Compiling Scrivener manual by myself looks different?


I am actually playing with the downloadable scrivener manual to try that project as a base for my own document. I do this, because it is a fully written technical document for a software and has a basic but nice layout with a well made front matter and toc.

But first, several images are missing and I recreated them by making new screenshots of these out of the ready compiled PDF. That is a one time job I just mention here: Why not prepare a full project including all images to be compilable on a different computer (that is not the developers one)?

But more important is that I am unable to get the same looking nice front matter with the first page and two column toc. Why is that do?

I am currently running the app on Snow Leopard that may the cause and multimark down is

stsmac:~ lothar$ mmd --version

MultiMarkdown version 5.4.0

Any ideas?

Thanks, Lothar

From what I remember, Ioa has some custom post-processing scripts that handle some of the formatting details that cannot be handled by Scrivener 2 or MMD. These are not included in the project itself.

The Scrivener 3 manual is even more beautiful (Ioa did an amazing job), and more of the formatting was done in Scrivener — thanks to all the new features in S3 compiler…

I will be very interested to see the scrivener project for the V3 manual, but I suspect Ioa will need some time to sort things out before (hopefully) sharing it.

Thank you,

I think I should getting it started. Today I also have seen some formatting youtube videos that demonstrate overriding font and some other stuff. And thus it is probably not that important in the beginning.


The main issue that you are running into is that this is a very old project—or rather it was initially designed a long time ago, when MultiMarkdown was at version 2—a substantially different beast than what it is today. I have a copy of MMD2 that I keep around just to compile that project. That said with the exception of maybe the logo image as you note, everything you need should be in the project (or obtainable from CTAN). It’s an XSLT file that generates the .tex, not MMD vanilla. Some of it is picky as well, but that project had a +40 item post-compile checklist of things I would have to do in a text editor to get it typesetting correctly. The idea with these projects has always been less about making a simple demonstration of how to make a PDF that looks just like ours and more about demonstrating the working techniques inside of the .scriv itself that can be used to create source material such as the stuff we use to make those PDFs.

But, if you want a better and more modern practical example I’d recommend the Scapple user manual. It doesn’t use multicols I don’t think for the ToC, but that’s nothing fancy:



The Scapple manual has the entire preamble in the compile settings and works with MMD 5 (I think, I haven’t tried in a while since I updated it to v3), though it may still be missing a few PNG files, I don’t recall.

The v3 Scapple .scriv will be ready and downloadable tomorrow. It’s not a really good example of what v3 is capable of—more a quick and dirty hack to get an old v2 project generating a PDF to our current design, but it works. And it’s sitll not going to be easy to create a PDF that looks just like ours. For one thing this time around we bought a brace of hand-selected fonts rather than using freebie TeX stuff. We obviously didn’t buy the rights to distribute those fonts though.

The new Scrivener .scriv is going to be where it is at, but I’m afraid to say it might be a few weeks (very optimistically) before it is in a presentable form. There is a lot of the duct tape you would expect from a project that has spanned back into alpha days—workarounds that are no longer necessary, etc., and the inevitable mayhem of a midnight-oil-for-months type endeavour. It has something like 700 lines of LaTeX preamble that could surely be optimised a bit.

I’ll have to give the images thing some thought. The problem is bulk—the new manual has close to 300 figures, and I work in high res (+180 DPI). It’s close to a gig of source material that’s all of course linked from the file system. Given how links work in Scrivener, it’s not like HTML where you can point to spot with a relative address. I could have done things that way, since with MultiMarkdown stock you can do just that, but I like having the thumbnail in front of me when I’m writing about it.