I feel dumb(er) when typing, and smart(er) when writing

[dgan]

I have noticed many times, writing down a problem and work on it on a paper is much more productive than staring at my screen and/or jumping right to code.

Last time it happened, I spent a day on a Coq exercise, when I finally I decided to grab a paper and pen, and solved it in 5 minutes.

Is this something that anyone else is experiencing? Is it something that has been (dis)-proved?

[PragmaticPulp]

Context shifts are actually great for changing your perspective and reinvigorating your ability to work around road blocks.

Switching from screen to paper is one way of context shifting. You could also switch a whiteboard.

You could switch to talking through the problem verbally with a colleague or, lacking that, a rubber duck.

You could just get up and go for a walk to think about that problem. Context shifting from indoors to outdoors can be a significant change.

It helps to learn how to recognize when you’re stuck, not making any progress, and have maybe a bit of “learned helplessness” slowing you down even further. Shaking up your context can help break out of it.

[dgan]

I think you are right to generalise it to "context switch". Countless times I deleted my yet-to-be-posted, detailed Stackoverflow question, when after typing (ha!) it, i would realize my mistake.

I never tried duck - debugging, maybe should give it a go too

[thenerdhead]

Writing in my opinion is the best way to think through problems. It takes concentration and effort.

Jeff Bezos says the "smartest thing" they did at Amazon was to ban low effort presentations and embrace writing long-form briefs for each meeting. That made sure the problem was thoroughly thought out by the meeting owner and is read by the group in detail for the first 30 minutes of the meeting.

In open source, many projects adopt a RFC(request for comments) process in which a problem is written about exclusively with solutions and asking for public feedback. This works effectively in ecosystems like Rust, Python, JavaScript, and C#.

If you think about knowledge work in general, 80% of the job is problem solving, creative thinking, etc. 20% of the job is implementing / coding. Jumping straight to code was something I did when I was early in career, but now it's the last thing I do until I understand the problem in detail.

[lambdaba]

Actual writing is the best form of pseudocode :)

I like the anecdote, this and the two-pizza rule are some nice bits of Amazon "management lore".

[pdpi]

It's useful to understand the difference between problem space and solution space. Writing code (in most languages) is all about writing in solution space, and sucks at talking about problem space. I find that most people tend to jump into their editor way too quickly and start writing the solution before they've thoroughly analyzed the problem.

When I jump from editor to pen and paper it usually means that I'm having trouble mapping the code I'm writing to a part of the problem Im'solving, and I switch to exploring the problem more before going back to the solution. I'm going to go out on a limb and say that that's pretty much what happened to you as well.

[karmakaze]

I've noticed the same when I write design notes in a Github issue vs a Google Doc. The latter results in a longer document with less technical focus => lower quality, more suitable for a less-technical audience. Even knowing this, my writing style and content drifts in those directions for each medium.

[drakonka]

There's a little notebook by my computer at all times for this reason. It isn't even just writing that helps when I'm trying to figure out a problem or pick a good approach, but doodling. Doodling while thinking seems to help things click into place and generate new ideas sometimes.

In my first job in games, my lead at the time took issue with seeing me occasionally doodle in a notebook in front of my PC. It wasn't a case of sitting there doodling all day and getting no work done, or not focusing. I'd only do it for a few seconds or a couple of minutes at a time, when thinking through a task or problem. After that interaction, I stopped doodling at work for a long time, to my (and I believe my work's) detriment.

[jbluepolarbear]

I’ve been a professional for 10+ years and I have always used a notebook to work out issues. I fill a notebook in about 3 months. It’s really helpful for working through a problem in a free form way. I draw diagrams, tables, pseudo code, etc. Often forcing myself to think of the problem in a different way (writing it down) helps me understand what needs to be done.

[Tarsul]

I just read an interview in the guardian with author George Saunders that has 7 tips for writing (and reviewing what is written). One fits here especially: " Print - Revision is “not meaningful unless I print,” he says. “There’s a visual difference in reading on the page versus the computer. I don’t trust it unless I’m reading a hard copy.”" So, basically if an award winning writer thinks like you do, there's nothing wrong with it :) (btw. his other tips are great too) actually, there are studies about this that agree (e.g. better retention when reading paper than when reading on screen) but I don't have them saved.

https://www.theguardian.com/books/2022/jan/08/always-wanted-...

[usrbin]

Anecdotally I notice the same. I'm curious about whether it's the look of print vs. a screen, or the physical form of the text being a physical object vs. stored on a device. Do we get a similar effect from using e-ink devices?

[dgan]

Yes, I have noticed that too, but it's less pronounced in my experience. Just the fact that I can turn the page, look forward, come back and underline in color, feels much more engaging

[periheli0n]

Pen & paper provide the opportunity to use the full 2D plane for laying out your thoughts. I find this extremely valuable during conceptualisation, for sorting ideas, finding and highlighting connections etc.

Typing however is one-dimensional. Thoughts need to be serialised in order to type them out. Sometimes this is exactly what you need. For example, a linear narrative tends to make it much easier to explain your ideas to others. And of course code is just a serialised version of a program.

[lambdaba]

Contratulations, you have discovered [Hammock-Driven Development](https://www.youtube.com/watch?v=f84n5oFoZBc)! (also https://github.com/matthiasn/talk-transcripts/blob/master/Hi...)

[saghm]

Honestly, due to my perpetual difficulty with writing both neatly enough to read and quickly enough to keep up with my thoughts (it's always been an either/or thing), I feel much smarter when typing (on a physical keyboard; I also can't type that quickly on a phone keyboard). Adding in the ability to arbitrarily jump around and edit stuff, I feel like I can always express myself in a way I'm happier with at a computer than on paper.

[rudolph9]

Same here! I have never felt comfortable hand writing things but the act of typing in an editor such as vim feels very liberating.

[wodenokoto]

I don’t know about you but I feel a piece of paper provides much more freedom than a txt file.

While writing text is slower than typing, it’s much, much easier to choose where to write, and to draw lines and boxes. All tools that helps thinking.

> when I finally I decided to grab a paper and pen, and solved it in 5 minutes.

In defense of your hard work, it’s always the last solution you try that works.

[pedrovhb]

Recently I started writing out all my ideas and designs of personal coding projects in a notebook of graph paper. It's pretty incredible how I can see and think about high level things (like general architecture) much more easily than if I'm in front of a screen staring at code or a project template. The cycle of thinking, building it out until you reach the next tier of work, and then stopping to think again is a lot more efficient in the long term than continuously solving the next problem without having a concrete target direction.

One of these days I'll start seriously building my Graph Paper Markup Language processor, which will use scanned notebook pages with special symbols recognized by OCR to automatically tag and add dates to my notes. It's all sketched out already.

[jonahbenton]

I can attest to having seen papers on this, but am not able to track them down right now. The theory is not dissimilar to rubber ducking- using your brain to work in the problem space- which helps transition your mental model to solution domain.

Like you and others this experience with paper has also been my experience for decades. Perhaps of interest is that I transitioned from using paper to a Remarkable tablet, which is surprisingly great from a writing experience perspective, but which is deficient in the "paging through" experience that is so important with paper. Its still preferable to me to work digitally, and to unify around a device I can read on, and I am able to use it to work in problem domain reasonably effectively.

HTH.

[wrp]

My problem solving ability is noticeably much stronger when working on paper versus at the screen. This is true for both mathematical reasoning and narrative writing. The effect is so great for me that I don't even try to do serious editing or study of a text on screen.

I have just a tentative explanation for this, that it is related to the apprehension and distraction of having a machine staring you in the face, waiting for you to do something. I had exactly the same issue in the days before personal computers, when I would sit facing an electric typewriter.

[GuB-42]

For me, it is the opposite. I feel much more productive when writing code, to the point where my idea is "when I am not coding, I am not working". I don't even have a pen and paper at work. I have a phone with a stylus that I use as a substitute from time to time, and I sometimes take notes on text files, but 99% of my thinking is done while coding.

Code is what keeps me down to earth. It may seem like an unnecessary detail to think about how to pass a variable when you are barely familiar with the subject, but if you can't get something that simple right, chances are that your are going it the wrong direction and when time will come to do the implementation, you will be in a world of pain. Of course, a lot of that early code ends up being thrown away and reworked, but in the early phases, it is rarely a problem.

It probably has to do with a top-down vs bottom-up approach. I am much more comfortable working bottom-up, thinking first about the details and how they fit the big picture, rather than to top-down approach, where you think of the big picture and work up the details later. I think that both approach will eventually lead you to the same destination, so choose the one you are most confortable with.

Also, context matter. I do mostly low-ish level programming (mostly C++), I don't know about coq, but it seems to be more maths than technical engineering, so maybe a different mindset is required.

[YetAnotherNick]

Without all the context, I think it is due to structure of the problem. Some problem can't be understood clearly with just text, but needs diagrams etc. Coq problem seems like one that could fall in this category. And if you are like me, rough diagramming in computer seems slower/harder than with pen and paper. For these kind of things, I found apple pencil+ipad to superior to both as you don't need to worry about space and page flip and you have all the advantage of free hand scribbling.

[dgan]

Interesting, i never thought about e pen and a electronic pad. Will think about that

[karmakaze]

In a similar vein, I've noticed that I'm significantly dumber when I use a laptop with the touchpad vs using the same laptop with a mouse. I mostly grew up learning to write software with a mouse and later in life consume media using touchscreen phones/tablets. Now I know how to get in the 'work' mode. [Also noticed that comfy trackpants are worse for thinking than jeans or other office pants.]

[sammycaution]

Alan Kennington wrote about the "asterisk method" to learning mathematics [0]. There is a line that I'm thinking about in particular when you mention the utility of paper and pencil over reading on screen (or solving problems in general): "Copying material by hand is important because this forces the ideas to go through the mind. The mind is on the path between the eyes and the hands. So when you copy something, it must go through your mind!"

Maybe there is something about owning a problem that is more personal when writing it down and working it out by hand versus reading it on a screen.

[0] http://www.geometry.org/tex/conc/mathlearn.html

[Syzygies]

Back in the day, a shopkeeper in Japan might realize that they'd forgotten to pull out their abacus, that they'd been playing this musical instrument in their head to do calculations all morning.

One guides a slide rule to roughly the right position based on intuition, then homes in on the answer.

Calculators make people dumber.

I'm torn between doing all my math on 40 lb laser paper with an assortment of Staedtler Pigment Liners, or on my iPad Pro using the architectural drawing program Concepts. Either way, I'm drawing. A digital interface is like a bird taking flight when one becomes algorithmically involved.

[bcrl]

Personally, I find that I remember things much better after writing them down with pen and paper versus typing. It's a very different kind of flow relative to programming on a computer with a keyboard which feels much more ephemeral.

[brewmarche]

What I experience is that I can remember things I write a lot better. I’m not sure why. I can type a lot faster than I write.

Also for problem solving I agree. I also don’t know why but I feel a lot “freer” when I have a sheet of paper in front of me.

[dgan]

Totally agree on last part