Hi,
The following is related to the changes to capitals and emphasis[1] as Michael
Gray described them a few months ago.
I was thinking about an opcode to specify which characters do not reset a word.
Currently, everything that isn't a letter will break a word. As far as I can
tell this is handled in a single function resolveEmphasisResets() that is
called only for caps. Thus, the behavior for emphasis is slightly different.
As an example, the word E.T.A. should have one capsign in front of it because
the period doesn't cancel caps. This doesn't work currently, however it does
work for emphasis, presumably because word_reset isn't handled the same way.
This conflicts with the Dutch braille standard because:
1. As Bert Frees pointed out[2] not all characters cancel a cap or
emphasis sign.
2. In Dutch caps and emphasis are handled the same with regards to
word_reset.
Since most characters cancel a cap or emphasis sign, it seems logical to
instead define an opcode for characters that do not cancel these signs.
Initially I was thinking of adding an opcode wordmodechars with very similar
behavior to numericmodechars. This would handle both caps and emphasis. But
that will not be correct for UEB if Michael's code is any indication.
So some alternative solutions are:
1. Have two opcodes, capsmodechars and emphmodechars.
2. Have many opcodes *modechars, where * is caps, ital, bold, etc.
I'm leaning towards (2). In either case, it looks like resolveEmphasisResets()
needs to be called for emphasis as well. This should be fine as long as that
function doesn't have any hard-coded logic for splitting words based on
character attributes. To my knowledge at the moment the function considers any
non-letters to mean a word reset.
I don't want to break UEB with this change though, so I am putting the proposal
up here for discussion. Any feedback would be appreciated. For instance, what
do other languages need? Is nocapsmodechars indeed preferable over
capsmodechars (i.e. the exact opposite)?
Davy
References:
1.
//www.freelists.org/post/liblouis-liblouisxml/CapitalEmphasis-update
2.
//www.freelists.org/post/liblouis-liblouisxml/CapitalEmphasis-update,1
-----Oorspronkelijk bericht-----
Van: liblouis-liblouisxml-bounce@xxxxxxxxxxxxx
[mailto:liblouis-liblouisxml-bounce@xxxxxxxxxxxxx] Namens Bert Frees
Verzonden: donderdag 25 juni 2015 16:05
Aan: liblouis-liblouisxml@xxxxxxxxxxxxx
Onderwerp: [liblouis-liblouisxml] Re: Chat between APH and Nordic Braille in
DAISY Pipeline 2 project
Christian Egli writes:
Procedure
~~~~~~~~~
- Create a branch in the master repo so that everybody can work off
the same branch and integrate via pull requests
typeform support for harness tests
----------------------------------
Bert will look into the harness test to see whether it can be made to
support typeforms
Opcode unification
------------------
- There is a proposal to merge all the emphasis opcodes into a few
generic ones.
- Bert will post a mail to the list with the proposal for further
discussion
- This is quite a sweeping change but it would be good to combine this
with the UEB changes
Backwards compatibility
-----------------------