among others I use LyX for translation work and as I saw on users list I'm not
the only one. There are few features which would help me a lot, namely:

- displaying the appropriate parts of texts in parallel in some sensible GUI (I
  had to run two windows or boxes together now)
- to have the possibility to typeset these texts in bilingual mode, i.e.  the
  corresponding parts of text facing on pages to each other with the line
  numbering aparatus around either for proofreads or for the final typeseting.

I'd be happy to listen some feedback, whether is this extension desirable/
acceptable for LyX.

Also I have already put some code effort in both gui paths, but still not sure
which one is the better so any additional comments to advantage/disadvantages
below could help me to decide.

Work can be split into relatively independent parts.

1. Ledmac support. 
The aim of this package is to typeset critical text edition. It has rich
possibilities of footnote aparatus and for our purposes is important
line-counting feature in the margins. This package is used by Ledpar then.

I have no ambition wrt the footnote stuff used in the critical editions, but
its easy to add it once ledmac support is implemented. I'm already in contact
with Adam, who need it for his work and admitted he would implement it once
the customizable insets work.

2. Ledpar support.
The aim of this package is to typeset paralel texts on facing pages/columns
using the features included inside the ledmac package.

GUI representation.
I see two basic possibilities how to represent the parallel text inside LyX.

1. Reuse/inherit the tabular inset with one/two fixed-width columns and each
cell dedicated to one paragraph of translated text (this is the basic atom for
ledmac/ledpar packages)

Advantages: relatively easy to get inset with basic properties we want. We can
get eg gui line counting for free, each paragraph can easily have its own
ledpar settings easily by adding new columns etc. Keeping synchronous
paragraphs in one row is gratis too.

Disadvantages: not so easy to tweak the machinery inside tabular code.
Things need also to be tweaked are e.g.
- copy/pasting of text needs to be cut on single paragraphs through the column
  (important feature as user usually have the original version which wants to 
  This is for free in Box solution. Is this portioning of pasted text 
  doable ? 
- different actions on e.g. enter

2. Hierarchy of three Box-based insets (eg
http://www.mail-archive.com/lyx-devel@lists.lyx.org/msg127619.html )

ie, one enclosing inset for ledpar and two nested of ledmac type for
the factual texts.

* It allows to use (and develop) the ledmac inset self-contained
  (there could be just one flag which discriminate insideness-in-ledpar and
  flex the latex output appropriately for that - the code is structurally
  identical, only the keywords change).
* Existing Box code is short and not complicated to tweak. 
* We get natural box inset which other users can use for other purposes
  as is line counting facilities of ledmac.

* There will be problems how to deal with preferences for the individual
  paragraphs (e.g. "dont count this one"). Don't see how to solve this
  elegantly now.
* We wont get row synchronization easily. One idea is not to implement exact 
  synchronization of paragraphs but to have two assistants: 
   - keyboard shortcut (eg Tab) which move the cursor inside the other ledmac
     inset and put the cursor on the corresponding paragraph 
   - kind of vertical-space inset, which can shifts the text appropriately.
     Such kind of inset may be needed anyway, as we need to have some inset
     for paragraph omission (command in ledpar). There could be only some check
     box  eg "Only visual space"...
* As stated in the post above it is nearly impossible to make Ledpar-box
  invisible for cursor movement and editations; thats not much aesthetical and
  the user could do nasty things eg deleting one of ledmac insets.


Reply via email to