Some degree of Word Style export for Scrivener 2pt0...

Hi, Keith. Oh dear. I know this topic is a sore thumb because Scrivener is for writers, not for formatters. I’ll get to my request shortly but I think you need to know why I’m requesting some degree of style sheets for Scrivener and export to RTF. And for full disclosure, you need to know that I have just downloaded Scrivener and haven’t even paid for it yet :slight_smile:.

The thing is that Tex language and that Multimarkdown thingy are just way too geeky for us book writers. For my first book, I did NOT use Word for page layout (god forbid), but I did use style sheets and took good advantage of them when exporting them to QuarkXPress. In fact, I’ve done most of my own layout in QuarkXPress (moving on to InDesign) and I can handle page layout programs a lot easier than Page Markup language, which makes me gag. I need some degree of WYSIWYG in my drafts, or at least some semblance of it, where “citations” look distinct from “standard text”.

I really need Scrivener because my second book is coming together much more “raggedly” than my first. The writing is much more stream-of-consciousness and I need Scrivener’s unique ability to adapt to my changes and inspirations.

All I need are a few style sheets, such as “citation” and “quotation”, “header” and “subhead”. My proposal is that since Scrivener already has the ability to export footnotes and a few other pieces to rtf, and since Scrivener already has a formatting tool in its screenwriting mode how difficult would it be to export those screenwriting styles to rtf styles of the same name? Maybe I could skip the MS Word step altogether and get things directly into Quark.

What do you think?

Part II: The rewrite problem.

When I produced the second edition of my book, I discovered some serious missing links in the production tools. Although my editor and I worked in Word, I did final editing in Quark itself after layout and proofreading, prior to creating the PDFs for printing. This means that there was no final draft available in Word. I discovered a program called “Copyflow Gold”, which was able to take exports from Quark, convert them back into RTF and then allow the rewrite in Word. It was not much fun (it didn’t work that well) it was expensive, but it got the job done. I don’t want to repeat that process. So, before Scivener version 3.0 it would be nice to see if we could use it to reimport rtf and convert rtf styles back into scriptwriting styles for the rewrite. What do you think about that?

I’m crossing my fingers, as I have great respect for those who want to keep Scrivener a simple and direct writer’s tool. The great thing is this proposal won’t disturb the look and feel of Scrivener at all! You already have a form of useable “style sheet”, it’s just the export and import we have to look into.

Thanks for thinking of this,

Bob Katz 407-831-0233 DIGITAL DOMAIN | “There are two kinds of fools,
Recording, Mastering, Manufacturing | One says-this is old and therefore good.
Informative CD MASTERING WEBSITE | The other says-this is new and therefore | better.”

No trees were killed in the sending of this message. However a large number
of electrons were terribly inconvenienced.[/code]


Thanks for your suggestions. The problem really is that Apple’s text system doesn’t properly support styles and Scrivener uses the Apple text system (and as a lone developer I don’t really relish the idea of writing my own text system!). Although RTF supports styles, Apples RTF exporter does not - and again, a custom-built RTF importer/exporter would be an awful lot of work for a single developer.

So, although I can understand why some users would like full style support, the main problem for me is a trade-off between the amount of time I would have to spend on implementing styles (a lot) versus how much benefit such an implementation would bring (a good deal for some users, none for myself - as I never use styles - and it’s also a step in the direction of word processing that I would rather avoid, as trying to turn Scrivener into a full-blown word processor would be insane - maintaining everything a word processor can do alongside all of Scrivener’s unique features is just not the sort of Herculean task I’m really up for :slight_smile: ).

Hope that makes sense.
Thanks again and all the best,

BobKatz, what sort of book are you writing?


I too am looking for the Holy Grail of the One Good Word Processor. I tried Scrivener today and it perfectly reflects the way my brain is organized. Having said that, discovering that it cannot export Styles was a huge disappointment. I totally understand and respect the one-man-band limitation (I am in a similar situation) - however there may be some hope:

  • There are several low-cost RTF-to-XML (and back) converters out there. Some are specifically developed for interface between Word and the Web. Is it something you might be able to incorporate? I personally would be willing to pay more for this feature.

Other items in my wish list:

  • Better implemented numbering of Chapters
  • Keyboard shortcuts - especially for navigating from the Browser to the text field and back.

These additions would make Scrivener a real contender in the field.

Many thanks

… what do you mean by ‘better’ exactly?


If you could point me in their direction, I’d certainly take a look. Though it would mean a lot of messing with the given text system in order to attempt to get it to recognise styles. This would be difficult enough, but given that in Scrivener the text is scattered across hundreds of files, it would be a nightmare…

As for your other suggestions:

• As Wock asks, what is the problem with chapter numbering? The auto-number tags (<$n> etc) work very well. It’s difficult for a program such as Scrivener to off live-updating auto-numbers given that the numbers may be contained in lots of documents that haven’t even been opened in the current session. 2.0 does make it easier to have all of these tags handled at the compile stage, though, so that you don’t need to put them in your document titles at all.

• Keyboard shortcuts. You can take control of all keyboard shortcuts via the System Preferences (Keyboard & Mouse - Keyboard Shortcuts pane). That allows you to override the current shortcuts or add new ones. You can currently navigate between views using the View > Navigate To menu. 2.0, however, adds Opt-tab as a way of cycling between the binder (browser) and main editor, which is what you want if I understand correctly.

Thanks for your interest in Scrivener and for posting your thoughts,
All the best,

Hi Keith

and thanks for the quick reply. I’m definitely going to take my time and familiarize myself with the Scrivener interface. Could be that I’m still stuck in old thinking modes and that I won’t need to go back to Word at all (which would be great). I found a lot of great info about shortcuts in these pages and I’ll try to get into the numbering before I bother people again.

If you’re interested, a Google search with XML & RTF in the box will bring up a number of sites that offer info and/or software that will convert RTF to XML (and Style Sheets) and back. My thinking was, none of this needs to impact the internal organization of your software. But for people who need organized headings, it could be something that happens at the compile stage, and the rules could be simple: Folder equals heading, text equals… text. You already have a strategy for setting up Folder level styles, so…

But the important thing is - this is a great piece of software and you should stick to what you want to do with it. Hopefully you can keep it solid and reliable and get some more traction with it. I’m glad I found it and it’s a great deal.

Again, my thanks and good luck.

Thanks again, Savvas, for the kind words about Scrivener.

Actually, if it’s a matter of converting RTF to XML, I doubt it will work, unfortunately. The problem is that although RTF supports styles and headings etc in theory (I believe), the Apple RTF exporter (which Scrivener uses) does not. So RTF exported from Scrivener, TextEdit etc has no styles information in it. My guess is that any XML converters would be looking for style information in the RTF files themselves, which won’t exist in this case. (I tried a search on XML & RTF in Google, but there are many hits and none of them seemed immediately obvious as having much information on this particular issue.)

All the best,

If you want XML output, I recommend looking at the MMD (MultiMarkDown) exporter. It doesn’t really read the RTF styles in your documents, but I believe it recognizes the basic bold/italic as well as Scriv’s annotations, and pulls headers from the document names in the binder to build a properly syntactical structured document.

From MMD/XHTML there are filters to convert your document to several other formats. Not including word I’m afraid, but if word can read XHTML with a stylesheet then there might be a possibility there, or one of those XML->RTF converters you mentioned.

It doesn’t do automatic bold and italic interpretation, due to the differences in philosophy between syntax and RTF formatting. With the latter you can create strange combinations of bold and italic which can cause unprdictable results. So the conversion tool is a manual operation that the user can execute in Scrivener and then immediately preview to make sure it all came out okay.

But yes, if you want XML more than RTF, there is little reason to not use Scrivener’s MMD features, as that is designed to produce a valid semantic XML file (and a few derivative formats off of it, using transforms).

I think MMD might even be useful to those who just want RTF and who have no desire to fiddle around with XML or LaTeX (read: the non-geeks). The example I’ve recently posted shows, that the MMD HTML export is almost good enough to suffice as an interchange format for OpenOffice. The support for styles is already significantly better than what you get with Scrivener’s default RTF export, so this could make the case for MMD.

Bild 1.png
There are still unsolved problems: First of all, Open Office does not automatically resolve the link to an image, which means that the image will show up in a document, but it will no get embedded (only the path is stored, not the actual image). Footnotes are also still unsupported. The good news is, that OpenOffice has a special syntax to store footnotes in HTML files, so this problem could be solved, either with some XSLT wizardry or by changing MultiMarkdown directly.

I would certainly agree with that statement, but I understand that most people have an aversion to old-school WordStar era word processing. Everyone wants things to look roughly like a printed page looks, now, while they are working. I don’t think that makes any sense at all, personally (especially when making something bold means nothing at all except a font change), but other people think asterisks and pound signs are ugly, too.

Fundamentally you are right. If everyone was using MMD we wouldn’t have all the gripe about Apple’s poor RTF engine when it comes to stylesheets and so forth. Hey, if Apple actually cared about their text system, we’d have rudimentary semantic styles in the ruler, proper OpenDoc export, Pages integration, the stupid text smear when scrolling bug that has been in there since Panther would be fixed, and so on. There are open source RTF editors that do a better job than Apple’s monumental kludge.

I’m guessing the problem is that Scrivener has gone way past its original ‘novel writing’ roots. Personally, I just use it for story writing, so I don’t really need to worry about formatting while I’m trying to write.

The gripes are quite legitimate when you’re talking about the text engine in my opinion. I was amazed (though not for long) to discover that you can’t even set the document language from the Apple text engine, so there’s no getting around it; the Apple text engine is poor. Fortunately, folk seem to understand this, and don’t blame Keith for not including the extra bits they need.

Yes! I’ve seen the smudge! I wondered what that was all about. The problem as I see it, is that fixing the text engine is fairly low priority, along with everything else that would enhance the desktop. If the problem affected the iPhone, then it would be a different story.

It’s a pity that the Mac doesn’t really have a market for third-party frameworks either, then at least someone else could pick up the slack.

There are rumours that the text engine is being improved for Leopard. I haven’t signed up for the developer program yet (I will when 2.0 is more complete so that I can test on Snow Leopard before the release of either), so I’m not breaking any NDA here, just reciting what I’ve read on rumour websites. Old news, but this is it:

That’s all there is, though. I am hoping that, in keeping with the whole point of Snow Leopard, they will also take some time to fix the long-standing underlying problems. What I really, really hope is that they aren’t just going to throw on auto-correction and replacement and leave the dreadful tables and lists as poor as they have been since their introduction in Tiger, etc…

All the best,