Large-Scale Info Managers and Scriv

Yojimbo for me.

The ‘Tag Collection’ feature allows me to automatically file away about 90% of the docs I want to keep.

One drawback: Yojimbo will only handle files that it can preview and this does not include Excel files. Not a deal-breaker, but a mild disappointment.

Got a question for you all.

As part of my job I teach theology. I have extensive notes on various aspects of theology. I recently came across a great numbering system for theology. It breaks things down into general themes ie God = 1000, Jesus Christ = 2000, Holy Spirit = 3000 and so on. It then breaks the major themes into smaller themes ie God-All knowing = 1020, God-Anger of = 1025.

As you can imagine I have a list of these themes that goes into the thousands. Is there any way of creating a seachable index locally on my machine? So that I could search for Anger and it would give me a list of everywhere that Anger is mentioned and the number associated with it? I did try a single text document, but then you have to cycle through the individual references. I also have Devonthink and wondered about that, but it seems I would need to create a single file for each theme for that to work. At the moment my listing looks like the following:

I also thought of creating a wiki, but am unsure if that would work as I have never created one.

All I really need is a searchable index, if the topic is there I can add the number reference to the document I am creating, if not, I can add it to the index and then to the document. It also means that if I am searching for a specific theme, I just need to find the reference number and can search all my documents.

Any help would be much appreciated.

At first blush, it seems DevonThink would be ideal (I’m thinking SmartGroups for a start), but I’m hesitant to respond in more detail as I am not sure I fully understand the system you are describing and how it might work in practice. Perhaps you could start a new thread as I am sure there are many ways your needs could be addressed.

You might try this experiment. Create a DTP database and import all of your coded items. Then match the codes with tags. You search for all the 1000 items and tag them (God), then all the 1310 items become (God) (as judge), and so on. I’m using parens ( ) to simulate the oval “pill” shape of a tag.

The tags appear in the Inspector panel. There may be ways to select items and apply a tag to all of them, which would speed up this kind of editing a lot. I haven’t used tags myself very much; Katherine (kewms) is a very sharp DTP user and may be able to correct/enhance my advice. Good luck!

Yes, that’s easy. As soon as you’ve defined a tag the first time, it appears in the tag collection (yellow symbol in DTPs “binder”/tree structure). If you select several items in the list, you can simply drag them on the tag you want to assign to them, and it will be assigned to all of them at once.

I followed nom’s advice and created a separate thread for the question I asked. You can find it here:

I’m really happy with my info management on my mac, well, that was until I bought an iPad. My current system gives me document names that tend to look like this:


R is for record. This has been fine, but on the ipad with it’s limited real estate and especially those dumb apps that only give you a grid view my list of documents tends to look like this:


Most of the bits after my categories is missing due to space. I now do not know what a document is without clicking on the title and this is unacceptable.

I now need to rethink my info management. Any ideas? I did wonder whether just to use a title and put everything else in a document header. This should still make searches work properly. Still, open to suggestions.

A partial solution: are all the dashes necessary?
i.e. Compare “2012-07-01-2026-R-Cat…” to “201207012026R-Categor…”

Not much difference, but it gains an extra 4 characters. A little harder to read on it’s own, but in a column it is still easy to see differences in dates and times.

If you want to gain an extras 2 characters, you could remove the millennia from the date (e.g. “1207012026R-Category-…”). I don’t know about you, but I don’t expect to still be writing in the year 2100. Even if I do become a centenarian author (and then some!), I can’t imagine using the same system I do now.

For my work, I use the four digit year, because sometimes the document relates to something happening or written in 1912 or 1812, etc., even though I wasn’t the one writing.

Just a thought from a very tired old codger…

Fair enough. Based on the content of earlier posts (and some of them are incredibly detailed, so I might be confused) my understanding was that svs was using the date & time the document was made rather than the content it contained. If the latter, then 4-digit years make sense.

Overall, I’m not sure my suggestions address the core problem, but they may help a little until a better alternative is discovered/derived.

You could reduce the century and the month to letters which would also leave out leading zeros. so 2012-07-04 would become T12G4, 1945-09-10 would become S45I1O — or you could use lower case to ease confusion s45i10. That is based on the letter position in the English alphabet of course, though you could do it other ways, and it would soon become natural.

Don’t know if that would work for you or be automatable.


I suppose the key question has to be “what do you search for when you are trying to find things?” Allied to this is the important question “what software do you use to do the searching?” I’m still working on this problem, but I suspect that I will end up using OpenMeta tags more than I have done (I certainly already have tags that are yyyy-mm-dd). Not everybody’s favourite way of doing things, I know, but is it possible to tag on the iPad? Just a thought.


Hi svs.

Have you tried GoodReader on the iPad?

The column with the names is half the size of the iPad screen and if the names get too long, the letters get smaller so you can still see the whole name.

Works fine for me.

Give it a try.


DTP never worked out for me, perhaps i didn’t put enough effort into it, but my point for these applications is that they should be effortless. I was a Yojimbo user for a long-time until i moved to Together, which has been doing great. Changing to Together was a decision taken after writing this long article. This article covers Evernote and Yojimbo, its pros and cons and it concludes by suggesting a new info manager (nicknamed Foxhole :smiley: ). Foxhole could be a great project, but i don’t know anyone willing to develop it. Its changes compared to major apps. Among the suggestions for Foxhole would be improved tagging, top readability and handle of web content and a whole different way to include files into it.

Actually, i’ve sent this article to Shawn Blanc and this is a quote from him:

So if there’s any developer out there willing to talk about it, i’d welcome that (:

The larger the data store, the less truly “effortless” any filing/retrieval system is going to be.

Compare, for instance, the Library of Congress to a small office bookshelf. For the office bookshelf, pretty much any organizational scheme will work: by author, by topic, by size and color. For the Library of Congress, you have a multilevel hierarchy of topics, with additional classifications (author, year, edition) on top of that. Developing and maintaining the system requires substantial technical expertise, as does efficient retrieval of individual volumes.

Very few “consumer”-oriented programs handle truly massive databases at all well.


I don’t think that sorting or categorizing is necessary at all. Your notes have to be ideosyncratic to be useful anyway, being full of interpretation and your ideas and the like. Why use a system that anybody else can make sense of, then? You’re not the Library of Congress after all.

So when it’s becoming important to make sense of stuff and be creative, which means: connect things and generate new notes, then you’d be fine with a system lacking any hierarchy at all. Links between notes and shared keywords/tags will create clusters over time, though. This way, your system will create topic clusters by itself and generate internal hierarchies you wouldn’t have imagined before. This is the foundation of card index-based information management from centuries past. In German, this is the foundation of a “Zettelkasten.”

Like David Allen said in [i]Getting Things Done[/]: bottom-up planning is more useful in your daily work. Top-down planning like deducing from principles and pondering hierarchies seems to be sound advice at first. In the end, though, going with the flow and focusing on the process will yield more manageable results. Hierarchies emerge, you needn’t impose them.


Exactly that’s the purpose of a Zettelkasten: to surprise you when you look for something specific and stumble upon unintended connections. A strict filing system wouldn’t be able to produce that behavior.

FWIW, I find surprises in the library all the time…

While it’s true that only I need to use my system, months or years might pass between visits to the same topic area. The clearer the structure is, the easier it will be to remember what I was thinking the last time.

Also, I think there are useful distinctions to be made between my own ideas and information I collect from other sources. If I have a collection of technical papers on various topics, the material will naturally divide on more or less hierarchical/topical lines. It would be silly to ignore the structure already inherent in the material.

My own ideas, on the other hand, are pretty random. At one extreme, I have notes from reading and interviews, which tend to fall in the same topical categories as technical papers. At the other, I have essentially random observations, which may or may not be relevant to any current or future projects. It’s tough to define a system that can accommodate both.

Can you point us to any further information (in English, please :blush: ) about Zettelkasten? It sounds like an interesting approach, and one I haven’t encountered before. (At least not under that name.)


New to me too - but I found this earlier today: … asten.html