syngin <[EMAIL PROTECTED]> writes:
> The time has come to determine an index system for the new help package.
> The entire help system will be changing (as previously determined) and
> initial discussions with Sven have leaned toward an XML index and
> precompiled HTML help files.
> The role of the index will be to eliminate hard coded file links inside
> The GIMP's mighty core which will in turn lead to greater i18n
> We have now reached a point in discussions where other's are encouraged
> to comment on how this might be achieved.
Everything that's needed from the gimp-help module side is an index
file which contains an identifier --> document_index mapping for the
chosen target format (e.g. HTML).
The GIMP core itself will not need to know at all about this index
file, as the core will only reference the identifiers. I suggest a
naming scheme with a prefix identifying the help domain
(e.g. "gimp-rect-select-tool", "gimp-file-new-dialog", "gimp-foobar").
The help browser will read the index file, which, for the HTML case,
might be as simple as:
This way we remove all HTML links from the core and are able to use
any target format like PDF.
The browser gets passed the help identifier together with the help domain
(which will be NULL for all pages shipped with the gimp-help module).
The help domain is used for 3rd party plug-ins where it's the file
system path to the plug-in's help stuff.
To be flexible enough for future extensions or help formats, I suggest
the installed help files to follow a standard hierarchy.
so the HTML help plug-in will know that it needs to look at
the PDF help plug-in will open the
IIRC, this roughly what we discussed on #gimp a few months ago, feel
free to comment or eek...
Gimp-developer mailing list