DEVONthink — Second Impression and some Tips
#Polymorphic Software, #TECH22
Technological advance is an inherently iterative process. One does not simply take sand from the beach and produce a Dataprobe. We use crude tools to fashion better tools, and then our better tools to fashion more precise tools, and so on. Each minor refinement is a step in the process, and all of the steps must be taken.
Chairman Sheng-ji Yang, “Looking God in the Eye”, in “Alpha Centauri”
I’m continuing to use DEVONthink (Pro Office version; see this posting for a very quick introduction of its strenghts). I am still transferring the information from my Wiki to DEVONthink, so this is a work in progress. While my main database is now at about 29 GB, it still works fine. Opening the database takes about a minute or so, but I only do this once every few days — the rest of the time it stays open — so that’s okay.
At the moment I think that it’s the best collection I have used so far — and it is very good to dealing with a lot of information/files you put into it.
According to the very helpful customer service (thank you ), the theoretical maximum number of documents is at about 200.000 and the maximum total number of words at about 300.000.000. So far, I’m at 50k of 200k files and at 1.2M of 3M words. So, there is still some room left. Coming with a very large collection (and it is a collection that I can import easily), I’m not concerned that I’m already at 25% of the amount of files, it would be akin to saying the apartment you just moved in was too small because with all your stuff from your old apartment there is only little space left. So, no sweat there.
So, how does my DEVONthink database look so far?
Note: I use the terms “groups” and “folders” interchangeably, because in DEVONthink any folder is actually a group, i.e., all documents in that folder are tagged with the group name and displayed in that group.
The first thing I do with new software is to check the Preferences (application name in menu bar, then Preferences; or press cmd + ,). The global settings are powerful and often the default settings are not ideal for me, so they are well-worth a look. The downside is that in some applications they can screw up work, so think before you click.
Personally, I have chosen in General (don’t know the original settings): Alternating row colors in views, Display number of items inside groups, Highlight Internet links in views, Mark duplicates and replicants in color (duplicates are blue, replicants are red — very useful!), and colorize icons with label.
For Import I have selected Filename without extension as I do not want the see .txt or .rtf each and every time. The icon works fine to tell me which file type I have.
For Backup I have selected daily — note: it’s only the backup of the database information, not the files itself!
For Update I have selected on startup.
The right view
DEVONthink offers different views (View – [first six entries]). Personally, I think the “as Three Panes” works best (for me) — you get the group structure left, the content of each group top right and the content of the selected file bottom right.
It also gives you two arrow buttons which allow you to go through the content you had displayed previously.
Very useful if you want to look up something (e.g., an image while writing something in an rtf file), have found it, and want to go back to the file, then simply click on the left arrow until your rtf file is displayed.
One References Database
I have put all my material that I still work with, PDF documents of articles, images, videos, etc. pp. into one database. Archived content of finished projects is in a Wiki. The reason for one database is that I think that different topics stimulate each other and if they are grouped (“put into folders”), there should be no conflicts. It also saves space, as some images, videos, documents, etc. are relevant for private projects and work projects. Disadvantage is that it gets large very quickly, but see above, it should work and for me (now) the advantages are greater than the disadvantages. If it gets too large it’s possible to split the database, although only some versions of DEVONthink support multiple databases.
Sort the material in DEVONthink
I strongly recommend sorting the material in DEVONthink. The reason is that DEVONthink is not a simple file/folder system, but offers some features that help you very much in sorting your data — most importantly recognizing duplicates (see below) and allowing files with the same name in one place (see below). I didn’t have problems importing a few thousand files in one go — simply drag them into the folder you want to have them. Even if it’s PDFs (and DEVONthink seems to analyze them to be able to suggest similar documents and to find duplicates) DEVONthink simply works — I noticed this when I accidentally dragged 8900 files, mostly PDFs, in one go into DEVONthink (cursed subfolders! ). In comparison to other programs like Papers it doesn’t even heat up the processor — it just imports, one file after the other. So drag your collection into it (make sure you have enough space on your hard-disk left, it copies them!) and go drink a cup of coffee.
Files with the same name in one place
One of the stumbling blocks of current operating systems is that they do not allow you to have files with the same name in one directory. For good reasons, after all, how would you differentiate between the files? But there are ways and one of them is looking at the file size. If you try to clean up your collection with Mac OS Finder you get crazy when each time you move documents with the same name into one directory. It warns your that you are going to overwrite it and if you decide not to overwrite it, the files that are possible duplicates remain in the directory where you wanted to move the files from. Then you have to manually check whether the file is identical, e.g., by looking at the file sizes or the date-time information. This is very inconvenient. In DEVONthink you simply have the files in one group and can quickly check whether they are duplicates — or rather, let DEVONthink do the checking for you. If the files are the same, they are marked in blue (depending on the preferences). It doesn’t get quicker and easier than this.
Putting Images from Aperture into DEVONthink
I had a few thousand images in Aperture that I needed in DEVONthink. Be very careful how you put them in DEVONthink! Do not simply select them in Aperture and drag them in DEVONthink. I did this and somehow Aperture did not export the masters but a downscaled version of the images. While some images were still usable, comic strips (a few thousand Dilbert’s, Non-Sequitur’s, PhD Comic’s, etc.) were unusable afterwards. You have to export the Masters (or the Versions in Best Quality if you have made changes in Aperture and want to keep them) into a Finder directory and then import that directory.
If you want a downscaled version of photos you have taken yourself, because you do not want to use the photo in a presentation or print but only have it as a reminder of an idea, you can use the eMail export option. It’s optimized for small file sizes that work well on the computer display (but nowhere else). Either use the Export Version and select the preset for eMails or select the images, click on send per Mail and then — when the eMail is created in Mail, select all images there and drag them into DEVONthink. It’s quick and dirty and it works — for these kind of images.
Note: If you have accidentally imported images in two versions, the original size and the smaller size, and you want to remove the smaller size, sort them according to name, then have a look at the file sizes. The smaller one is either always the first or the second one with the same name. You can additionally right click on the header line of the view (where name, modified, etc. is displayed) and select “width x height” to get that information in the list view. Hmm, although thinking about it, making a smart group with type image and Date Added with the information when you have added the images should work like a charm … damn, that idea came to late for me.
Get rid of the duplicates
DEVONthink has the ability to recognize duplicates — by comparing the file content, not the name! This means you can put three images with the same content but different names and time information in DEVONthink and it will point out that these images are the same. This makes sense as files often have the same name but different content, e.g., two word documents representing different stages of the writing. It doesn’t go so far to recognize scaled versions of the images, but recognizing duplicates is a huge boon. The best way to deal with duplicates in the early stages is to go to the smart list “duplicates” (should exist, if not, make a smart group with “Search in: Database” (or where ever you expect duplicates) with the terms: “All of the following are true: Kind is Any Document and Instance is Duplicate”. You can also limit this to specific file types.
Next, if you have not already put files into a structure, you can select all duplicates in the smart group and use the script menu of the menu bar (looks like a scroll, the one belonging to DEVONthink on the left side, not the system one on the right), go to data, choose “Move Duplicates to Trash”. This leaves only one instance in the database*. The reason why you should do this only if you have not yet put the files into a folder structure is that I am not sure whether DEVONthink puts the ones in the Inbox into trash. In case you want to do this manually, sort the files by size (by clicking on the size header) or sort them by name. If you do not see such a window, use View and select the “as Three Panes” View.
So far the only problem I had with the identification of duplicates was with a two TED Talk videos which are regarded as the same although they are different, even having different file sizes — but they share the same beginning. I think I simply edit out the trailer at the beginning in one file — it should work.
* This is an example of how DEVONthink makes it hard for you to use its full power — judging by the name I had expected that it places all duplicates, i.e., all the files, into trash. In fact is leaves one instance. It should have been named “Move all duplicates except one instance to Trash” or “Move all duplicates save one to Trash”. Ideally there would be a “Move all duplicates in Inbox to trash” version, which would leave all sorted files intact but remove duplicates only from the Inbox. Selecting only duplicates in the Inbox would not work here because it should remove the one file in the Inbox that is also already in the database.
Find a good structure that works for you
I think you need a clear mental model of how you want to use DEVONthink — until you have that, you cannot use it effectively. There was a request for training in DEVONthink’s forum and while a single case isn’t reliable data (actually, only a datum), it’s usually a bad sign. But let’s be clear here: A bad sign regarding the communication of how to use it, the software itself is top.
Currently I have some input folders (or groups) and then a structure that makes sense to me, among others:
- current: Holds documents or replicants of documents I work with at the moment. I need to edit them quickly and I do not want to open folders/subfolders to get to them. One of the strengths of DEVONthink are the replicants. If you right-click on a file anywhere in your database and select “Replicate to” you can create a replicant (think file alias) anywhere else in the database. After that it doesn’t matter which file you edit, the changes will be in “both” files (it’s the same file that is displayed in different places). Very useful. Be careful not to duplicate it, because this will create a copy of the file and your changes will not be in the other file. If you have selected the color highlighting in the preferences, replicants are red (easy to remember via the “re”), duplicates are blue (mirror the “d” and you get the “b” for blue). The current group is the only folder/group that I have assigned a red label to, so I can find it quickly. If I want to go the the group where the ‘other’ file is, I go on the information inspector (Tools => Show Info; the i in the Task Bar or simply shift+cmd+i), at the bottom of the window is the Instances information, click on it and select the other file and you jump to that group.
- Projects: Contains groups/folders of my projects, ranging from publications to teaching to career to advisory, etc.
- Handbook: Contains a collection of files that all deal with doing my work as a scientist. Information on publication, APA standards, writing quality, some notes I made on how to improve what I get out of conferences, etc. Files from Sources, Images, or Videos are either replicated there or available via smart groups. For example, I have tagged a few of the very best TED Talks I have seen with “presentation” and “best_practice” and a Smart Group shows me these videos in “Presentations” as a reminder of what a good presentation can do.
- Topic Notes: My notes on the topics I work on (duh! ). It includes a Glossary with smart groups that show the files with the explanations sorted in Grammar, Graph Theory, Statistics, Theories, etc. and normal groups for the main topics I work on (mobile media, reflection, etc.)
- Community: Contains files with Information about interesting Journals, People and Companies, Tools, Places, Conferences, etc. For example, the rtf files in Journals contain some information about the target audience of the journal, the topics, which issues I have read/scanned for interesting articles, etc. Each journal file is tagged with “journal” and a tag that indicates how relevant it is for me. Of the > 200 journals I have there, I can quickly check the journals with a relevance of “very high” by a click on the smart group.
- “Private Projects” and “Private Topic Notes”: Similar to the “Projects” and “Topic Notes”, only the information is private, i.e., definitely not work related. It essentially prevents (or I hope that it does) that I sit in a meeting and accidentally have photos I took at my last bondage shooting appearing on my screen. Like written in the beginning, I think that work and private interests stimulate each other (although not necessarily in the way of the current example), which is why I do not split them into different databases, but at times this can cause problems. Quite large ones if the notebook is connected to the projector in that moment …
- Sources: All my literature that is in PDFs. Given that I have transferred almost my whole library to PDFs via a very fast document scanner (and by gutting the books, yeah, I’ll end up in hell for that) and that nearly all my research literature is in PDFs, it is quite large (2981 files). So I have used the Data – New From Template – Registers – A-Z and created groups from A-Z and sorted them there. Each PDF also has a text or rtf file in the same group with some notes about the article (if I have read it). I’ll write more on the in a later posting, but it is crucial to have a clear structure here and it’s very, very easy for anything written: use the authorname_year style. If there is more than one author, use authorname1_authorname2_authorname3_[etc]_year up to seven author names (because this is the amount of names you have to write out in APA style, after that it’s first author and et al, so a paper with 8 authors would be firstauthor_et_al_year).
- WebCapture: Using DEVONthink’s Browser Plugins I can easy transfer a Website to PDF and have it appear in the Inbox. I put it into this group, tagged with the relevant information.
- Presentations and Lectures: Notes from, well, Presentations and Lectures, not sorted at the moment, but I’ll probably use subfolders/groups like year. Tagged of course.
- Images: Images are a difficult subject and I am still struggling how to organize the >40.000 images I have there. Yup, forty-thousand! In the best of all worlds, I would know the source of the image (e.g., the photographer or the artist/designer/painter) and would group them this way. But I didn’t get to that amount of images this way and the only way this worked is for comics, comic strips, and some artists I know very well. So I have tried to additionally sort them according to categories, e.g., Astronomy, ads, apps, design, Drawn Art, Jokes, and for most photos I used categories like Decisive Moment, Portraits, Street Photography, Landscapes, Cityscapes, Documentary, Animals, Plants, Underwater, etc. Problem is that they are not mutually exclusive, meaning that the same photo should be in different groups. Usually that’s what smart groups are for. So, at the moment I am using static (normal) group to make sense of the images I have and then create smart groups of the main categories (like Street Photography, Animals, etc.) and tag the photos — and this way I can give one image multiple tags to have it appear in more than one smart group. If I got time that is, 40.000 is daunting. Still, I think that keeping images in DEVONthink works better than keeping it in Aperture or iPhoto. Reason being is that I do not want to edit them or use them in a print project. I want to have images available for presentations, for stimulation (not only this kind of stimulation, thank you), inspiration, etc.
- Videos: Easier than images because I have fewer (about 100) videos at the moment, usually small clips. Tagged so that is it easily available where I need it.
Use the right view for sorting images
While the “as Three Panes” view is ideal for working with the collection, for sorting images I prefer to open the group into a new window (double click on the group) and select the “as Icons” view. You get a small display of the image and can easily select them and sort them into the groups they belong to. Use cmd to select multiple images in different locations or shift to select a group of images that are next to each other. Given that I am interested in portrait photos I had a large amount of these images and I used the following setup to quickly sort them:
Note: It was pointed out to me that the image here was NSFW, so I replaced it with a censored version. If you want to see the uncensored one, simply click on the image.
(Don’t expect much, I think the images are artistic, that’s why I saved them — as an inspiration, but tastes differ. However, I don’t want to get anyone in trouble with their IT department.)
In the middle is a open window (group) of the images I had to sort, on the left is an open window (group) of all portrait images, and on the right is an open window (group) of images that are not portrait photos. I selected the portrait photos in the group on the middle and dragged them left. From time to time I selected the images above the point I was and dragged them right. If you have only two categories (A and not-A), this works very well.
Use Smart Groups to have the information available where you need them
The killer feature of DEVONthink are the smart groups and the criteria they can be based upon. It’s like a smart list in iTunes or Finder, only it can contain anything depending on your criteria. You can base the contents of this group on: All, Content, Name, URL, Comment, Metadata, Author, From, Recipient, To, Title, Headline, Subject, Description, Keywords, Organization, Copyright, Album, Composer, Tag, Flag, Unread, Locking, Label, Width, Height, Duration, Size, Word Count, Kind, Instance, Item, Date Added, Date Created, Date Modified, Date Opened. You can use one or multiple criteria which either all have to be fulfilled or any of them. This allows you to get a list of all images with a width of less than 600 pixel who were added before date x as easily as getting all text files with the word “important”. Usually, I use smart groups based on tags, as I can easily specify files that I want to appear in a certain place.
For example, I use tags for the literature relevant to a course I am giving (tag: “sgl_kurs_2011″) so all the literature (pdfs) that is stored in the Sources Group in an A-Z index is available in the smart group SGL Literature (Searches Database with all of the following true: tag is sgl_kurs_2011). If I stumble upon an image or comic strip or video that is relevant for the course, I simply tag the file with “sgl_kurs_2011″ and it appears in the smart group as well. Very, very convenient. Think about it, you do not need to copy the file or replicate it, it is just there — easily available where you need it.
Personally I would not use tags for categories that apply to all elements of a group. For example, I would not tag my 4263 Dilbert Comic Strips with “organization”, although they all deal with organization. It would make the tag on its own worthless, because whenever I use the tag I would have 4263 images in that smart group — not so smart anymore. Instead I remember that all Dilbert Comics deal with organization and use specific tags (e.g., teamwork) on the few very good Dilbert Comics where this applies to. Additionally, if I am working on a project (let’s say a course) and I find a comic that might apply to the course content, I tag the comic with that tag, e.g., sgl_kurs_2011.
Conclusion (so far)
In short, at the moment I am highly satisfied by DEVONthink’s ability to handle my data, to collect my files and to support me in my creative work. But I still think that DEVONthink has two disadvantages:
- It is not intuitive — it is very, very powerful and flexible, but it doesn’t show off what it can do for you. Which is a very strange disadvantage. If they would put up a few fully fledged case studies of people using DEVONthink (you can find some videos on their site and on YouTube but they don’t really cut it) it would help others to get a better impression of what you can do with DEVONthink.
- It is offered in different versions with different functionality. Personally, I think this is the Microsoft ex(?)-Windows still-Office way — and it’s a very bad way. It complicates things. With very few exceptions all Mac programs I know of offer only one version. It doesn’t ask you to pick and choose beforehand. It just gives you a program to work with, after all, it’s not like the additional functions would cramp up the application or fill up the hard-disk. Even Adobe, which I would regard as an untypical company regarding its Mac software, offers different packages but these packages contain different combinations of stand-alone software. There is no InDesign, InDesign Pro, and InDesign Pro Office, there is only InDesign and that’s a core strength of the company and those working with the software. Other examples are iPhoto and Aperture, which are very different and address different audiences. I cannot understand why they do not offer only one version and price them differently for students and teachers (educational license) and offices (business license). They still would get a fair price (based on what the person can pay for it) but would reduce complexity in purchasing, support and how-to’s. Hmm, perhaps a historical legacy embedded in the company, I don’t know, but it gives the impression that they themselves don’t know what should be included.
But still, these are disadvantages that other companies would dream of having. Currently my view of DEVONthink is like the Enterprise in “Star Trek XI” diving up from the gas clouds of Titan. It doesn’t look like there is much, but there’s something powerful — and it blows you away (yeah, and I’m a nerd ).
Scene from “Star Trek XI”
PS: I normally don’t write disclaimers but no, I’m not sponsored in any way by DEVONthink. And no copyright infringement is intended by the cuts from Star Trek XI, I hope it falls under fair use, it’s just an association I wanted to convey.
Other Postings regarding DEVONthink
- A very quick introduction to DEVONthink Recommended for getting an accurate mental model of what DEVONthink can do.
- DEVONthink Another good overview posting.
- Literature Management with DEVONthink Not only important if you are a scientist or student.
- From DokuWiki to DEVONthink If you have previously used DokuWiki, changing to DEVONthink is very easy.