Re: Programming Lout Backends
Jeff Kingston <[email protected]>
| Newsgroups | gmane.comp.type-setting.lout |
|---|---|
| Message-ID | <[email protected]> |
On Tue, Dec 16, 2008 at 07:07:35PM +0100, Christoph Angerer wrote: > > My idea would have been, to use lout as the "lingua franca" and > provide different backends in lout to generate the formats directly > instead of trying to convert postscript files into mobipocket files > > My question is now: does anybody think this is feasible? How difficult > is it (in practice) to write another lout backend? As I said, the > input documents are?at least for now?rather simple texts with sections > and subsections and maybe sometimes boldface font or a quote. > > In the lout source code, backends look pretty well modularized; I am > thinking of the typedef struct back_end_rec in externs.h that seems to > provide some plugin functionality. So is it as simple as providing > implementations of those 26 functions to produce a file in a > completely new format? Or are there some caveats or am i missing > something? (Since the PDF backend is kinda discontinued it makes me > think that this might be harder than it looks...) I don't think Lout is well suited to the job of being a universal document markup language. I can't think of any language that is well suited to that task, except Nonpareil, which does not exist yet. There are a lot of issues hiding behind the 26 backend functions. My suggestion would be to make a design that allowed multiple formats and plugins that convert those formats to other formats. In the end I think you want to be displaying PDF, because it can display all kinds of fun things, it is a moderately reliable standard, and there are converters from virtually every format to PDF. But of course you can't store documents in PDF format, because they are already laid out in a fixed format. Jeff