Generally, my pattern goes like this: 1. Read summary/abstract. 2. Then skim intro. 3. Then read discussion/conclusion. 4. Then methodology. 5. Then read f

Generally, my pattern goes like this: 1. Read summary/abstract. 2. Then skim intro. 3. Then read discussion/conclusion. 4. Then methodology. 5. Then read front to back. 6. Then read front to back again: but now we're being extra skeptical: we're questioning everything, looking for common bad practices, checking doubtful references to see whether they're true (the number of references that I follow where I don't agree with the author that the reference says what they claim is too damn high). 7. on loop: read from front, hit and study topic/part you don't understand, focus on and study that part, upon understanding continue read through to see whether understanding elucidates any other parts you didn't understand. 8. repeat step 7 loop until you understand everything. Now running concurrently with most of these steps, there's two escape/break decisions to be made at almost every point: a) do I understand paper sufficiently, if yes, then stop reading. b) is paper actually worth continued reading: if no, then stop reading. overall, I'd say the vast majority of papers aren't worth reading. Some are easy and only require ~one read. A very very very very small amount are both worth reading, and take a substantial amount of time. multiple reads, several weeks/ months to grok. I don't think I'm being hyperbolic when I say the number that justify that probably are less than 10 in your entire lifetime.