[liblouis-liblouisxml] Re: Issue with emphasis

  • From: "Michael Whapples" <dmarc-noreply@xxxxxxxxxxxxx> (Redacted sender "mwhapples@xxxxxxx" for DMARC)
  • To: liblouis-liblouisxml@xxxxxxxxxxxxx
  • Date: Thu, 20 Nov 2014 15:57:15 +0000

It would be useful to know why pass2, pass3 and pass4 are done after applying emphasis markers and if it is required. This is the sort of thing I need to know if to make a suitably informed decision.


Michael Whapples
On 20/11/2014 15:02, Bert Frees wrote:
Michael Whapples writes:

I think this is sort of what I was meaning by an external option. Well
may be I was thinking it could potentially go further or it could be as
you describe.

May be it could be C may be it could be another language, if it is not
mixed in then this does not matter too much, and whoever writes it could
use which they feel is most appropriate. "Appropriate" including
ensuring that it is usable by those who need access to the
functionality, so realistically it probably would be a choice between C,
C++ or other language which can be used from C and has no large
dependencies.
Important consideration is that the liblouis table, which should contain all the
information needed to perform the translation, that means including emphasis
markers, should only need to be parsed once.

May be it could be called near the end of the liblouis translate
function, then consumers of liblouis need not deal with additional
function calls and that complexity.
Absolutely! We definitely need to avoid introducing additional API calls, we're
talking about implementation details here and they should be hidden from the
user.

I'm not so sure this can just be a post-processing function though in the sense
that it would be called near the end of the translate function. The way it works
now I think is that pass2, pass3 and pass4 are applied *after* emphasis markers
have been added. Not sure if this is really a required feature, but something to
keep in mind.

If it is C or C++ may be it could be statically linked to liblouis and
then no additional dependencies are to be handled.

I think I am getting more convinced that it might be the quickest option
to implement.

The only bit I cannot quite work out is whether the Braille string and
typeforms would be enough to apply the emphasis indicators or whether
one needs to have access to the original text.
For a description of the software, to download it and links to
project pages go to http://www.abilitiessoft.com

For a description of the software, to download it and links to
project pages go to http://www.abilitiessoft.com

Other related posts: