Re: new help browser

[Andy Smith <ams70 cam ac uk>]

Alexander Kirillov <kirillov@math.sunysb.edu> writes:

> The only problem with this is that then the help browser needs to grep
> through all the words on screen and compare them with the glossary all
> the time.

This might not be too hard.  When you're scrolling through the
document, the browser only needs to check what just disappeared off
the screen, and remove glossary entries if the last reference
disappeared (refcount?), and similarly add entries for words that
appeared at the other end of the visible area.  And if it's being done
with a dynamic Docbook->HTML generator, presumably it could remember
where the <glossentry>s or whatever are in the vertical continuum, and
keep a sorted list or something.

> Another version of the same idea is that for each <sect1> (or the
> whole document?) of a document, make a list of all the words in it
> that might need explaining, so that when you read, say, a section on
> GMC, in the left pane you have a list of words in this section that
> appear in the glossary (symlink, mount point, etc); you can click on
> any of them, and it will show you a definition? I think, it is a
> better idea than actually having all occurences of the word symlink
> somehow highlighted so that the reader needs to click on it.
> 
> 
> Now if all the docbook wizards and developers could tell me if this
> could be implemented...

I think this could be done with a DSSSL.  Have a stylesheet or mode
that ignores most elements, but generates a bit of text for each
glossentry and section header, so the final result could be something
like:

    <id of some sect1>
        symlink
        mouse
        device
        ... etc
    <id of another section>
        more words...

Or generates some XML if that's easier to deal with.  Then for each
section the help browser could look in the list and pull each
definition out of the glossary and put it in the left pane.

    Andy