[liblouis-liblouisxml] Re: Capital/Emphasis update
- From: "Susan Jolly" <easjolly@xxxxxxxxxxxxx>
- To: <liblouis-liblouisxml@xxxxxxxxxxxxx>
- Date: Tue, 24 Feb 2015 12:53:08 -0700
Thanks so much Bert for the feedback.
I agree with you that it would be nice if braille systems could avoid
context-dependent rules. However, I expect that achieving that would be a
much more difficult problem than it might appear for both braille-specific
and general reasons. Braille systems have many constraints given that they
are intended to be human-readable which I believe to be the reason for at
least some of the context-dependent rules. New braille systems also have
some obligation to be both backwards compatible and future proof in order to
avoid unnecessary difficulties for already-fluent braille readers.
As for general reasons think of what we've learned about the complexity of
digital encoding of documents. The history of the Text Encoding Initiative
(TEI) illustrates this as does the fact that even DocBook has been under
development since 1991. Recent examples of the discovery that problems can
appear more complex the more we learn include I18n, ePub, and the Semantic
Web.
It is important as far as understanding the current situation to remember
that many braille systems were defined long before the possibility of
automatic translation became a reality. Also, at least here in the US, many
early transcribers enjoyed manual transcribing and were very proud of their
skills even sometimes to the point of hindering the development of automated
translation.
From a historical perspective braille transcription was sometimes viewed as
publishing. So just as a quality print book requires a human editor, it
shouldn't be surprising that a quality braille book would as well. My own
feeling about certain rules, such as the definition of a passage, which are
intended to make braille reading more efficient and/or enjoyable is that
these rules should be tagged as recommendations and not requirements.
And, of course, I'm sure we could never get universal agreement on what
makes a transcription more or less readable. Just remembering that there
were riots in the UK when they decided to require braille capitalization
indicators in all contexts underscores this. In my opinion some of the
arguments both against and for that change were very well-reasoned.
I do think it extremely unfortunate that UEB wasn't designed with
considerable input from persons experienced in modern digital encoding.
Here was an opportunity for a new braille code that looked toward the
future; instead we are ending up spending huge resources changing to a
system little different from what was originally proposed as UEBC back in
1991.
However, given that we seem to be stuck with UEB, I think it would be a
worthwhile exercise to determine what, if any, additional markup would need
to be added to the latest ePub specification in order to support fully
automatic transcription to UEB translation plus some standard formatting
prescription. (Or has the DAISY Pipeline effort already completed this
exercise?) I have two reasons for this suggestion.
First is that there is growing expertise outside the braille community for
producing rich ePub documents including developing efficient methods for
generating information not present in the starting electronic document.
Second is the standard argument for separation of concerns. There is more
than enough braille-specific detail for braille table designers and software
developers to handle without having to compensate for avoidable deficiencies
in input.
HTH,
SusanJ
For a description of the software, to download it and links to
project pages go to http://www.abilitiessoft.com
Other related posts:
