Interesting. But if the OP is going to be writing a thesis, they are going to need to cite quite a bit. I would therefore suggest that at this point they need to see the reference manager as the hub and starting point of everything they are doing. You do not want to be accused of plagiarism, nor do you want to find yourself hunting around for the source of something when you have a deadline to meet.
If I were the OP, I would put everything in Bookends, allow Bookends to manage and rename all the PDFs etc, and index the attachments folder with DEVONthink. I would tag the PDFs as a way of categorising and organising the material. (I suppose I should say that I use this method myself, and it has usually worked pretty well.) I find that tagging files is more flexible than putting items into static folders or groups, and it goes well with smart folders/groups. And of course tags are visible across the system and appear in the Finder as well. For some time now I have used a system of tagging devised by Macdrifter which consists of merely doubling the initial letter of any word you use as a tag: so instead of Freud the tag is ffreud. This makes it very easy to search for tags. Back in the days when I was doing my PhD I might have had a PDF that was tagged with cchapter_1, ffreud, ddefence_mechanisms and so forth.
But as I said before, devising your own system is advisable, in my view.
My point is I find it easier to locate a pertinent reference with Devonthink’s search feature than using the search capability of most reference managers.
And I think it is fine during the writing/editing phase of almost anything in medicine to use a web URL link or a Devonthink Item Link as the reference.
At least for me, this makes my writing process much more efficient to find the relevant publication, and it assures that I do not plagiarize. If something is going to publication or some other very formal work product, I simply need to convert the URLs and/or X-item links to the exact MLA format or whatever format is needed. Sometimes a reference manager is not even needed at all in medicine since most medically related citations are on PubMed and PubMed has pretty good support for creating a bibliography in the necessary format.
Most notably, we have just given an excellent example of how a workflow that works for one person may not work for another. Just think of each person’s use case as an example and then mix/match ideas from those examples to come up with your own workflow.
And I suppose my point is that if you have everything well organised in Bookends, you don’t need to search in DEVONthink
I bow to your greater knowledge of the field – but what if the OP decides to develop the academic side of their career (is taking a PhD a sign of this?) rather than the practising medical side?
In any case, it often seems to happen to me that what helps in finding material is my own memory. I will have a vague recollection of some form of words, a theme, an idea, a subject, and then it doesn’t matter what program I happen to be in. I will be hunting inside my own head as much as anywhere else. But perhaps that is different from medicine, too. And perhaps I am an oddity in that I don’t find DEVONthink’s sophisticated search all that useful, and rarely use it – perhaps because I am looking for themes or ideas more often than words.
I am also a fellow physician and finished my PhD last year. Like many I used a combination of Bookends (for managing the citations), DEVONthink (for making links, searching for references and organising notes), Curio (for brainstorming and planning my thesis) and Scrivener (for writing). I kept two seperate databases in DEVONthink - one for my literature (the indexed Bookends folder) and notes and a project database where I kept all of the organisational material related to my PhD e.g. ethics documentation, data, emails, supervision minutes etc. A lot of what I did I have written about, including the various scripts that I wrote to automate many components of the PhD.
Regarding one database / multiple databases, I suggest it doesn’t matter so much with DEVONthink 3. The “how many databases” question is really a meta question about “how do I organize my materials”. Of course the answer depends on personal skills, but it is not something that needs to be answered definitively up front.
I usually recommend initially having one database. I like hierarchical folders (“groups”), others like tags or hierarchical tags, others like strict file naming conventions, others not. Just don’t lock into a set method up front unless it is a method that worked for you in the past. A beauty of DEVONthink (easier than Finder) is that reorganization is relatively “easy” – of course that ease diminishes as the collection grows, so do not wait too long to settle into a method.
Do not forget the DEVONthink AI “see also” and “classify” features. “Classify” can be helpful for organizing some kinds of documents. It works best when there are a lot of grouped documents in the database(s) so it takes a bit of growth in the database before the value of “classify” kicks in.
My own preference: Start with one database. Do organize with groups or tags or both. Do not dump everything into one group – that has a lot of downsides with any database, especially a research database. Take advantage of Smart Groups and Smart Rules.
@sanadsaad I haven’t seen much in this discussion about document annotation and how it relates to your note taking, and later on to organizing and drafting your dissertation. What is your experience – what has worked for you in the past? That will influence the choice of tools and method.
All pdfs are captured, renamed, and organized in Paperpile. Paperpile is a powerful reference manager, simple to use, and can be used in both desktop and iPad.
All pdfs are stored in Google Drive (this is how Paperpile works). You can also index them in DT to have the best of the two worlds.
I read, add comments and highlighted the pdfs using Paperpile on iPad. Everything you do will be synced, and you will have all your work available on many platforms, including DT.
Then, I export all comments and highlights to a markdown document (yes, you can do that in Paperpile) and index the markdown document in DT.
I process the markdown document by writing permanent notes using the Zettelkasten approach (there is a book on how to do that in DT, including how to make links, etc.). When the markdown document is processed, I delete it.
I use markdown for my permanent notes, keep them in a Dropbox file, and index them in DT. You can always import them to DT if you want.
When writing a paper, I search all the relevant notes in DT, export them to Scrivener, create an outline, and then write the article.
When I fell the first draft is good enough, I export it to MS Word (because journals ask an MS document), add the references using the Paperpile plugin, and complete my task.
Some additional thoughts:
You do not need Scrivener to write your paper. You can do everything in DT.
You can use Bookends instead of Paperpile, but after testing both apps a lot, I found Paperpile easier to learn, maintain, etc. Besides, reading pdfs on the iPad with the Paperpile app is an excellent experience.
By following this workflow, you will increase your productivity. Do not waste too much time overthinking how to organize your permanent notes. Just set a time everyday to process your pdfs, write your permanent notes, and link your notes in meaningful networks. You will build a knowledge system that will help with your dissertation.
DT has everything you need to be successful. Good luck!
Well, given these are two completely different products with different purposes/features … I would never try to write a thesis with DT only. Scrivener highest priority, with Word/Pages/WhateverWordProcessor next. Possible in DEVONthink? Yes, but why?
I would never try to write a thesis with DT only. Scrivener highest priority, with Word/Pages/WhateverWordProcessor next. Possible in DEVONthink? Yes, but why?
I would - just to prove it can be done efficiently
Writing a dissertation has two steps: writing and editing. Writing can be done in DT without any problem. All tools are there. Editing perhaps need an additional software, but I am sure @BLUEFROG has a solution for that as well :). Again, up to the writer!
Jim are you still working on your Theory of Everything dissertation using DEVONthink? The one that will gain you the Nobel Prize? We’ve been waiting for you to publish.
Jeff Taekman has published a comprehensive workflow for academic writing at his website: Workflows in Personal and Professional Productivity <://wippp.com>.
A lot to chew on, including his use of 15(!) Apple Ecosystem applications including Devonthink 3 and Devonthink To Go, which he uses his webpage to explain.
"Some criticize my workflow for its complexity, but I can assure you that it is amazingly functional once set up. I can’t imagine working any other way.
I’ll start with the “meta” infrastructure programs that make the whole workflow run, including the set-up. I’ll then cover how I collect information . After that, I’ll write about how I read and synthesize the information to help with sense-making. Then, I’ll go over extracting information . Finally, I’ll take you through the writing process , from concept to manuscript/grant submission. I’ve included links to past articles to show how I’ve set up my infrastructure."
Jeff’s “workflow” (a word that makes me cringe) is really not complex. Reading closely he’s not using all that software for every bit of writing. He’s mainly mentioning how he uses a collection of commonly-used software. (Other than OxMD – and there are other means to get info from PubMed.).
Pay attention to his task groupings: “Infrastructure” (stuff in the background); “Collecting and Extraction” (got to get the documents somehow); “Synthesizing” (got to make senses of a pile of annotations and notes); and “Writing” (the point of all the above). Anyone doing the work discussed in this thread will do the same.
Not having heard of Paperpile before I have just checked it out. Personally I don’t want to let Google pry into my life any more than it does already. Also a pretty interface for organising etc pdfs when really all that’s needed, at least all I need, is to rename a pdf on arrival and put it it in a folder in Dropbox which is indexed to DT3 where I can further do whatever is enough. Even if not renamed an ocr using DT3 will access the content.
I think this is good example of why different academic fields (or maybe better termed different readers of academic publications) are adapted best to different workflows.
It sounds as if your work may be such that a scholar in your field can assemble a curated collection that is nearly complete with the classic or seminal articles in the field. The field may grow at a slow enough rate that an active scholar can continue to read and collect new key articles in the field as they are published. In such a field, organizing these classic articles in a reference manager may well be an ideal solution.
But in clinical medicine and many research areas of medicine as well, published knowledge grows so fast with a doubling time measured in months that no scholar or clinician can fully master even small areas of specialty.
As a practical example, suppose I am accessing the literature on knee bracing for cruciate ligament injuries and I want to share with my students or colleagues a few pertinent articles as part of a class lecture or a draft of a paper or presentation or for my own use while writing a draft article. I will typically share such articles via a URL:
Note this link is basically a dynamic website updated regularly which contains not only (a) Full text of the article but also (b) Links to similar articles in the literature, (c ) links to other articles which have cited this article, and (d) links to references cited by this article.
Now contrast that with the classic output of a reference manager which gives simply static references in AMA or MLA format etc:
Now which is more useful for research or teaching or clinical use in medicine - a collection of dynamic URLs (stored in DT3), or a set of formal MLA-styled citations ready for publication (stored in a reference manager)?
Note that if/when you get to the point of actually preparing a manuscript for formal publication, the conversion from URL link to formal citation is easily done. But I suspect even for active publishing scholars in medicine, that represents a minority of cases; the active URL links in DT3 are generally more helpful as a primary means of saving citations.
So that is my dilemma as to why I presently consider DT3 to be my primary repository and reference managers to be secondary. I am interested in learning other approaches to this issue.
(By the way- it is not exactly an either/or situation. It is of course possible to link a URL to a citation in a reference manager, just as it is possible to link a formatted citation to a DT3 record. My main point is that citation manager software excels when your main desired output is a formal list of formatted citations; DT3 excels in almost every other situation.)
What I really hope is that someday academia gets over AMA/MLA/APA/NLM formatting and just accepts URLs as valid citations. That makes so much more sense to me. At that point, I would see no need at all for reference management software; DT3-like or generic bookmark management software will replace it for all.
Good point. All apps have tradeoffs. At any rate, a reference manager is much more than just rename PDFs. You need it to capture references, cite references using differ styles, read the papers, take notes, keep these notes, etc. in my case, paperpile helped me to capture, manage, read, and cite all kinds of reference without any friction. As always, there are other apps available for testing.
You present the scenario very well, and the use-case is very clear. Incidentally, you illustrate why DEVONagent is so useful in some disciplines, but not in others. I could never think of a use for it myself.
I have worked in various fields over a zig-zag life. Most recently in history, psychology, and latterly in counselling/psychotherapy. As you observe, these are disciplines that move rather more slowly than medicine. Historians typically write books that take several years to complete, and often refer to archival material. There is no URL, and I suspect there never will be, for a page in General Rawlinson’s diary (held in the Churchill Archives Centre in Cambridge). I spent long hours photographing the diary, and even longer hours transcribing it. (And were I to write an article about it, I could imagine that it could take twenty years before anybody read it, and maybe a century before anyone wrote another.) Archival research is a marathon, and a very uncertain one. I remember that my professor (not a historian) expressed an interest in going to the National Archives in Kew and researching the use of machine guns during WW1. “Will it be in a box marked ‘machine guns’?” he asked. Sadly, I had to tell him that it could be in any box in the whole damned archive, and any nugget of information could as easily be in the middle of a discourse on women’s petticoats as anywhere else. “Seek and thou shalt find” was not written by an archival researcher.
So, I’m afraid academia as a whole is unlikely to move to URLs as valid citations, because so much material in so many fields will almost certainly never have a URL. Even if every page in all the archives in the world could be digitised and given a URL, that still leaves material in private hands. DEVONthink was very useful to me for storing photos of archival material together with transcriptions of it, but Bookends was better for references, in my work.
As to finding things using DEVONthink’s undoubtedly sophisticated search functions, there were times when it occurred to me that a particular idea (e.g. deportment) expressed in a source work was important, but the word “deportment” never appeared in the text. It was implied, not stated explicitly. That is when human memory, and different tools, like Tinderbox, become useful, because they allow you to map concepts and connections rather better. But I have strayed off topic.
Anyway, I am glad that medicine is moving fast . We need it to do that.
I know this gets off point - but that’s interesting. Do you not see a trend toward having most archival sources digitized? I realize that does not necessarily mean they will be available publicly on the web - but are there really major library collections without a plan to scan and make it accessible remotely?
Indeed - just as you have pondered what the use case is for DevonAgent, I keep pondering the use case for both Tinderbox and for the more recent interest in graphing connections such as Obsidian can do. They are all fascinating pieces of software, but these application depends very much on the nature of the underlying discipline being studied.
At the risk of continuing off the point (apologies to all) – a trend – yes, of course. But just to take the case of the Churchill Archives Centre, it has material relating to about 600 people, each of whom has anything from one box to several hundred boxes of material (for Churchill himself, it is about 3,000 boxes). When I was working there I was aware of about three or four staff in attendance. I cannot imagine how many years it would take for them to digitise all the material. I recall that at another archive, when I asked for a photo of a document, it was considered so fragile that it first had to be sewn into a transparent envelope made of an inert material to protect it. It took several days to complete the request. Then again, there are customary practices when working in historical archives that may not be obvious to outsiders – such as preserving the order that the documents are in within a box. It may not be important, but in some cases the order itself is a potentially valuable piece of information about the material. I once showed a professor a box of my own family heirlooms, and I noticed that he was careful to put them back in the box in exactly the same order as he had found them. It was a tell-tale sign of someone used to working in archives. Digitisation would only give us part of the evidence. Indeed, the opportunity to go directly to a document in a database is potentially problematic, because the researcher might miss something from the context – the other documents in the same box. So it is not as simple as it might seem at first.
. We need it to do that.