After my rather disappointing experiences with askSam I am still reluctant to trust any information manager again, but then again, I come across a lot of praise for the DevonThink products here and there and have started to think that maybe I should at least give it a closer look. I have indeed a very large collection of web downloads, texts, images and other things, and if there’s a better way to tame it than by Spotlight plus Finder alone, maybe I should know about from my own experience… :neutral_face:

So, maybe some of the seasoned DevonThinkers here might want to enlighten me a bit?

What I want to know:

  • The differences between the different versions - Personal, Pro, Pro Office… and what is DevonNote? And what is DevonAgent?? I have studied their website, but I still have only vague ideas what might be the one I need. If I need one.
  • That built-in AI, that seems to do the categorizing almost by itself, just like a reliable old archiver: Does it work with texts in other languages than English as well? Half as well? Not at all? (I have to handle a lot in German.)
  • Are there limits? In number of documents, size of documents, types of documents? (From their website you get the impression that you might put in the whole internet without problems, but then again I’ve read that once you try to handle texts of the length of a book (not uncommon for writers), it might become very slow. So, what are the real-life limitations?)
  • How many years do you use DevonThink already? And do you expect to still use it in ten years from now? What would you do should DevonThink Technologies cease to exist and Apple publishes a new operationg system Mac OS XII under which the old software won’t run? (Or if Apple should stop manufacturing computers at all because they turn into a phone company completely?)

Any comments welcome!

Hi Andreas

Am pushed for time, but will try to deal with your points briefly one by one (based, incidentally, on about twelve months’ use of DT Pro):

AFAIK, DevonNote stores notes you’ve written. The other DevonThinks - Personal, Pro and Pro Office - in addition store, index and search (or just index and search) most document formats you can throw at them (though not MS Word or Excel at the moment?), with the usual escalating degrees of “un-disablement”. In addition, Pro Office will take document scans in from a scanner, OCR them automatically and store them.

DevonAgent is very cool IMO. It is essentially a website crawler. Give it your criteria and it will use various engines to search websites and any links they contain, pursue the links as far was you want, rank the results in terms of semantic relevance and then download the pages you choose from its list. It has found material for me when I despaired of other means.

I’m certain the AI will handle German, though I have no direct experience of it doing so. After all, all the AI is doing is comparing arrangements of letters. It cannot “know” they are one language or another. And the company behind the software is German.

I’m the wrong person to provide evidence of limits. However, as you say, some posters on the website (e.g. Bill de Ville, who works for the company) talk of storing millions of words and gigabytes of material. I’ve not heard of the slowdown problem (though I have heard of it with rivals, such as Eaglefiler). The particular reputation of DevonThink, which first attracted me to it, is that its storage is rock-solid in its reliability; that is is my, albeit limited, experience.

The software is built around a proprietary database format, so this may be a justifiable anxiety. A number of critics have pointed to the possible problems of getting documents out, if required. Personally for this and other reasons, I don’t store much in DevonThink itself. I mainly use it it to index folders in the Finder. There are issues with this; for example, you shouldn’t move files around. But I do gain the benefits of the AI and DT search. Also, in the new version 2, promised for this year, exporting (or just indexing external folders, as I do) is said to be much more straightforward.

Overall, DevonThink isn’t AskSam (part of the Windows past I moved to the Mac to forget!). The expense is similar, IIRC. I know there are people here who are longer-standing DT fans than I. But some may suggest competitors that are less expensive, whose UI is less old-fashioned, which use tagging explicitly (as some favour) and which don’t store documents in a proprietary database (such as Eaglefiler and Together). But IMO DevonThink is still more sophisticated, reliable and potentially useful than any of them, especially with large and growing volumes of data and when coupled with DevonAgent.

The release of the long-awaited Version 2 may well be a watershed. When DT was first developed, (I think) it was very clever for its time. Some of us are waiting now in the belief that the people who came up with Version 1 can’t fail to produce something interesting for version 2.

But, given the responsiveness of the developers and the fact that the company is based in Germany, why not email them directly with your questions?

My (quite rapidly devlauing) 2p.


I own licenses for many applications that may be used in this note taking/information management business. Each have their strengths ad weaknesses. I’ll give you a rough overview what I like using them for, perhaps that gives you some information what it is you want.

Yojimbo - came bundled with Tinderbox. I did not think I would need it as such because it has only a very limited structural hierarchy (no subfolders etc). Then I found out that it synchs nicely with .Mac. For the moment I am keeping serial numbers and stuff that I regularly need on several computers in Yojimbo
Good luck if you put a lot of stuff in there and want to find it again later. It forces you to follow the tagging concept but if you don’t you will be lost.

Tinderbox - a strange beast. Very powerful but not easily classified. It handles information that is representable as text most gracefully (multimedia support is scarce and useable as an annotation). It does offer you more options than you will be able to explore in terms of what you can do with this information. It pays off in the long run IMO but prepare for a loooong run :slight_smile:

Omnioutliner - an outliner that I also use to take notes if they refer to a single event (e.g.a meeting). Anything more complex or if I want to keep track of recurring events and I will resort to other programs.

Curio - has a map view similar to Tinderbox but is less text-centric. I use it to quickly put things together that benefit from a visual understanding.

Devon Agent - it is some kind of a customizable meta search engine for the net running on your computer. Whenever Google turns up a lot of false positives that you have to manually inspect to pick up the hits that really interest you, it is worth instructing DevonAgent to do a search that will likely match your intention a lot better. The “NEXT” operator alone is highly useful. DA is used to GET information, not keep or maintain it. It can also inspect websites and download all things of a certain kind in a batch operation (pdfs, movies etc)

Devonthink Pro (not Office) - this is where I put all the bits and pieces that are of relevance even the ones that are not related to any particular project at the moment. To get some criticism out of the way: Yes, the interface is a bit daunting and not the visually most attractive but it gets the job done. It is scriptable and is perhaps the only application on my Mac that I have not been able to slow down to a grinding halt by putting a lot of information in there. And I mean a lot. DTP lets you gather information in al kinds of different formats, from webpages including partial webpages while retaining the information where this piece of information came from. I make use of this feature daily.
There are a couple of misconceptions, the most common being that you have to import everything into the DTP database thus losing the ability to work on those documents in the finder. You can import local documents from your computer in which case they will reside inside a UNIX bundle (still accessible by “reveal content…” in the Finder BTW). This will make the content of what you keep in DTP self sufficient and easily transferrable to other computers. I use indexing a lot more often. This keeps an index of everything I ask of it and allows me to perform very configurable searches just like as if I had kept all the documents inside the database. I keep Gigabytes of research articles in several folders on my harddrive. The most amazing thing, however, is the “see also” feature. Its artificial intelligence will search related items in your database by analysing its content and comparing it with the rest of your database WITHOUT asking you to categorize, keyword and tag everything manually first. It works well but is most impressive if you can feed it with content that you have collected yourself.
From time to time I update the index to add newly found material to my index. I may add that I work on two Macs and have yet to run into a problem keeping them in sync even when using indexing (i.e. keeping the bulk of the material outside the database) as opposed to import everything.

If I had to name the knowledge management application I rely on the most for my work it would be Devonthink with Tinderbox becoming more important as I learn to handle it better. Devonthink is where I put everything and look for it and Tinderbox is where I put stuff for a larger project I want to tackle. They both have the advantage of not relying on obscure formats that may become obsolete in a very short time. Devonthink does not alter the original documents and Tinderbox is plain XML.
Long post, hopefully of some use to you.


Oh, one thing I forgot: There is no problem indexing german documents at all. The only thing to keep in mind is that the power of the "see also " feature drops somewhat if the most characteristic words that are used and weighted internally to classify the document are very different in all the languages you mix in your database. As long as there is enough “overlap” (i.e. some words identical or similar) it still works well.

After years of trying out different info managers, DTPro, Yojimbo, Together, Eagleflier, you name it, I have returned to the most basic and simple. I simply create folders in the Finder and organize my info accordingly. I use the Finder window and have my folders in the sidebar for easy access. I have pdfs, media files, pictures, videos, rtfs, most of which is fodder for potential projects, so it all needs to be quickly accessible and able to add to a project database. Since DTPro and others use their own proprietary db, it made it hard to attach things to Scrivener project files. With this system, I know exactly where things are, can use Quick Look to do just that, Spotlight for searching, I can use pretty icons for my folders (can do this in DT and others as well). And most importantly, I can swiftly and easily use Scrivener’s excellent reference features to attach research items to my project dbs without having to load it all in every time I need it. This works for me, since many files are shared among different projects.

The only time I find myself wanting a quick and easy way to store something is when doing Web research, which is quite often. I tried Together for a while. It stores things in Finder folders so they can be shared. Though not consistently and in a way that suits me. But it does have quick ways to get info into the program. Others do this too. But then I’m stuck with its way of organizing things.

No, for me, I found all these programs to be more cumbersome than my now simple organizing system and a distraction to actually getting work done. I’m sure others would disagree heartily, but I can’t stand having to use a gazillion different programs to do things. While I still have to use different programs for things, I want to keep it simple and straightforward and transparent. After years of trial and error, this is the simplest and least obtrusive method for me.

Hi Hugh,

thank you very much for this hint - I haven’t taken much notice of DevonAgent, but this indeed sounded so good, I downloaded the demo immediately. And already the first steps blow me away. At least this one I gotta have!

They disguise themselves very well. I thought it was a US company.

I conclude DT Pro Office would be one step too far, because if there’s anything I do not want in this world, it’s a paperless office. I will testdrive DT Pro, then.

I happily used DevonNote for a year or so, then switched all my project-based info organizing to Scrivener. For storage and cataloging of info that can be used in multiple (or undetermined future) projects, I, like Alexandria, now just use the Finder for everything – web archives, RTFs (usually read in Bean), pdfs. The appearance of Spotlight and EasyFind (and now QuickLook) made using the Finder effective, simpler (one less app to buy, learn, and fuss over) and future proof.
Maybe it’s just that Alexandria and I are frugal Portlanders – the city that recycled an old armory into a theatre, any number of old warehouses into lofts and corporate HQs, soon a post office into a college – and we’d rather re-use existing assets than buy new ones.

thank you for your enormous post! I agree with you a lot - I use Tinderbox, but I consider it as a tool for computer-aided thinking, not as an storage for random information. Yojimbo’s on my hard drive, but I have no use for it. And OmniOutliner is for outlining singular things (from preparation of parties to jotting ideas for an article).

Its interface can surely not be worse than askSam’s clumsy appearance. Although askSam was unbelievable powerful at its strong points, I would not return to it.

This I believed indeed. I welcome the idea that this has not to be so.

Gigabytes? No problem. I have what I call my “InfoBase” - a huge tree of folders in folders, a vast archive system with a lot of keywords, in which I feed whatever information I come across that appears worth keeping to me. Problem of course: A lot of things fit in a lot of categories - and sometimes they are not where I am looking for them. Solution up to now: Spotlight and it’s clever brother SpotInside (a Spotlight plus the ability to look within the files that meet the criteria).

Sound great, anyway! Thank you very much again.

I understand. That’s exactly the way I am handling it in the moment. A folder system full of RTF, TXT and HTML-Files (plus GIF, JPG etc.) is rock-solid; you can even move from one OS to another without problem (I took with me the yield of at least ten years surfing, downloading and collecting, when I went from Windows to Mac OS X).

Simplicity, yes. This is where I am standing, and I will stay here unless there are very good reasons to go elsewhere. However, I had the feeling I should have a closer look on DT, keeping in mind my experience with some big data collections (maintained as askSam-files) I almost lost by switching. (I finally found a way to export at least the text, but all images, internal links etc. got lost, which is a painful lesson as for as durability of databases is concerned.)

That’s a likeable trait, I must say. :smiley:

Andreas - glad to be of help.

brett - it’s about data mass, isn’t it? Somewhere, as one collects documents, there’s a tipping point where a tool like DT becomes all-round better as an information manager than Scrivener plus Spotlight and Quicklook plus the Finder - and as a search tool and as a “see also” suggester of hitherto-unrecognised relationships.

In my current project, I thought that tipping point would come at about the 1,000-document level. Hence the investment in DT Pro. I’ve no idea whether that assumption was right.

Maybe if I lived in Portland I’d think differently - but all those ornamental cabbages… :wink:


Agreed – if I maintained a large database of info for continuing research (if I were still in academia, for example, and publishing lots of journal articles in a certain area), I’d certainly still be using Devon. But my current work is pretty much project to project, and the data mass for any one of them (including my book, which has hundreds of note cards and years of research accumulated) is small enough that the Finder and Scrivener are plenty. For the most part, Devon is an excellent and intuitive app, with a strong user community and responsive developers.

It is ‘Skunk’ yre referring to, isnt it? :open_mouth:

We Portlanders do think differently. It’s what makes us so special. :wink: (okay, so I know I’m inviting more flaming responses about Portlanders, especially from vic-k!)

Seriously, I have thousands of research documents and files (having spent almost 20 years as an academic!). My DTPro file was hundreds of megabytes large. And DTPro handled that load flawlessly. But it is actually for this reason I moved away from a proprietary format. There were too many files dumped into DT and much of it got lost. I was able to find things according to DT’s great find features, but with Spotlight, really, it’s working almost as well. Ultimately, I have found the system I described to be much more useful for managing this much lot of data, especially since the type of data is so varied. I think Brett implied it was Scrivener that really changed things for me and the need to share files among multiple project files.

Hey, different strokes and all. I would never claim my system to be ‘the’ system!

I’ve found that Leopard’s Spotlight (much faster than previous versions) and a good folder hierarchy work best for me. That said, a little while ago I had to do a search for a set of phrases through literally thousands of PDFs. Spotlight just didn’t cut it. What did cut it was FoxTrot Professional Search. I was extremely impressed with its interface and feature set. I’d strongly suggest you try it before settling on an information manager that adds yet another layer of abstraction on what are basically files. :smiling_imp: :slight_smile:


Devonthink features “Replicants”, i.e. you can have the same entry appear in as many instances and contexts as you want. Although they look like copies they are not, i.e. if you change one the other instances stay in sync (unless you expressedly want to leave an unchanged copy which is also possible.

For example, you run across a website with some information about Piranesi which you haven’t seen before. Now, where to store the information? In the branch of your info tree concerning itself with “architecture”? Or “drawing”? Or “18th century”? Or “nightmare”?

  1. You could make one copy of the website and deposit it in any of these categories, hoping that you will still think of it when you need it in another context sometime in the future.
  2. If you were diligent enough you could deposit copies of this note elsewhere in the tree.
    Problem: It is rather tedious plus the copies reside independently of one another from now on. But what if some scientist discovers tomorrow that one of the drawings on this webpage was actually created by somebody else? In this case you will have to hunt down all the copies individually and change them, one by one.

What I do is this: I highlight the part of the webpage that interests me (excluding ads, if present, for example) and press a hotkey that will create a note inside DTP including working hyperlinks and a metadata field containing the URL of the webpage for future reference.
Inside DTP, I highlight the relevant items in the folder hierarchy I have created over time (“architecture”, “drawings”, “Piranesi” etc) and press the button “classify”. Done.

What this does is deposit the SAME note inside these three folders at the same time. No matter which of the three folders I loook into in the future, the note including the pictures will be there. Spotlight or other search utilities are nice but they still force me to follow links somewhere else or having to remember or search for them elsewhere. Not to mention choosing the correct search term…

When something needs to be corrected, I do it once knowing that all the other places containing this very note will reflect this change, too. The added convenience is that these multiple replicants do not occupy more space than a single note inside my library.

Actually, I find this much more convenient and simple than other ways I have explored before but hey, maybe that’s just me. Try it out, they have a demo.


Nope. Portland may be noted for its whacky baccy - I wouldn’t know about that - but I’m led to believe it’s particularly well-known for its ornamental brassicas. In window-boxes… :unamused:

Maybe it’s the rain… :wink:

Thanks to all who anwered (isn’t this the best forum of all?) - just to keep you updated: I downloaded DT Pro now and run an index on my infobase, without any import. (The possibility to do this takes at least away my fears about stuffing a lot of things into a database that won’t let me take it out again…)

Impressive! It took - well, about 20 minutes to index 6,5 GB (around 64.000 documents), and really, this function they call “artificial intelligence” finds connections I wasn’t aware of but that make sense.

Right now I found out that you can create a list of words, weighted, counted, sorted in all directions, and by doubleclicking on a word start a search for other documents containing the same word… Now, if that’s not cool?

What I still have to test is what it will not find. Which connections I know that exist go undetected. Exactly because it’s so impressive, I feel a strong persuasion to rely completely on its results.

Anyway - at least in the moment, it’s more fun to surf my own infobase than to surf the internet. :mrgreen:


One discovery I made about the trial version. If I remember correctly, Devon operate a trial period based on time launched, rather than calendar time - credit to them for this.

I notice you’re indexing rather than importing. But… if you do import anything, every time you download or tuck a document into the application via “Services”, it launches automatically (of course)… and will stay launched, sitting quietly in the dock and eating up your trial period, to no purpose.

Worth remembering.


FoxTrot looks like exactly what I need regarding searches. It is the only thing I miss is really potent search capabilities. The Pro program is WAY too expensive for me, however. I can’t afford $99 Euros, which is probably close to $140 for us U.S. types. But I’m still going to check it out. And the personal version, which is much more reasonable. A good search engine would complete the picture for me.