Using Circus Ponies Notebook to collect and sort information for your thesis (or article)

Ipsa scientia potestas est.
(Knowledge itself is power.)

I got an interesting question regarding the use of Circus Ponies Notebook for thesis notes.

I am actually looking for a program to organize and to code my papers for my thesis. I would like to code my paper regarding different topics, so afterward I can see which paper I can use for each topic. Would this program be able to do this?

The answer to this question warrants its own posting. The term “code” is a bit ambiguous, but if I understand the question as it was meant, the question is whether you can use CPN to collect notes and mark or tag papers where the information should end up (e.g., theory, results, discussion; or chapter 1, chapter 2, etc.).

The short answer to the question is: “Yes, but …”, so let’s have a look at the “but”s.

First off, there are different ways to deal with papers, depending on what you actually want to do.

Classifying whole papers

If you are just interested in classifying or tagging whole papers, you might do this far more easily by just taking the PDFs and add tags to them. Mavericks (OS 10.9) supports tags, so does DEVONthink, and almost any reference manager. However, I would not recommend such a solution — it is far too unspecific. So, paper X with 3-50 pages belongs to the introduction, or to chapter 4. But where exactly does it belong there? And which part of the paper belongs where? It’s a rather superficial sorting that does facilitate a little bit of writing, but not that much.

If you want to do this with CPN, take a page for each paper, and add a keyword or sticker (the graphics attached to an outlining cell). You can also use a single page, create the structure outline of your thesis (introduction, theory, results, etc./chapter 1, 2, 3) and just write the name of the paper beneath the respective header. I would not recommend it, but it is easily possible.

Classifying parts of papers

More interesting is to tag parts of a paper as useful for your thesis — and where these parts should end up. After all, writing a scientific text is just building with LEGO bricks. You need to correctly cite the bricks for your argumentation from different articles and add your own bricks from your own thoughts, arguments, data, and interpretation. Looking at a paper this way … screening it for parts you can use … I found it really helpful. Of course, there are limits. You can’t just cherry-pick the results and arguments that speak for your thesis, and ignore the rest. Likewise, you always have to make sure you preserve the spirit of the article when you home in for the details: If author X strongly argues for position A overall, but mentions an argument for B (because you have to address counter-arguments and rebuttals), you can’t cite author X’s argument for B as if author X was for B.

But taking a “LEGO brick” view is really helpful when dealing with papers — and using them (correctly cited and correctly interpreted) for your work.

Here, CPN can help. No matter what you do, you need to copy-paste the text that is useful (the LEGO brick) into your own notes. Make sure you tag the cell with the source information. CPN supports keywords for cells, and using the citation information (authorname_year) as tag/keyword is really helpful here. Take care how you copy it — some copy & paste methods of CPN remove the keyword. So make sure you have the keywords always visible (ensure that “View” => “Keywords” is checked). If you cannot copy & paste the information from the article, you have to either use OCR (optical character recognition) software or manually transcribe it. Acrobat (the full version) does have OCR (“Document” => “OCR text recognition”), so do other programs. Have a look at this posting if the PDF file does not cooperate. I also strongly recommend taking a look at digital reading — either in this book on page 349 (“Reading Digitally”), or in this posting.

Then, there are multiple ways to deal with the selected parts of the paper. I focus on three I still use.

CPN with one page per article

You can create a CPN page per article and just copy and paste your notes about this specific article there. Then you have the information available. From an article of 3 to 50 pages, you only have to deal with the notes that are relevant for you. You can also add some needed context information. For example, what does the article argue for overall, how good were the studies done? Where do the arguments break down? What are confounding variables?

If you put the article in focus, this might be useful. You can still add stickers for the information in cells to “code” where this particular piece of information should end up. In writing my dissertation thesis, I used tiny graphics like “Th” assigned as stickers to cells to “code” information I needed in the theory part of my dissertation. Later, you can either scroll down each page, looking for this sticker, or use the “Multidex” “Stickers” page to see all the information beneath each sticker. It worked, but I would not do this again for writing a thesis. But I do still have CPNs for the information I copy-pasted out of the articles I have read. As a backup.


Simply use a CPN (or multiple ones) to collect the interesting parts of the papers you have read. It’s a kind of backup repository. It’s got one page per paper you have read, using the author’s names and the year as title. Add a, b, c, etc. when these people wrote multiple papers. Ensure that you tag each cell on the page with the source — easily done by adding the authorname_year as keyword and then expanding all outline cells and adding it to all of them.


Thesis Notebook

I think the much smarter solution is to start a thesis notebook. Create an outline how your finished dissertation thesis should be ordered. Usually, there is a structure that you have to use (e.g., Introduction, Theory, Methods, Results, Discussion, sometimes with multiple studies in an empirical part). Then copy-paste the relevant information (with the keyword) from your read papers CPNs into your thesis outline. Ensure that you remember that you have copied this information and that the information you copy is understandable on its own. This might mean you have to add some context information, e.g., in […] what the position of the author was, or that this was one of many arguments, or that this information is speculative. You are removing the information you use later — when you likely do not remember all of the paper, so provide some context information (trust me on this).

This has a number of advantages: You have all the information still available in the papers CPNs. Useful if you want to create Topic Notebooks. You can handle the information from the papers individually. Place it where-ever you need it. It’s not just that information x belongs in theory, you can place it below w and above y. You are also developing a “content outline” — not just a mere “structure outline” of your dissertation thesis. You can play with the LEGO bricks, yet always know a) where they came from (author_year information tagged/keyword to the cell), b) the context of that information (in […]), and c) flexible in its positioning in your outline.

And once you have created that content outline, added your own thoughts and data, you can make sure you have a coherent “narrative” and put it next to a good writing program (like Scrivener) and start typing.


Topic Notebooks

If you work with the information beyond your thesis, think about creating topic notebooks. They are similar to a thesis notebook, but instead of focusing on a single work, you create a structure to handle the information regarding a whole topic. You find more information on topic notebooks here.


So, yup, you can code your papers with CPN.

One caveat though. CPN is very powerful, but it still got a rare “page not found” bug. So do backups. And for Science’s sake, do keep the old versions of your backups. If you ever try to create a page and CPN acts strange, e.g., you do not see the “Untitled” or just two overlapping circles, press Undo (cmd + z or cmd + y), quit the program and reopen your files. Otherwise you might lose some or all of your data.

Other than that … yeah, I think CPN is extremely powerful and extremely useful for writing a thesis.


  1. Hi Daniel,
    nice article and thanks for the helpful tips.
    It seems that I always encounter new ideas to improve my workflow in CPN.
    I noticed that you copy a lot of information between notebooks, e.g. from literature notes to a content outline of an article or thesis. To keep track of the information and where it comes from, would it be useful to include links between the different notebooks? And, if you have any experience with links, do you know how stable those links are, when it comes to large notebooks?

  2. Hoi Tim,

    merci, and I agree, you can improve your workflow forever. One reason to focus on the actual consequences — the productivity in quantitative and qualitative terms. Improvement without better work isn’t really an improvement, even if the work would be done perfectly (if work were done) 😉 Just too easy to fall into the trap of optimizing for optimizing’s sake, esp. when it’s fun to do so. Anyway, that’s something I frequently have to keep in mind …

    Regarding your question, yup, I copy & paste a lot, but given that the source information is always available as keyword I don’t need links. The cells (and subcells) are self-contained and understandable on their own, so I don’t need to go back to topic notebooks. Regarding links in CPN, personally, I would be skeptical. Not sure whether links between notebooks work (you can link to webpages and within a notebook), but I would only use links to webpages. Links reduce flexibility. Once you have made a link, you cannot simply remove the information where the link lead to, or the cell where you did put the link. If you do, the link will be dead but probably still be visible, and that leads to confusion (made some bad experiences this way). Also, in most cases just going to the target page via the Table of Contents (card) or the nearest tab is almost as easy. So I avoid links as best I can.

    But that’s just me. Links might work (very well) for some purposes, but in my system they would show me that I have a problem with the structure and finding information. I rather just write a note where I find something …

    Best regards


  3. Hi Daniel.
    Since Scrivener has specific files for research, couldn’t you do all this inside one single thesis project and have all in the same place? You can use keywords, collections, colours for similar purposes. Since Scrivener outliner is not as developed as CPN you could use an outliner specific sw and export to Scrivener as opml. I’ve read your CPN posts but I still don’t understand the nead for that extra step between the reference manager and Scrivener. But that’s me (to use your words)

  4. Good point — I think that Scrivener it excellent for structure outlines, and actual writing. The ability to easily do version controls with snapshots — waow. But for planning a paper — CPN has much more to offer. I go into some of the differences in this posting, but among others, in CPN you can refer to individual units of information from papers and use them like Lego bricks — and always have the source information available. That’s invaluable, especially if you store the information from papers in CPNs, either a thesis notebook or topic notebooks. When you create a new CPN for the article/thesis, you can play around with the structure for a long time until it’s “right”. Then you can put CPN next to Scrivener and write the article. However, this is for scientific texts. If you are writing a text where you do not have to reference every idea or datum, Scrivener might well be enough. Personally, though, I really like the outlining features of CPN to plan the story and then write it down. It’s not that Scrivener does not have a research area, it’s that it’s much more higher level than what you can achieve with CPN. For some people this is … not enough, but actually the best way they can work with. Some people need this higher level perspective and their creativity would suffer if they go below that. But for detail-fans like me, Scrivener — while (almost) perfect for writing — would not be enough for planning a story (or article).

    Not sure, but I hope it answers your question. If not, please ask again. BTW, I still struggle with making links visible in the comments. This posting got a few, but they are only visible when the mouse if over the links (mouse pointer changes).

  5. Hi, Daniel,

    I’ve been learned a lot from your blogs and thank you very much for sharing this wonderful app with us. I’ve been tried CPN for a while and it does make a lot of contribution to my work.

    I just have a small concern about the software. What if the software not longer updated and not compatible with later computer OS? I know I could export the entire data as webpages or docs, but I found the “key word” function which is in your workflow (and mine too, thanks again) as reference is missing. That’s could be a big problem when we are referencing. How should you deal with this problem?

    Anyway, how should we avoid losing information or entire database from future risks?

    Thanks for your time!


  6. Hoi George,

    good question, that would be a problem. But hopefully not a showstopper. You can export the file, e.g., as .opml (Outline Processor Markup Language) which gives you the cell information which keyword (“category” in the file) was used. I think it should be possible to use this .opml file and either import it in another program or use a script to extract the necessary information in a usable form (e.g., recreate the lines with the keyword part in [] at the end of each line). I haven’t tried it, but I am optimistic. I do not consider CPN a data island and if (or rather: when) it folds, it will take some work and trials to find another solution, but it should be possible.

  7. I’ve “fast-read” your book and agree with the ideas. I use Scrivener and agree with your comments. Now I’m looking for a software to accompany a history research project to handle idea collection and inter-relation and hyperlinking of notes to pdf collections. So that having structured my notes, I can easily open pdfs and write it all up in Scrivener.

    Mind-maps are not the answer for me, I think. But your two key recommendations, Circus Ponies Notebook and Devonthink, are both Mac-only! Do I have to switch to the Mac and stare at its irritatingly non-matte screen?

    I’m tempted to fall back on the solution I use for my 8-person property research company’s institutional memory.

    We use Google Sites for our “How-Tos”. We have 9 tabs, housing all content, with the How-tos hyperlinked to a deep pool of knowledge. If we create a new process, we then document it, and in half an hour we’re sure we won’t ever forget the details. Search is great, too.

    Anyway – any thoughts? Since I can’t get hands-on experience of Circus Ponies Notebook and Devonthink, it is hard to tell what I’m missing. Scrivener is such a massive aid to writing, and your description of Scrivener is spot on. So do I have to buy a Mac? Or are there really viable Windows alternatives to these two Mac-only programs, or does it sound as if they (and any alternatives) would lead me to do what I’m anyway already doing with Google Sites, so I needn’t bother?

    Simplicity is a virtue – but accessing better functionality at an early stage is also important.

  8. I’d say that CPN would probably work, giving that you can insert PDFs by not copying but just linking to the files. But as you said, that’s Mac only. If it’s only the screen that turns you off, you can use an external monitor 😉 but I’d be hesitant to recommend such a change for one program alone. On the other hand, why not ask a friend/colleague who uses a Mac to try out CPN on his/her device? You can trial it for 30 days I think for free, and it gives you an idea whether it would work for you. Otherwise, I’m not sure. I don’t know the Windows world that well, but perhaps Microsoft OneNote might be worth a look.

  9. I’ve looked. I think it will do. It’s not dissimilar to how I use Google Sites, just faster and jazzier (though it feels more vulnerable to accidents). Thanks!

  10. Hi, just an extra note for your info: see final comment with a link to your site.

    Actually I find Reference Managers like Zotero puzzling, I can’t figure out whether they are intended only (or primarily) for tracking bibliographic references, or also meant to store pdfs. If the latter they do a poor job. When I imported locally-stored pdfs into Zotero it recognized none of the bibliographic information, clearly it relies on something else apart from the pdf itself to do this job.

    So if Reference Managers aren’t meant to store pdfs, how do other people store them?

  11. Thank you for the link 🙂

    As for reference managers — there are reference managers who try to identify the paper automatically using the title, author name, etc. Ideally, they use the unique doi. For example, Papers (for Mac) tries to identify the paper and offers you citation information (based on what other users have entered, so you have to check it). I think solutions like Mendeley with a network behind it might be more accurate, but I lack experience using them. As for Zotereo, I think it works best when you import the citation information as well (available on many websites). But given that it is an effort to enter (and check) the bibliographic information, I suggest importing only the papers you actually cite.

    As for how other people store them, some use reference managers and enter or import the bibliographic information, others simply use folders (with files à la authorname_year_title), etc., perhaps A-Z folders or per topic area (which has the disadvantage when a papers fits within multiple topics, here tags are usually better). My impression is that many people use (more or less) reference managers and/or refer to what they have cited in prior writings of them. As for me, the division between DEVONthink (all material), Papers (what I have read and want to cite), and CPN (notes with source information) works really well.

    But dealing with literature is an individual issue and everyone has to find the solution that works best for this person — and at that point in the career. Just don’t spend to much time on it — decide which papers are actually important and focus on being able to work with them (understand, use, cite) — not every bit of information you come across. After all, it’s only a tool/building material, not the work itself.

  12. Hoi Mio,

    thank you for the link. There are some people questioning whether CPN has a future and they might be right to be concerned. Personally, I would not be concerned. Yes, I have lots and lots of information in CPN, but I can get it out. But at the moment, no need to. And BTW, the website seems to be up again.

    I am more concerned that the talk about CPN folding might lead to exactly that event, a kind of self-fulfilling prophecy (akin to how a rumor can lead to a run on a bank and thus to its ruin). Personally, if they go for less frequent releases because it’s stable I’m okay with that. But one danger of this “we don’t need every feature” strategy is that people think a software is dying if it doesn’t do monthly updates (as if that were a good thing).

    So I’ll watch and wait unconcerend and continue to use it (and recommend it). If they do fold in the near or far future, so be it. It would be a loss for the Mac because it’s one of the best and (for me) most useful programs I own.

    Best regards


2 Trackbacks / Pingbacks

  1. Designing a Personal Knowledgebase – A Curious Mix
  2. RIP, Circus Ponies Notebook | ORGANIZING CREATIVITY

Comments are closed.