CP NoteBook v. DevonThink again (with apologies)

I know from doing a search of this forum that this is a FAQ but I have a question slightly different from previous ones. I want to put all of my working papers (courses, notes, papers, and even personal stuff) in one or two big databases. I can use either NoteBook or DevonThink Pro. But it occurred to me that virtually all of my materials are in outline form. That means that most of my files are essentially outlines. So why should I use DT Pro for this project? If the lowest unit of the database is an outline, wouldn’t CP NoteBook (or OO Pro) be a smarter choice?

Thanks for any info (esp from Bill).


When I “work” my database I’m looking for information.

That covers a spectrum from looking for a phone number or a bibliographic citation – what I’ll cal “flat” or archived information, or “mere facts” – to prospecting for new ideas and insights, which I’ll call “live” or interactive information that helps me develop new approaches and understandings about a topic with which I’m already familiar.

DEVONthink Pro, CP NoteBook and OO Pro excel at storing and providing access to archived information, although I must say that my appreciation for the powers of DEVONthink Pro to help me find stuff keeps growing. If all I need is a phone number or bibliographic citation, these three applications (and a number of others) will do the job for me, although I must sat that DT Pro can handle more such information than most.

But DT Pro blows away those other applications at the level of interactive assistance in exploring database contents and finding interesting relationships among individual words and collections of words (documents) in the database. In the past, I had graduate students as research assistants. I really valued the bright, inquisitive ones who suggested new ideas and information relevant to a research topic. Now I use DEVONthink Pro as my research assistant, and I must say that it’s better than most.

Example: “Smart” search features in DT Pro. I use the Tools > Search window when I’m doing searches. A DT Pro search produces a list of search results, initially ranked by what DT Pro “considers” the relevance of found items. Often, that relevance ranking is useful, but it may not be. I’m responsible for understanding the topic and deciding whether a suggestion by DT Pro is useful to me.

Let me walk you through some of the things I may do on that Search page, especially when I’m trying to go beyond “mere facts”.

Phrase search for “hazardous waste”. In 100 milliseconds, 112 results were listed. By sheer coincidence, the top-ranked result was one of my papers, “Case Studies of Hazardous Waste Problems In Louisiana” which focussed on (United States) federal and state approaches to developing regulatory programs, with a discussion of the role of risk assessment. (Risk Assessment of Hazardous Waste Sites, American Chemical Society, 1982.)

The Search window Context button: If I click on that button, DT Pro will slide out a panel that lists terms found in the search results that I might want to further explore. The terms (in this example) are fairly obvious: hazardous, waste, chemicals, toxic, etc. If I click on one, that initiates a search for that term in my database.

The Search window Spellings button: If I click on this button, DT Pro suggests a list of related terms, which turns out to be quite interesting in this example. The list includes, of course, “hazardous” but also included “nonhazardous”, “ultrahazardous”, “biohazard” and “geohazards”. I might well want to explore some of those terms using new searches.

Opening my paper, I can click on the See Also button. Suppose I want to explore discussions of the use of risk assessment (and methodologies used) for environmental policy and regulatory development in the United States, the European Union and other areas. The list is a treasure trove! Not all the items on the list are relevant to my interest, but many are. So I’ve got a good start on that topic. And I might want to explore how risk assessments are made in non-environmental topics, as well.

So DEVONthink Pro has done its job as a good research assistant, starting me off with a good collection of material relevant to, e.g., a writing project on risk assessment uses in policy and regulatory development. Of course, I’ll keep pressing these buttons in doing exploration of suggested documents. And I’ll probably use DEVONagent to look for additional relevant material on the Web, and dump interesting results into my DT Pro database.

After this excursion, let’s answer your original question. Here’s my answer:

I don’t care if the material I put into my database is an outline, a book, a Web page, or whether it’s poetry or prose. If it’s useful to my interests, I want in in my DEVONthind databse, where I can do more with it than I could in CP NoteBook or in OO Pro. :slight_smile:

Bill and others: thank you for your reply but…

  1. Crazy as I might sound, what if I don’t use the Search and See Other features of DevonThink? Is the DT advantage still greater than NoteBook?

  2. To repeat my point, the common or basic unit of organization in DT, unlike NoteBook, is NOT an outline. So

if (a) a user doesn’t need advanced search or see other and (b) manipulates outlines then it seems to me that DT is overkill compared to NoteBook.

I am not critical! But I am still trying to understand if DT has other features that I am missing. Again, thanks for any input.

I use both DTPro and an Outliner (I own NoteTaker, OOPro and Notebook). I do a lot of presentations and writing. I have found that DTP is the best for gathering my information and notes. All the notes, clippings, PDFs, etc can be put in one spot and easily organized. They are more accessable during the writing process. The outliner is the best for crafting the expression of my thinking… logic… pattern of presentation. I don’t see it as an either/or proposition. Use both tools. I used the outliner tools by themselves for a while but still found large data organization to be cumbersome. I didn’t like wandering all over my NB/NT/OO document looking for what I wanted. DTP keeps it all right in front of me (use the three pane view) never fearful of misplacement. I use the outline to stay focused on my thinking, reasoning, writing and presentation.

In the development of DTP in the months ahead… from what I read here… you will be able to reference documents like OO, NB etc more easily in DTP making them part of your project file. You can link them now… even storing them in OPML or HTML format.

Hope that helps…

Thanks Bob and this is exactly what I wanted to here. Just to clarify: DTP can NOW

  1. Link to external Notebook files. So you can click on a NoteBook (.nb) file in DTP and it will launch Notebook and the .nb file.

  2. But! the .nb file is not indexed, right?


  1. You can export as .opml and have DTP read a notebook file as one big outline.

  2. But this will be (always) be a two step process in going back and forth between NB and DTP.



I interpreted your original post in this thread as asking whether, because a document is in outline file, it should be stored in an outliner rather than in DT Pro (which may not have been the best interpretation of your post).

Anyway, my response indicated that for reference purposes, the content of a document is more important than its style or format, and DT Pro is a champ at working with content. I mine my databases for ideas and references, so if I’ve got a file that was originally developed as an outline, and its got information pertinent to my interests, I dump it into a DT Pro database.

Personally, I occasionally use an outliner to lay out the structure of a writing project, then just start writing (but usually the writing isn’t done in an outliner). More and more, I’m actually writing inside DT Pro. And with the considerable improvements of Cocoa text in the last couple of OS X updates, the List styles in the DT Pro Toolbar usually satisfy my outlining needs, which are very simple. When I’m satisfied with an outline covering the topics I need to address, I simply create a new rich text document in my database for each outline segment and Link To it (contextual menu option) from the corresponding section of the outline. That creates a Table of Contents which could, using File > Export > As Web site, produce an HTML version of the finished project!

When I’m writing, it’s in rich text format and may include some images, URLs and links to other material such as other documents in the database, reference citations or whatever.

For final output, I’ve been using Pages a lot for the last few months, especially if I have to do a MS Word version. Most RTFD documents from my DT Pro database copy/paste perfectly into Pages’ Blank template. Images are properly set up and hyperlinks work. It’s easy to add a header and footer (whole document or sections), footnotes, and additional hyperlinks to reference citations, etc. If you’ve every tried to assemble a Word document from bits and pieces of RTF and RTFD documents (especially placing images) you know that’s a headache, and you will love Pages, which makes this a breeze and does the job much better than Word itself, through the Word Export feature. Links and bookmarks in the resulting Word document are perfect. I’ve opened the resulting Word files under Word just to check, but have never had to do further editing. Pages does a good job with PDF Export, but I’m not satisfied with Pages’ HTML export. For some projects where I have to produce very similar layouts of PDF and HTML output, I use Create to do the layout.

I understand now. I think I remember Erich, the CEO of DevonThink, saying in an interview that DT Pro was probably overkill if one is only manipulating a dozen or so files. More and more I find myself preferring to dump my stuff (research papers, lesson plans, clippings, etc) into several big files rather than hundreds of small one. I don’t think there is a clear advantage or disadvantage of the Few v. Many issue but I think DT Pro is much more approriate for people who like to split their materials rather than people who like to join their materials. Perhaps, the issue I am hung up on.

Thanks again.

I don’t join small files into bigger files, nor do I split large files into smaller files. My database contains many thousands of documents ranging in size from text snippets to PDF books and manuals running into hundreds of pages.

There’s a really neat supplement to See Also, which does essentially the same thing but works from a selection (word, phrase, sentence, paragraph or larger) within a document. It’s See Selected Text, which is a contextual menu option when one has selected text in a text or PDF document That makes large documents more useful, as one can ask DT Pro to suggest related documents based on a selection of particular interest in the large document.

What a cool thing to point out, Bill. I learned something new and great about DTP today! (I had never used See Selected Text before).

But I must say that while See Selected Text works well for narrowing the origin of your search, it does not necessarily suit selection of the destination…

For example, one of my bugbear documents is a dictionary of the Persian language circa the 1800s. This is an awesome resource for translating poetry, since most classic poems are from before that time.

No matter what my topical database, I’ve found that I need to keep that database OUT of DTP, since whenever I include it, it always ends up being #1 on every documents See Also list. Simply because it has so many unique words in it, it’s likely the match nearly all of the words in almost any other document I have. (Strangely, I encounter the same thing with certain novels, like those by Dosteyevsky).

I find this to be the case in general with larger PDFs and text files: they start coming up as a candidate match for everything. But if I segment them into smaller pieces, the “hits” that DTP suggests get more and more accurate. Sometimes uncannily so.

I have one database full of religious literature. Initially it was just a bunch of texts, because that’s what I had. But I found that although DTP could find things fairly well, it couldn’t relate them. Most of the books were like most of the others – which really wasn’t an interesting statement.

So I spent a few days and broke them up so that no entry was greater than 5 pages in length, and most were 1-2 pages. I tried to never break up an idea, but to represent the flow as a series of internally-coherent points.

The result was that DTP was sometimes able to find exact correlations I never knew about. It was almost spooky! But this could only happen once the data itself was properly “cued” for the DEVONthink AI.

One feature I would absolutely love is the ability to say “split here” in DTP. The way I ended up doing it was very nasty, involving hand-coded Emacs and Python scripts.