Tino Schwarze wrote: > On Wed, Aug 23, 2006 at 10:42:36AM -0400, Nicholas Vettese wrote: > >> Can, or will Scribus be able to edit a PDF directly? Can I now open a >> PDF that someone sends me and change the information, or will this be >> available in future versions. > > Probably not - PDF is not designed for editing. Therefore, it is very > complicated to build something close to a reliable PDF editor.
Indeed. In fact, even Adobe can't do it well - Adobe Acrobat's PDF editing facilities are incredibly limited. The PDF editing and preflight add-in PitStop from Enfocus adds a lot more capabilities, but can be quite unreliable (I've had it subtly corrupt PDFs in rather nasty ways). If you know of a capable _and_ reliable PDF editor I'd love to hear about it. > Scribus would have to support every feature of PDF so no information is > lost. Not necessarily. It could potentially extract content streams it understood, and store the others unchanged in the document for output straight back to the final PDF. Unfortunately even what you'd expect to be quite simple in PDF - like text - is a royal pain. There are so many ways it can be structured and encoded, so many different ways to format it, etc. It can be quite hard to take a block of text and figure out what order the parts go in, what the shape should be (columns, flow-around, etc), and sometimes even what the characters associated with the glyphs are. In fact, it can be hard to tell it's even text if it's been sent as outlines. Also, if fonts have been subsetted, it might not be possible to edit the text without replacing the font, and that means trying to adjust the text layout without messing anything else up. I think there's a good reason there are no decent PDF editors out there for any money. It's a format that's not designed for editing in detail - though it's quite good if you just want to remove and replace whole objects, or if you want to add to it. Personally I expect that Scribus's eventual PDF editing capabilities are likely to be limited to embedding pages from PDFs in its PDF output natively (without rasterisation), and possibly some document assembly. Both of those are quite hard enough (and you need a good, fast, reliable PDF parser too). Even that's an unknown distance off into the "it'd be nice to have" future, since everybody has other scribus and non-scribus related things to work on that take higher priority. -- Criag Ringer
