[Scribus] pdf or .sla to html

[Craig Ringer]

Lee Wiggers wrote:
> Hello
> 
> I am almost too new to ask a question, but I will
> anyway.  Flame at will. :)
> 
> I have spent the last week reading, learning, and using
> Scribus.  Amazingly enough, I am happy enough with the
> catalogue that resulted, that I would like to add it to
> our site.
> 
> I cannot find any reference to a working utility to
> transform a .sla or .pdf to presentable html.  I have
> tried pdftohtml with less than great results.

There isn't one, as far as I know.

Frankly, desktop publishing apps just don't lay out documents the way 
that is necessary to produce decent quality HTML. It'd be possible to 
use some CSS2 absolute positioning etc to achieve vaguely sensible HTML 
output from Scribus, but you'd have all sorts of problems with people 
with different screen sizes, with font availability, with vector 
graphics, with overlapping objects, etc. And that's assuming we relied 
on the *very* *latest* web standards, which 90% of browsers would 
display incorrectly.

All that, for output that still won't look that good, and will totally 
suck from the point of view of good web design, accessibility, support 
for very large and very small monitors, etc.

The fundamental issue is that Scribus (and other DTP apps) lay out 
objects by specifying exact positions and sizes, and they do so in 
specific real-world units. Good web pages must, in so far as is 
possible, specify every objects location relative to other objects, and 
must use flexible relative units that can adapt to the user's display 
device. On top of this, DTP is very image friendly; on the web you're 
often better off using some CSS rather than an image. DTP uses lots of 
custom fonts; on the web, you're unwise to rely on the user having more 
than a very small set of fonts, and are best off using general selectors 
like `serif' and `sans-serif' instead. They're pretty much opposite 
media, and it shows - web page editors suck badly for DTP, and DTP 
programs suck badly for making web pages.

SVG is a marginally more realistic option. It supports vector graphics 
well and is much more friendly toward the absolutely-positioned style of 
layout used in DTP. That doesn't mean it'd be good for web use:
   - Many browsers have little or no SVG support
   - The SVG text model doesn't map well to a DTP-style one
   - The output would probably be very poor from an accessability PoV
     (screen readers, etc)
   - The output would probably still not be friendly to users with
     very large/small screens or very high/low display resolutions.
   - HTML+CSS is just plain better for web pages.

This has been discussed before, by the way. I strongly suggest you 
search the archives. I've presented my view above, but I know Craig 
Bradney at least disagrees with my sentiment that a HTML exporter would 
be pointless and probably harmful. You'll find more discussion in the 
archives if you feel like digging.

For now, the short answer is that I don't think there's a good way to go 
from Scribus documents to HTML. Posting a PDF is probably your best 
option, though making a JPEG (or set of JPEGs) from the PDF for use as a 
fallback would also be wise. Even better, design the web page right 
using the same basic source material (the document text and graphics).

--
Craig Ringer