Ask HN: Working Memory And Understanding

[mpg33]

I think i have the same problem with programming as i did with math. I have a hard time understanding what each variable/piece of logic represents/is used for and then when i finally do understand it, i forget not long after finishing with it.

It's almost as if my short term(working memory) does not move over to long term memory.

[pasbesoin]

Don't neglect / fail to pay attention to:

- Your physical health

- Your environment

When I'm feeling off, I have great difficulty tracking items and getting a "holistic" picture. And things like allergy or another chronic illness can leave a person feeling "perpetually" off, to the point where you think it's "you" rather than the result of these symptoms.

If you are in an environment that is constantly calling your attention, it's difficult or impossible to build such a mental model. Even when you are not being specifically called upon, loud noises, people in your peripheral vision, etc. can tax you to the point of disfunction. We've evolved to pay attention to such things; some people more than others seem to find this very difficult to "overcome".

An anecdote: I had a friend who was getting crappy scores in chemistry. A large problem for him, as he wanted to go pre-med. He always studied with music on. I suggested he turn the music off. After the next test, he thanked me profusely -- his score had jumped a grade level or more.

He hadn't consciously experienced any problem with the music and his studying. But, anecdotally -- and with a strong supporting opinion on his part -- there was one.

P.S. I'll add that fMRI and the like are beginning to show that stressors literally descrease or "shut down" areas of brain function. When you feel threatened, you brain restricts "higher" function and strengthens more "basic" function. Researchers interpret this as a survival mechanism; in dangerous situations, immediate action is paramount and intensive analysis (and delay) can be deadly.

Get stressed, and you will never remember "those function parameters". Your brain simply isn't in a place to do so.

[dholowiski → pasbesoin]

+1 Physical exercise. Getting more oxegynated blood to your brain will do wonders for your memory, IQ, understanding etc.

[hvass]

I recommend using Anki. Active recall (meaning you are asked something) is much better than passively reading and spaced repetition have been shown to help learning. If you dig into memory you will surely read about the 'forgetting curve' so usually a program such as Anki has an algorithm to ask you at the point you are about to forget something.

And by now you should wonder what is Anki. Well, it's basically a software flashcard. The neat thing about it is if you find something easy to remember you just click that you find it easy, so it wouldn't ask you as often, but if it's hard, it will be repeated often until it's drilled in your memory.

Here's the link: http://ankisrs.net/

P.S. For learning languages the same way I would recommend http://www.memrise.com which also incorporates in a lot of their lists visual/auditory cues, which I cannot stress how helpful they are in remembering words. Try the SAT vocabulary list, it's fascinating how easy is to remember words when you associate them with a visual cue.

[mpg33 → hvass]

visual memorization/associative memorization i do not have a problem with..

when i am trying to understand code visual memorization does not seem work..

[gahahaha → mpg33]

I am reminded of this article on learning math: http://www.ams.org/notices/201010/rtx101001303p.pdf

"""The alternative route to acquiring problem- solving skill in mathematics derives from the work of a Dutch psychologist, De Groot (1946–1965), investigating the source of skill in chess. Research- ing why chess masters always defeated weekend players, De Groot managed to find only one difference. He showed masters and weekend players a board configuration from a real game, removed it after five seconds, and asked them to reproduce the board. Masters could do so with an accuracy rate of about 70% compared with 30% for weekend players. Chase and Simon (1973) replicated these results and additionally demon- strated that when the experiment was repeated with random configurations rather than real-game configurations, masters and weekend players had equal accuracy (±30%). Masters were superior only for configurations taken from real games."""

And what I'm thinking is that maybe (just maybe) you just haven't programmed enough to become a "master programmer" with the ability to remember your program like master chess players remember their games.

[gwern → gahahaha]

There are other results on chess, you know, like grandmasters spending more effort on falsifying their suggested moves: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.117...

Correlation is not causation, and of the two correlations, I think falsificationism is more useful than quick board-encoding - you don't play chess from memory, you play the position the game is currently at.

[Gormo]

I've read a lot of articles that suggest that multiple n-back exercises[1] work well to extend the working set of short-term memory. I've been using Brain Workshop [2] for a few weeks to test it out, and I've certainly noticed an improvement in my performance scores from doing the exercises - I suppose I'll see in the near future whether there's a noticeable impact in my ability to "cache" information in practical contexts.

[1] http://en.wikipedia.org/wiki/N-back

[2] http://brainworkshop.sourceforge.net/

[rumfa]

More about N-Back you can read here http://www.gwern.net/DNB%20FAQ also check out nootropics.

[xxjaba]

I struggle with this problem as well. I often feel like the people around e learn more with less effort - perhaps because their brain's RAM has more space. Oh well, I supplement my lack of a large memory with stubbornness.

I find going back and re-reading (often manny times) helps tremendously. The first time through you may only commit 30% of the concepts to long-term memory. The second time through you will pick up a bit more, and the next time even more. As the % of concepts in your long term memory grows, the easier it will be to pick up the ones that are giving you trouble, since your brain has a larger number of memories from the prior readings to make concrete relationships.

[tokipin]

I have a similar problem but it doesn't stop me from being a good programmer. What it means for me is that when I sit down to work on a large project, it can take me a while to get up to speed with it as I reload/reaquire the knowledge. And I think that's totally normal.

In your case it might mean that your memory isn't "flaky," but rather that the way you perceive the content of a program is high-bandwidth, which would imply deeper understanding of the content even though such perception has the side effect of making your memory seem worse.

Still, you should check with a doctor on whether you might have some weird anemia or something.

[brador]

The time spent comprehending could be better spent building your own.

I found this early on after suffering similar problems to yourself with lack of recall.

For example, Early this year I used 3 days trying to find a good multiplayer server side system for my hobby flash games, Didn't understand the ports/sockets jargon and eventually said fuck it and built the thing from scratch in PHP with GETs in under 2 hours. It just works and I know how to fix it if it breaks.

So my verdict - build it yourself, your way. It's faster and you'll learn more. Don't be afraid to try.

[mpg33 → brador]

this sounds great...except my day job is bug fixing of an existing software

[mattmanser → mpg33]

Yeah, don't do this. Your co-workers will rapidly come to fear you touching their code if you keep rewriting it all the time.

Try and pick up their 'style' if there's a consistent one and code like them. The way they name their variables, etc. If not it'll just take time. The more you touch particular parts of the code base, the more familiar you'll get with it. The more you work with someone else's style, the more you'll be able to parse it.

[danielharan]

Not sure what the ask is.

Check your diet, sleep, exercise.

There's lots of stuff about n-back in the comments, which is definitely worth investigating.

If your memory has markedly deteriorated, go see a doctor and get a full check-up.

[SkyMarshal → danielharan]

>If your memory has markedly deteriorated, go see a doctor and get a full check-up.

And make sure to ask him to check for pernicious anemia, among other things.

http://news.ycombinator.com/item?id=3294598

[SkyMarshal]

There's no easy answer to this. You have to rewire your neural pathways till they're more effecient at working memory, and the only way to do this is with lots of hard work doing things like math and programming that require high working memory. It will be painful for a few years, but keep at it and it you will get better at it. But there's definitely no silver bullet here, at least not at this point in history.

[lukifer]

Make copious notes, add lots of comments, and don't be afraid to refactor for readability/understandability. Write both comments and code like you're explaining it to a third-grader. When each chunk is easy to understand, it'll be easier (or possible) to see how it all fits together.

It's not YOUR brain that is deficient; it's the human brain in general. Code should be written around that fundamental constraint.