Outliner with replicants

I am working on my new novel, and this time around I am trying a different writing method. Rather than rushing through the first draft, I am writing one chapter at a time, and outlining the reminder of the novel (refining said outline each time). In this process, I discovered that it might be useful to do this with an outliner that has the “Replicate” feature in DevonThink. This is handy when the storyline is linear but the telling is not. DevonThink seems too much for such a simple task. Is there any outliner which can do that? Omni Outliner, which looks great, does not have that feature.

When you say “replicate” are you talking about cloning? That is, do you want to keep two copies of the same item in the outline, and when you update one, it updates the other?

If so, you should check out Tao:


Tao is a very powerful outliner, and can be a little intimidating. But it isn’t really all that different from OmniOutliner and it only costs $10 now. (There’s a replacement in the pipeline, that’s in beta now.)


Actual node cloning is difficult to find. They are considered to be somewhat of a power feature, mainly because they can be confusing as they contradict nearly every convention. Things either exist in one place, or a secondary thing points the way to it. Those that do “get” them though, definitely do like them, I’ve found. Tao (now Neo) is the only pure outliner I know on the Mac that does node cloning (replicants are, as far as I know, DEVONthink jargon, so if you are Googling for the feature, you’ll get further searching for cloning). Mori, the notepad/outliner application, also supports cloning, it calls them aliases, but this is a little misleading as they don’t act like aliases, they act like bi-way clones. I’m not sure I can recommend Mori though. Ever since it was sold, development has tapered off and support has been flaky at best. Neo is probably your best bet. It’s a fairly complex application with a weird interface (there are multiple ways to select something, and where you click in the row will do different things), more outliner than most need, but if you need some of those power features that nobody else touches, chances are it has it.


I find “replicants” (or clones) very useful. It is true that things, in the reality, exist only once, but this is the magic of computers: making ubiquity readily available!

Here is a case I’m facing these days: I’m writing two essays on different theses, but with some common reference materials. Without clones I would better put the reference material in a place separated by other materials used for the essays. Then, an alias would le me go to them when needed:

  • Common reference material
  • Essay 1
    – Materials for Essay 1
    – Alias to common reference material
  • Essay 2
    – Materials for Essay 2
    – Alias to common reference material

With clones, I can keep all reference materials (either specific or common) in a single place:

  • Essay 1
    – Materials for Essay 1
    – Common reference material
  • Essay 2
    – Materials for Essay 2
    – Common reference material

Just think to real life: it is true that you have only an encyclopedia (or a single set of each encyclopedia…), but each time you work on something different, you think at it as if it was a different collection. When writing of Bach, you think of it as a music reference; when writing of Baudelaire, it becomes a reference on poetry.

Clones let you separate the different faces of the encyclopedia. Better than duplicating: contrary to printed books, software reference materials can be ‘alive’ and grow as you use them. And any annotation will be readily available on both ‘faces’ of the same ‘book’.


I find it really puzzling that genuine cloning has pretty well disappeared as a feature in modern outliners - it was incredibly useful, but hasn’t seen a first rate implementation since "ThinkTank and “More” in the 80s.I was hoping to see it in the next OmniOutliner, but Omni seems to be pulling back from that plan and focusing on the Ipad.
They managed to have cloning in Mac 512s, but apparently it’s just too big a coding job for a computer with 2000 times the power.
I own “Tao”,but have never found it particularly stable and the interface is a mess.
The closest, best solution for me nowadays is Tinderbox, which has improved enormously in the last year and is now a Universal Binary. Its clones are just aliases, and don’t carry “children” of the original note as MORE etc did. But the program is very powerful and provides lots of workarounds for that issue.That said, it’s definitely the deep end of the pool and many users are confused by its myriad bells, whistles and farts. Anyone who uses it has to start slow and gradually adopt the more sophisticated features. Otherwise they’ll spend more time fiddling than writing. I should know.

I am sorry to say the original poster is faced with a choice between “intimidating” (TAO or NeO) and “very intimidating” (Tinderbox).

Faced with that choice, I’d plump for TAO or NeO. The bumps and crashes of TAO can be successfully worked through (I’ve recently put together an outline with more than 70 nodes), and NeO promises to be more stable. And if the OP is familiar with OmniOutliner, then TAO or Neo won’t seem too strange. Tinderbox with its note-centric approach starts in a slightly different place.

(Omni Group’s recent postponement of OO 4.0 with the likelihood of the cloning feature it had been talking about for four years, in favour of iPad development, was, I suppose, commercially sensible, but for me, irritating.)


OmniOutliner is a beautiful outliner, but it misses a lot more and more basic features than replicants/cloned items - for example the possibility to filter entries according to certain conditions, to name something rather simple (for example: show only entries with value “John Doe” in column “POV”).

Looking back, I have the impression that the art of making outliners is already on the decline again… :frowning:

Paolo, you don’t have to convince me of the usefulness of clones! My comment you quoted was meant strictly to speculate upon why the feature is so exceedingly rare.

Jowibou, the problem isn’t technical. It’s actually very easy to program for clones—especially if you have it in mind at the start and abstract display info from data. Since it isn’t really a technical problem, this is why I speculate that it is primarily a support problem. Consider a different feature entirely: Scrivener’s ability to represent containers of items as index cards. The idea is pretty simple once you know it: corkboard is just a fancy folder view, basically—like Icon View in finder, or better yet, like Album View in iTunes. But not everyone grasps that intuitively, and you see a lot of people on the forums—most of whom have not read the tutorial yet—confused because index cards are not acting the way they appear they should be acting. It’s difficult for a veteran Scrivener user to go back and realise difficulty in fathoming that the index card is the document, but a lot of people have a very hard time getting over that hump. It’s a feature which requires you to visualise data you cannot at the moment see.

Clones are even worse in that regard. You have to think about the interface in a way which most are not accustomed to. The majority of computer users, not being computer experts, have a very hard time with the realisation that when you are editing in one place and data in another place entirely is being manipulated. The computer does a very good job of hiding this for the most part, it erects powerful metaphors to make it look like what you are doing is simple.

Tossing in a feature which lets people open a node that makes twenty changes to the document at once is just crying out of lots of tech support, no matter how bold you make things in the manual or how many times you reiterate in a FAQ.

Tinderbox I wouldn’t recommend for this. It doesn’t support true clones anyway, but rather the alias model where pointers can be erected guiding the way to the master node. True clones can be deleted anywhere and you won’t lose data until you delete the last one standing. Delete the master node in Tinderbox and it along with its twenty aliases pack it in. Plus, in more recent versions, aliases have been acquiring a greater degree of attribute autonomy. Most of it is related to map appearance directives, but crucial separations are now available which cause aliases to report different than masters.

I wouldn’t hold your breath on OO. That program has been usurped by other priorities for many years now—about as long as TAO had been out of commission before NeO stepped up. OO has always been more of an outline for clicking than typing, if you know what I mean, too. It’s not that it doesn’t have its own sort of power, it does, but it’s not the type of power I need or want.

If you don’t mind going the cross-platform route, and installing some UNIX libraries, my favourite outliner on the planet is Leo. It’s got real clones (including children!) can do Scrivener-esque children compiling. You can put @others into the text body of a parent node, give it a file name, and it will dump a text file to the drive with the contents of that entire tree. Not only that, it will keep itself up to date with that file too. Edit the file in TextMate, open up Leo, and your outline body text will update. It also supports cross-wire inclusion for common headers. You can clone a common header set and include them manually into the document with << section >>. It’s a great little document outliner. It’s definitely an outliner, not a note-centric application, but the body text area is prominent and easy to get to. It’s not heading centric either, but rather pretty balanced. It was written for programmers so they could marry an outline with source code. The source code literally becomes outlined since you can keep an outline synced with the filesystem. It’s fantastic for that kind of thing.

Thanks for the usual thorough answer, AmberV. Of course, I was being facetious in calling it a coding problem, but I take your point. Perhaps it’s yet another rhetorical problem: an unwillingness or inability to explain a different paradigm to an audience that is stuck in its metaphors. I tend to discount visual metaphors in most software and actively try to think outside their narrow boxes. That said, unlike you I don’t have the programming knowledge to fill the gap with a coder’s mindset - so I just muddle along with an eye firmly focused on the functionality I need. Using (recyclable) printouts and a red pencil to edit for instance. That’s also why I barely use something like the index cards in Scrivener - they may LOOK like index cards, but to my mind they don’t behave like them in that they are inflexible to rearrange in different views - which strikes me as the whole point of cards in the first place. I much prefer the outline views, despite their limitations in comparison to something like MORE. (For me the single feature that sets Scrivener way above the fray is “Edit Scrivenings”, the rest is terrific but not essential.)
Thanks for the heads up on LEO: looks interesting, but UNIX seems weirdly - how shall I put it - insectlike to me. I know myself well enough that I won’t be investigating further until I have a LOT more time to spare.

Corkboards will be enormously more faithful to the philosophy of real-world corkboards in the next version. You’ll be able to do spatial non-outline-impacting arrangement, just like Tinderbox allows, and there will in fact be a clone-like approach. I can’t say much more about that, but for some applications, it will elegantly address the problem. I agree with what you say about the current iteration though. I usually find myself in Outliner mode for the same reasons. It packs more information into the display and essentially accomplishes the same task. But, for people that thrive on visual aesthetics, it can be a creative interface to work in. I think in 2.0, it will also become a vital tool that addresses certain things Outliner cannot.

You’d be doing well to wait for a pocket of spare time before tackling LEO. While not as complicated as Tinderbox, it definitely has a learning curve and you won’t get much out of it without going through the manual and playing with it for a while.

UNIX applications, or really to be more specific, the whole open source community, thrives on a different underpinning philosophy than the Mac. On the Mac, everything is in front of you. If you cannot see a feature in the application, then you cannot do it. Exploring a Mac application is thus a visual experience. You go through buttons and menus and once you’ve exhausted the repertoire of visual metaphors, chances are if you didn’t find it you’ll never be able to do it without a feature request. The only real exception to this is AppleScript, but support for that is spotty and requires the developer to implement it. Even then, AppleScript is often just a scripted amplification of the visual features.

UNIX takes a different tact in that quite often the interface (if there even is one at all, at least what a Mac user would call an interface) doesn’t even give you a tenth of what you can do. For people coming from that background, it’s not even an issue—it is what they are used to—but for someone coming from a Mac background it can often lead to moments of thinking that applications are woefully underpowered or lacking in the most rudimentary of features—just because there isn’t a button or menu for it. LEO is a good example, if you hadn’t of read my post or the manual, there would be absolutely no indication that you could write out ten different files off of ten different nodes, or that you could include the children of a node by typing @others or @all into the body text of the parent. There is no button that inserts that, no menu command that sets it up for you.

LEO has a completely open, Python architecture. The preferences file is a LEO outline which you can browse and edit. LEO itself is a LEO outline. You can browse the source code and tweak it at will, even inject code into the running thread. All of that is somewhat at odds with what the Mac is all about. I’m not saying one is better than the other—both philosophies have their merits, but you can definitely tell when those merits become stretched. Even geeky Mac applications follow this guideline, sometimes to a fault. TextMate is a good example. Everything you can do in TextMate has a menu, and it can a lot. There are a trillion menus and sometimes finding what you want to do can take minutes of time, whereas in the UNIX world with VIM, you just type the command in at the bottom of the edit session. One can be explored, the other cannot, one takes five minutes to find an obscure feature, the other has every single feature at a flat level of accessibility. It’s difficult to say which is “best”. It certainly can be grating to those deeply invested in one or the other though. I know when I first started using a Mac, I was in a constant state of cussing beneath my breath.

I digress. Edit Scrivenings: emphatically agree. There is nothing else like it out there, anywhere, and that in great part is a technical difficulty. Tinderbox’s Nakakoji view is the closest thing I’ve found to it, but it’s read-only and procedural as opposed to directed—a huge handicap. With Scrivener 2, some important abilities will finally be possible, and I think for power users looking for better brainstorming and outline handling, it’s going to be a treat.

But anyway, on the Mac and most other platforms, I agree with Andreas. The outliner market has been dying for a long time. There hasn’t been anything like MORE in decades, and the two closest contenders, OO and TAO, were neglected for many years, occasional bug fixes aside. TAO was close in its early iterations, but it got cluttered really fast. The developer understands what a good outliner should do, but with all respect, is not a GUI designer! OO is the opposite: great UI but beneath the surface it’s pretty shallow and does indeed miss a lot. Much of OO’s power has to be harnessed with scripting, which is a shame. For example, filtering is possible, you “just” have to program it to do it.

I should note that I am generalising a bit, of course. There are examples to the contrary on both sides. I guess Tinderbox is a good example of a Mac application that doesn’t put everything up in front of you—and I think that is a big part of why it has such a learning curve.

I am glad to see that others find clones useful too. After reading the first replies to my post, I did a quick search, and found that one of the developers of Omni Outliner considers clones an arcane feature.

As usual, the replies in this forum are very enlightening. I will give NeO a try, and then report back.

AmberV, Leo sounds intriguing. And, in another life, when I used to write code, I would have plunged right into it. But now I am afraid that it may require too much time to understand and setup. Thank you for your thorough, informative answer.

AmberV, I should be working, but just wanted to make sure you didn’t think your exhaustive post went unnoticed or under-appreciated. I simply can’t do this as fast as you can, but I do very much agree with everything you had to say. My impression of what 2.0 will be like is a bit vague, but I must say your hints are making me hungry. That’s as much time as I have to spare today for dreams of future tools, must get back to working with the current ones. Time to tell my brain to “hush” and go a bit deeper - that’s the best tool/tactic I’ll probably ever have.
Thanks again.

As always, Amber, your thoroughness and simplicity in explaining these things is very much appreciated by us less-educated “laity.” Sometimes I wish I weren’t traveling 40% of the time I would follow your advice, explore more of what you write, and learn much more. It all sounds fascinating.

I promised that I will come back to report on my experience using an outliner that allows to “replicate” or “clone” some of the nodes in such a way that they appear in more than one place and every time you change one instance such change is reflected in the other instances as well. Among other reasons, the main use for this when you are structuring a novel in which the telling is not necessarily linear.

First, I tried Neo, as it was suggested above. The application does handle “cloning.” But the interface is a bit clumsy. My fingers didn’t seem to adapt to the interface.

Some googling revealed that MORE was still alive an well, though it required a special environment to make it run under OS X. It was provided by the emulator SheepShaver which creates a perfect OS 9 under OS X. MORE, indeed, does cloning and works as you might expect. The slight problem is that it is difficult to make your outlines interact with other OS X applications.

So, after some soul searching, I went back to DEVONthink. As it turns out, you can simplify the interface until it looks and behaves like a regular outliner. DEVONthink handles clones gracefully, although, in a Blade-Runner fashion, it calls them “replicants.” DEVONthink has two added bonuses. The first one is that it plays well with other OS X applications. The second is that I can create one folder to hold all the research for that project. This solves a problem that I had for a long time.
I need to “see” certain things as I write about them. Scrivener has the capability of showing your research side by side with your document, but, alas, not in full screen, which is the mode I much prefer to work with. With DEVONthink, you can place your research on one side of the screen, go full screen in Scrivener and make the background transparent, and voilà, you have both applications side by side. This is a screen capture of DEVONthink as an outliner with replicants.

Of course, now that I have an iPad, I wanted to be able to see my outliner as I revise the drafts (I like technology, but when it comes to revising, I still prefer pen and printed pages). DEVONthink, unfortunately, is very limited in that respect. I exported the outline as an OPML file, opened it with OmniOutliner, and exported it as an HTML dynamic file which I uploaded to Dropbox. Now I have access to the outline through the iPad. When I change the outline in DEVONthink, it takes about a minute to update the new version in Dropbox. Seeing it with Safari, it looks like this:

Have you tried CarbonFin’s Outliner for the iPad? It’s a decent, but very minimal, outliner which through their free online sync service lets you upload OPML files to the device. Once there they can be edited and then re-downloaded later on the desktop. You’ll lose clones of course, but that might be a good two-way route for outlining ideas.