A very hesitant suggestion for Scrivener 4 (or maybe 5)

As I’ve followed the recent discussion of the viability of the current beta of Scrivener for Windows 3, I’ve gotten the impression that many of the most recent improvements to the beta have to do with the compile function, which is being made ready for public distribution as one of the last phases of completing the update. Reflecting on this has stimulated me to offer the following, whether wisely or not.

What I love about Scrivener is that it is for organizing and writing, and is exceptionally, even uniquely, designed to be adept at that. Final formatting and readying for a publisher can be done in a word processor, we are told, after compiling the manuscript to Word or RTF format. Except … you can format to an unbelievably precise level of detail for electronic publishing, also through the compile function. Am I right about this? I don’t write for epub (yet; I suppose the day may come), and I don’t do finalizing with MS Word for various reasons, so I’ve never used the compiler for anything, and may be off base here, in which case this next bit may be rubbish.

It just seems to me that huge amounts of programming effort have to go into compiling for epub, which has nothing to do with organizing and writing. I appreciate the need for epub and for software that converts fully written text into fully formatted electronic output. I’m just not convinced that it has to be the same as the software for creating the fully written text. Would Scrivener become a more nimbly and frequently updatable program if the compile-to-epub function were spun off into an entirely separate piece of software? Let Scrivener exist as the great tool for writers that it is. Let it export to a separate L&L program that formats the writing for electronic publication.

I’d love to hear what those who have experience using Scrivener for epub think about this. The rest of us no doubt also have opinions, but perhaps not as well informed, or just differently informed.

And happy not-eating-turkey-till-your-eyes-glaze-over-and-you-pass-out day to our friends in the UK!

It’s true that the Compile function has become much more capable since the early days of Mac Scrivener 1.0.

But I wouldn’t say that compiling for ePub offers any more options than are already available for the other formats, or that it is “unbelievably precise” relative to purpose-built ePub generation tools.

Much of the complexity of the Compile function is due to its flexibility on both the input side and the output side. It can take anything from a collection of poems to a patent application, and output it in any format from plain text to a publishable ebook. That’s an inherently challenging task. But it’s one that appears to be very valuable to a large fraction of our user base.


From an editorial POV, the sheer flexibility of Scrivener’s Compile function is its biggest flaw, at least for the mythical “average” user. Were it a book on, say, general boatbuilding, I’d probably suggest the author place all his general theory of keels and displacement and hull speed and power or sail and suitability for certain seas in a large general applies-to-all section, then sequester all the specialist information on final output–an aluminum fishing boat? a carbon-fiber cup racer? a faithful reproduction of a Falmouth Quay punt?-- into silos of heavily exampled individuality and label them with something specifically useful. Those Falmouth Quay punt folks care toss-all about those carbon-fiber spendthrifts and race-abouts, and very much vice-versa.

I’ve been at this since before Post-It notes replaced marginal paperclips, and find Scrivener superior for drafting and writing compared to all its predecessors, and its basic skeleton very simple to understand yet massively flexible, if you want it to be. But after I’ve drafted and written something using all those wonderfully flexible tools, I have but two needs: to print out a hard copy that lies ruthlessly unread for six or so months while I forget my wondrous brain’s outpourings and can see with fresh eyes that vast stretches have poured forth from a rather less wondrous orifice. Then, when I’ve fixed my idiocies, I’ll want to email a plain-Jane .docx file to my agent, which is what he expects to see, and he will then pretty it up to suit himself and send it along to suitable editors in the forms they expect to see. Every one of which, in my experience–editors mostly know other editors–will be reading my manuscript on-screen, almost certainly in MS Word, probably on a notebook or a tablet while sitting in bed next to her irritated husband, or sandwiched between noisy teenagers and menacing malcontents on the subway. One hundred percent of my target audience wants MS Word, Times New Roman, One-Inch Margins, a TOC, and nothing more.

None of those flexibility options related to typography and fonts and kerning and indentational operatics will ever mean more to me than do online advertisements for private jets or rock-climbing equipment. They’re just confusing.

Not confusing your readers (for which read Customers, is a Thing, as the young folks say. It is, in fact, The Thing. Scrivener partially does this now: Want to write a Manuscript with Parts? check. Compile to Standard Manuscript Format? Check. Then come the options, and options, and options, and options. A good chunk of this Board, since 2006, has dealt with those options, from wanting more to wanting different to wanting something that they can’t quite express but desperately seem to need.

The thing is, virtually any output imaginable is possible with Scrivener, no matter how arcane your needs. But getting to those options leaves many of us in the weeds.

Apple and Word both do do this Output Siloing fairly decently–though of course their metric reverses Scrivener’s by making you choose your Output before you’ve got anything to Output. Or even, as is the case with many acts of writing, not yet sure just what you’re trying to write at all.

Perhaps a rethinking, in Version 4 (or maybe 5) might address the big questions of What do You Want to Write? and Who Will Read It? by having the writer choose Specific Examples (again, this is partly done now), including the most Important choice: I Dunno Yet, but I’ll Decide Later. That Deciding Later–Oops, never mind about that Novel With Parts, I’m totally writing a Screenplay–should be just as easy. And choosing Output even easier: Wrote your Novel with Parts? Sending to an Editor? Here’s what a Greeked version looks like. Don’t care for that look? Here are six or seven others (at least half of which will piss off your editor, who’ll instantly change them 12-point Times New Roman with 1-inch margins all around). ePubbing? Here’s some likely looks, all fiddle-free and ready to go. Don’t care for those? Then On You Go to the Fiddle Section, where options are endless. Just drag and drop and pinch and squeeze until the Greeked text is just the way you like it, and you’ve shown those know-nothing book designers at the Big Five how a real book should look.

Writing a heavily illustrated and massively footnoted History of Etruscan Hairstyles? Here’s your Greeked example. Want those footnotes as chapter endnotes? End-o-Book endnotes? Click here. Something more outer yet so stupendously transparent that next year every historian will adopt it as her own and half will convince themselves they invented it? Drag, pinch, squeeze: Ta-Da!

Again, all these things can be done in Scrivener now for those willing to wade through a sea of fiddly options. But wading is way easier if some clever programmer (who started as a writer, let’s remember) thoughtfully places stepping-stones across the narrow shallow bits possessing firm easy banks with well-marked paths.Want to go wading somewhere more scenic or more challenging or you simply believe that, while that’s quite a nice safe place to wade across, there must be someplace, somewhere, that’s ever so much grander and possibly even safer, well, off you go. Miles of options before you cross; meanwhile, everyone else has picked that nice safe spot and is on to the chippy before it runs out of cod and they’ve run out of patience.

FWIW, there are two problems.

The first is that the rise of self-publishing means that Standard Manuscript Format For Sending to a Publisher only meets the needs of a relatively small fraction of users. And it was never really intended for anyone but fiction writers to begin with: scripts and technical publications have always had different requirements. So, for a program like Scrivener that tries to help many different kinds of writers, the number of “basic” options needs to be much larger than people who only do one kind of writing probably realize.

The second is that as soon as you get away from Standard Manuscript Format, there are a lot more “fiddly bits” than most people realize. While we don’t do citation management, it’s a good illustrative example: programs like Bookends exist in part because every single journal and technical book publisher has a slightly different reference style. Every single journal and technical book publisher also has a slightly different standard formatting template. UK scripts are different from US scripts, and movies are different from stage plays. Some books have endnotes, some have footnotes, some have both(!). And so on, endlessly. Our “basic” options are essentially a set of guesses about what formatting different niches require, and as soon as we guess wrong, that particular user is going to end up in the Compile Format Editor.


This is pretty much what specialized formatting tools are for though. There are plenty of those. Just output something decent with enough retained metadata that they can do their job properly, and it’s done.

Most people with specialized formatting needs will use a separate formatting tool anyway, since they need to format more than just output from Scrivener.

I’m very much in the same boat as above, with exactly two export needs; simplest possible Word (usually RTF) file, or a markdown export. All my actual formatting is done much later in the chain, and I have no use at all for Scrivener’s formatting options.

I very much believe in a tool focusing on its strengths, and the strength of Scrivener is researching and creating content. Formatting is as big a topic as all of the rest of Scrivener, and even at the best of times, Scrivener is not the best tool I can go buy for that. There are even free tools which are as good as Scrivener for formatting, if not better.

I absolutely disagree.

There are enough tools out there similar to Scrivener, that export to all kind different formats. Why would I want to buy two tools instead of one? Learn how to use two applications instead of one. Complicate my workflow by using two apps instead of one? Why? If I want to change something in the Scrivener project, I would need to redo the formatting in the second tool. Why would I want that? I makes no sense at all.

I would like Scrivener to go back to its root, and be the best software for fiction writing and stop pretending and trying to be what it is not – screenwriting, science paper, or multimedia organization tool. Focus on novel writing, give us tools for collaboration, revision, etc.

At the moment, it tries to be too many things at once, and that is a huge problem.


This is an argument for Scrivener to do more, not less.

This assumption, that Scrivener’s primary audience is novelists and Scrivener is only “pretending” to address other genres, is mistaken.


Then don’t use them. But understand that every single formatting option in Scrivener exists because a large enough number of users requested it.


Going back to the original post:

Speaking as a self-publisher, no, I don’t particularly care whether the compile function is more loosely integrated, as in a separate application. Speaking as a retired software engineer who occasionally had to consider such splits, it would be a technical nightmare.

Since at least 2010, when I bought it, Scrivener has been marketed to novelists, attorneys, academics, playwrights, screenwriters, and anyone else who had a long-form work to write. Loosening the tight integration of all the various output formats with the research, organising, and drafting functions is… well, consider it like surgery to separate conjoined twins. I can’t imagine that it will make the release of Scriv 4 faster.

Further, as Scriv is marketed to all the above groups, releasing “Scriv 4 the Organising and Draft Writing Tool (ODWT)” separately from “Scriv 4 the Compiler” would likely only serve to annoy anyone whose needed output format was not in the favoured “basic” few contained in the Scriv 4 the ODWT. Imagine the outcry if Scriv 5 the Compiler was late and Scriv 5 the ODWT was released but didn’t work well with the previous Scriv 4 the Compiler…

To summarise, in my opinion :

  1. Separating the code base would be fraught and likely add months to a development schedule.
  2. The limited output formats for Scriv 4 the ODWT would annoy many customers, even as the similarly limited output formats for iOS Scrivener already do.
  3. The two pieces would need to be released together anyway in order to meet the needs of installed users, thus negating any increase in “nimbleness”.

Again in my opinion, this is the sort of thing that a large corporation might do upon acquiring an application, in order to charge separately for each piece, thus increasing its revenue from its new asset. (Been there, done that.) A small firm with limited development resources (such as L&L) would be much less likely to consider it.

I appreciate all these replies, which help me get a better handle on compiling to ePub and its uses. I suppose that if I eventually self-publish electronically, I will appreciate being able to do it in Scrivener. Right now, it just seems like extra stuff holding up the core function. But what is “core” depends on one’s use of the program!


krastev, you make many useful contributions around here. As a nonfiction writer, however, I would like to humbly suggest that, once you have extricated your head from where it is currently lodged, you may want to go soak it. :stuck_out_tongue:

I don’t, no worries. But it’s a bit sad to me that the development of all those things for the Windows version is taking up so much time and causing so many delays. It’s not even what Scrivener is good at.