bslatkin

17 ms·

121.
It is ridiculously easy to refactor Go(onebigfluke.com)
149 points by bslatkin 14y ago|88 comments
122.
by bslatkin 14y ago
(i love this thread) Brogramming's lasting meaning? Being Goofus in production: pushing code live from your laptop with no commits, no warnings, whatever, mostly in situations that feel like this: http://youtu.be/O_HyZ5aW76c?t=1m
123.
"Commit then flip"(onebigfluke.com)
1 points by bslatkin 14y ago|0 comments
124.
Why is Map Reduce faster? -- A practical example of data joins(onebigfluke.com)
9 points by bslatkin 14y ago|0 comments
125.
by bslatkin 14y ago
Couple months old but a great article I didn't see here. The highlight for me: "What the smartest innovators and smartest marketers I know have in common is not asking a lot of questions but making sure the most important questions get aske
126.
Learn More By Asking Fewer Questions(blogs.hbr.org)
2 points by bslatkin 14y ago|1 comments
127.
by bslatkin 14y ago
While learning D3 I was surprised that it's a good general system for manipulating DOM. Using SVG is just the prettiest application of it.
128.
by bslatkin 14y ago
Also check out the Globe visualization: http://consumersurveys.googleusercontent.com/insights/consum...
129.
Google Consumer Surveys Election Data: Quest for Best Visualization & Analysis(google.com)
39 points by bslatkin 14y ago|5 comments
130.
Distillation: Small Data is the Purpose of Big Data(onebigfluke.com)
3 points by bslatkin 14y ago|0 comments
131.
Why cohort analysis?(onebigfluke.com)
43 points by bslatkin 14y ago|1 comments
132.
Show HN: Cohort Visualizer - A handy tool for browsing cohort datasets(bslatkin.github.com)
1 points by bslatkin 14y ago|0 comments
133.
Google or Gallup? Changes in Voters’ Habits Reshape Polling World(nytimes.com)
3 points by bslatkin 14y ago|0 comments
134.
by bslatkin 14y ago
"Perhaps it won’t be long before Google, not Gallup, is the most trusted name in polling."
135.
Which Polls Fared Best (and Worst) in the 2012 Presidential Race(fivethirtyeight.blogs.nytimes.com)
30 points by bslatkin 14y ago|13 comments
136.
by bslatkin 14y ago
I'd say your variant doesn't work because there's a hierarchical relationship between white-collar and blue-collar workers (one is hired by the other). This is not true for frontend vs. backend engineering in my experience. They are suppose
137.
by bslatkin 14y ago
Thanks for clarifying.
138.
by bslatkin 14y ago
What you said: "When I'm doing back-end development, I love having a front-end guy who can make it look pretty" How that sounds to me: "When I'm done with a hard day's work, I love having a wife at home who can make me a drink" Seriously! R
139.
by bslatkin 14y ago
Adam Smith was pre-industrial revolution. He had no idea how far this would go. But he probably understood diminishing returns or hidden costs.
140.
by bslatkin 14y ago
Really? What have you been doing?
141.
Over-specialization is the curse of our profession(onebigfluke.com)
25 points by bslatkin 14y ago|52 comments
142.
by bslatkin 14y ago
Why didn't they just run Hadoop on a sample of the logs instead of the full dataset? It seems like they were just looking for popular related queries anyways. Any ideas?
143.
Doing Cohort Analysis with Map Reduce(onebigfluke.com)
5 points by bslatkin 14y ago|0 comments
144.
by bslatkin 14y ago
Yes. You are responsible for replication, failover, disaster recovery, etc.
145.
by bslatkin 14y ago
Uptime comes from how you design and build a system, not which IaaS provider you choose.
146.
by bslatkin 14y ago
OP doesn't realize that big companies like Yahoo cannot sign API contracts; they get terminated immediately. Remember Ping/Facebook? I think the ToS termination issue is the single largest risk facing start-ups operating in another company'
147.
by bslatkin 14y ago
Two things: 1. As the researcher you can set a range of validity for open-ended numeric if you know it ahead of time. We recommend doing this. 2. The system automatically detects outliers in all cases and puts them in a separate bucket to r
148.
by bslatkin 14y ago
The results are a representative sample of the US Internet population.
149.
by bslatkin 14y ago
Most people I know have a FB account for their dog. Is it just me?
150.
by bslatkin 14y ago
This. Finding rigorous engineers is hard.

More ›