Replicants: the ability to file a document in several places

DEVONthink has a feature that Scrivener should consider: “Replicants”. Essentially, it is the ability to file a document in more than one place, and yet it is still just ONE document. Why would Scrivener benefit from this?

  • Let’s say I have written a collection of short pieces. Having the ability to replicate a document into several different folders in the Binder would allow me to look at that document in several different collections, in different orders. Or, put together new collections, without losing the old, for different purposes.
  • Perhaps, a novelist (which I am not) would want to experiment with scenes in different orders. He could replicate several possible structures of his documents, and continue to write. Examining each possible order from time to time with the Edit Scrivenings feature.
  • Scrivener is pretty good for writing poetry. Being able to replicate the poems into different possible orders would be a great way of examining different possible collections.

For those unfamiliar with replicants, you might say, “Why not just duplicate the docs?” The problem with that is that they then become their own individual docs, and when you edit one, the others don’t change. Replicants allow you to view and edit the same doc within different file structures. Combined with the “Edit Scrivenings” feature I can see powerful and abundant possibilities with this for writers examing their work in several different orders, on the fly, as they work on a project.

So, aliases/symbolic links by another name?

This has come up before. I like the idea, myself, but I think Keith was never keen on it – at least for round one. To answer your question, popcornflix, it is actually a bit different than aliases. Mori has this capability as well. A replicant or clone is different in that both instances of the document can be accessed as though they were the only instance. With aliases, there is a clear “parent” and “child” instance. The “children” point to the parent. If you delete the parent the children become broken, as in the Finder. With clones, there is no child/parent distinction. Each instance of the document can be accessed and edited from the position that you access it in, where as with an alias you would actually just be editing the parent. If you delete a clone, the other remaining clone(s) stay intact. You needn’t worry about which is parent and which is child. It is a great deal more flexible, but I imagine it is technically more difficult to create, as only a small handful of applications support it.

Incidentally, in the past, the workaround has been suggested as using reference links, or even dropping the file you wish to access into the text area to create an internal link. So if you have some notes you wish to appear in multiple places, you would simply link to them from the folder you would otherwise place it beneath as a clone.

I have to admit, I’ve wished for replicants in Scr. myself. But not enough to mention it. I guess, knowing where Keith is in the development process, it has to be a really BIG wish for me to post it. That said, I did want to lend support to this idea. It would be quite handy to be able to replicate files and apply them to different locations and in different ways.

I’m with you. cloning would be a great add.

One of the truly revolutionary features of MORE was its cloning ability. No other outliner to my knowledge been able to develop anything nearly as elegant or intuitive. The folks at Omni have been toying with the idea for years, but it’s never come to fruition. TAO has clones, but aside from being aesthetically challenged its interface is all but impenetrable.

One vote for clones / replicants here.


PS BTW, the development of Tao may also serve as a caveat. It started out as a great little competitor to OO, but became increasing complex and bloated. Seemed like every five minutes, Hamada was adding new features, switching menus around, renaming functions. So much so that it just became too much of a chore for me to keep up with it.

Another outliner that does good, true cloning is Leo, which is a really swell outliner written in Python. Hog Bay Notebook felt a lot like a slimmed down version of Leo to me. It had document assembly, clones, and so on. But, it is one of those applications that sits rather uncomfortably in OS X. It only runs in Python/Tk, which is even more clunky than Java. Before I switched to the Mac though, that was the outliner! One amusing aspect of Leo is that it has a bit of a chicken and egg problem. It’s document assembly feature is so powerful, they actually use Leo to write Leo.

I’m with you on the topic of TAO. I am a registered user of that application, from back when it has just turned 1.0. It had its rough edges, but was a nice alternative to OOP at the time. I think, think that 1.1 is even better? But I’m not sure, because like you I got tired of having to re-learn the entire application every time a new beta came out. What TAO really needs is what Keith did. Take the code base, sit down and write a detailed design document and re-code the entire interface from the ground up. It is supremely frustrating!

Since I’m a long time Devonthinker, Replicants/clones would surely be the number 1 item on my wish list if Scrivener was fated for extensive development in the near or medium term. I imagine they are a major undertaking and well beyond the present - and most likely future - concept of Scrivener’s purpose and design. Consequently I haven’t even bothered to make a request, much less bug Keith about them.

Instead I’m always trying to find a kludge/workaround that will satisfactorily replace clones.

The references lists is OK but a little hard to work with conveniently given its isolation in its own pane and the fact that the active document changes so easily as you try to select other docs to drag in. One improvement would be the ability to open all the internal references as a card set (or ES session) in the other split. That might be much like having clones. But the fact that you can’t search references means that you can’t find all items that have a certain internal reference - another disadvantage.

Using scrivener links in the edit field is one option, but my thick head always balks at putting too much non-draft material in that field. Given the isolation of the notes field (and its lack of Scrivener Links :wink: ), the lack of space, links and easy returns in the Synopsis field, and the aforesaid references issues, I suspect I’m just going to have to accept that the main text field is where most of my non draft material - notes, links, todos etc is going to happen.

If I can just get my stubborn head around it.



I agree 100 percent.

Hey, sounds like our experiences are clones of each other! My other frustration with TAO is the look of the interface. Using it is just not aesthetically pleasing. Scriv, on the other hand, works well and looks good. To me that’s a big deal. If I’m going to sit in front of an app all day, I want to see harmonious colors and forms, not the jumble of confetti that TAO presents (nor, by the way the, juji fruit look of Avenir). Keith was wise to spend time focusing on Scriv’s look. I wonder how many of us were sold, at least partially, on the aesthetics of Scriv?



Again, as a Devonthinker, I felt quite at home in Scriv from the start and the look was a big part of that. Cieth has certainly put his own quite pleasing stamp on it, and made it nicely adjustable (I’ve even ridiculously gotten to the point where I’ve colored my scriv elements with the designers (beautiful!) palette for the next production - kinda keeps me in the visual feel of the play :unamused: ) But overall what I like about the look is how cocoa, how MAC it feels.


Hi, thanks for your suggestion. Just as a matter of fact (from an implementation point of view), I am afraid that replicants (much as I love the term, reminding me as it does of Blade Runner) are unlikely to appear in Scrivener in any 1.x release.
All the best,

LOL!!! I had the exact same thought, since BR is among my top three sci-fi movies. But that’s fodder for another thread…

I assumed as much re implementing this in Scr. At least you qualified that with ‘any 1.x release’! So there is hope. :slight_smile:

Thanks for the response Keith. I have no idea how difficult it would be to implement. I think you can see that a lot of us would find it useful, so maybe further down the road…


Maria, Very interesting. I’ll consider that. One difference here is that we are talking about the ability to view parts of a written text in different orders on the fly with the Edit Scrivenings feature, it’s not just a file management thing.

My DEVONthink database is also littered with replicants, and am not always sure that’s a good thing either, but in general I like them. DT just needs to improve some aspects of it – if they ever get around to making DT v2 then maybe we’ll see it improved.


the EditScrivenings point is interesting, thank you for stressing that.

I wonder whether we should reconsider Scrivener’s purpose and the meaning of the draft. Scrivener is not for storing information at one or more appropriate places, but for writing text and keeping some essential material close at hand in some additional folders.

The heart of Scrivener is the draft, which is to produce text. Nobody would produce the same text several times again (only some nasty technical manuals, which are not worth reading). EditScrivenings is meant for the draft, so a writer (poet or scientist or what) would not need the replicant feature here.

It is just that one might want to try to read a text in different sequences that this becomes interesting. Then of course, tags are less interesting than replicants.


As for DT, I agree with you: We have to look for 2.0. In my eyes it was a wrong decision to start with thhe Office version before publishing the new file system which should solve so many basic problems. I am sad about that because I like DT and the company. But of course, they should know better, and perhaps they can sell a lot more Office versions with the old file system than Personal and Pro versions with the new one…

All the best,

Here is a scenario that I would find useful, if clones were available: I could in the Draft have the narrative of the story, but in another folder outside of the draft, have the individual parts of the plot arranged chronologically, and separately. Since the two branches would be linked together, I could either edit in the way the book will read, or as a linear single plot-line for easier chronology checking, flow, and consistency.

You can do this now dynamically by assigning a plot-line keyword and saved searches. But I do think it would be nicer if both were static. Then you could load two different sets into splits, for instance. Now you can only work with one set at a time.

Amber, Your scenario hits my point exactly, this would open up a wealth of interesting ways of writing, which is what Scrivener is all about.

Maria, I’m totally with you on the DT v2 thing. To me, the file system thing is so fundamental that if I were the developer I wouldn’t sleep well at night with it as it is. Not that it’s buggy, it’s just not as good as it could/should be. The Office version looks nice, but I haven’t even tried it. If I could afford a ScanSnap than I think I’d be all over it. I saw DT demoed with the ScanSnap at MacWorld two years ago and it was really very nice, the ScanSnap is a great little machine. Though I’m bummed about the dely of v2 I can see the market potential with Office though, because there is some sort of alliance I think with Fujitsu. And strengthening their market will be good in the long haul for all DT users.

I used to work with Avenir (which I still like, by the way, and which is an excellent app… Scriv just suits my needs better). I bought Todd Ransom’s license about two years ago and have been updating it since. As a result, I am on his e-subscription list.

Todd recently wrote about Smart Views in Avenir, which works exactly the same as Smart Folders in Mail and Smart Playlists in iTunes. I assume that means it must be an OSX implementation. I don’t think Scriv has an equivalent (though I could be wrong, there are probably tons of things I still haven’t explored in Scrvi). In any case, I wonder if the smart folder concept might not address some, if not all, of the needs underlying this discussion of clones and replicants.

Observations any one?


It wouldn’t for me, they are two different beasts.

One problem is where the folder gets its smarts from. In a sense, current saved searches function are not entirely unlike smart folders. But for a folder to be really smart it would have to allow for better, more complex search criteria than the current search engine allows. Since those sorts of searches/smart folders are fairly common on the mac (Itunes, Mail, Journler etc) I’ve also alwys assumed they were a a fairly easy-to-program thing. But this was discussed a while back and it turns out I’m wrong again, apparently. :slight_smile: