# Dyadic scanning allows for better processing of information When reading a scientific article or any textual work that is dense in information, the natural approach is to do a sequential scan, ie, to read it from beginning to end and try to memorise things on the way, as illustrated below. ![Sequential Scan|300](https://radimentary.files.wordpress.com/2020/05/gif_sequential_2x.gif?w=500&zoom=2) This is particularly inefficient because you have to sacrifice agency; you are forced to follow along, not knowing where the proof is headed, and this makes you more passive than you should be. On the contrary, what you should strive to do is to scan the article dyadically; one can also see it like a breadth-first search. ![Dyadic Scan|300](https://radimentary.files.wordpress.com/2020/05/gif_dyadic_2x.gif?w=500&zoom=2) Basically, you make a first pass looking at the paper's organisation and the main results. Then, for your second pass, you locate the main lemmas and the innovations of the paper. Only for your later passes do you delve into the technicalities of the work. This way of reading makes you much more active, allowing you to predict where the paper is going. Information compression is akin to prediction, so this also increases your information retention. Furthermore, the more you perform dyadic scans, the more you will develop your intuition of where to focus your attention (see [[The brain is like a muscle]]). --- ## 📚 References [‘Of Math and Memory, Part 3 (Final)’.](https://radimentary.wordpress.com/2020/05/16/of-math-and-memory-part-3-final/) Radimentary, 16 May 2020.