Apart from sleeping, majority of my time is spent working on comp or laying in bed watching yt/scrolling thru Twitter
Usually goes:
- 4hrs on comp
- 1hr mobile
In repeated cycles
Usually goes:
- 4hrs on comp
- 1hr mobile
In repeated cycles
Thinking about making a github repo called "bounties" where I'll pay users who create open source projects that solve desired tasks
50k datapoint backtesting
before optimization:
- generate timelines: >366 seconds
- calculate gains: 20 seconds
after optimization:
- generate timelines: 48 seconds
- calculate gains: 3-4 seconds (moving to frontend)
And I think I can still cut processing time in half If I
- Used lexical analysis to process different parts at the same time
+ (would reduce runtime but cpu time would be about same)
- added bulk operators that act over all datapoints instead of specific instances
+ converts type checking from O(N) to O(1)
- prevented duplicate operations for identical datapoints for static operators
+ reduces size of operation set
+ overhead would only be worth for cpu intense things like mentions and nlp stuff
before optimization:
- generate timelines: >366 seconds
- calculate gains: 20 seconds
after optimization:
- generate timelines: 48 seconds
- calculate gains: 3-4 seconds (moving to frontend)
And I think I can still cut processing time in half If I
- Used lexical analysis to process different parts at the same time
+ (would reduce runtime but cpu time would be about same)
- added bulk operators that act over all datapoints instead of specific instances
+ converts type checking from O(N) to O(1)
- prevented duplicate operations for identical datapoints for static operators
+ reduces size of operation set
+ overhead would only be worth for cpu intense things like mentions and nlp stuff
tldr, mysql doesn't follow JSON standards
https://dev.mysql.com/doc/refman/5.7/en/json.html#json-normalization
https://dev.mysql.com/doc/refman/5.7/en/json.html#json-normalization
learning how to do er diagrams is even less useful than UML diagrams but somehow also 10x more complicated
I think that go is like java in the early 90's-2000's where it's popular moreso because of marketing and not because the language is that great
google is to go as sun is to java
google is to go as sun is to java