialog box
that embeds LyX work areas. The infrastructure for this exists already.
Someone just has to do it.
rh
