4 ms·

Funny enough, the proposals here remind me an article on site metrics I read recently. To rephrase the relevant bit, "99.9..% uptime is highly misleading if a

Funny enough, the proposals here remind me an article on site metrics I read recently. To rephrase the relevant bit, "99.9..% uptime is highly misleading if a small number of your customers are simply down all the time." Unemployment has a long-known flaw cited in the article, especially relevant in this economy: the significant number of adults who have simply given up looking for work, considering it hopeless, and who are therefore not counted in the unemployment statistics. Any metric gives you just a tiny, keyhole-like view on the world. A big part of the battle is then to understand what your metrics are blinding you to.

makmanalp 8y ago
This is a nice connection and it's a kind of what's known as a structuralist argument - what you're saying is that the approach of "uptime is uptime is uptime" is wrong, and distribution of /where/ downtime happens actually matters a lot. Economics did run into similar issues with GDP because it assumed for a long time that "money is money is money", until it found out that just because both Iraq and Vietnam have similar GDPs doesn't mean that they're at all equivalent - one is heavily dependent on oil and has few development prospects, the other has a wealth of productive know how in a ton of different areas (shameless plug, I work on this tool): Iraq http://atlas.cid.harvard.edu/explore/network/?country=108&partner=undefined&product=undefined&productClass=HS&startYear=undefined&target=Product&year=2016 http://atlas.cid.harvard.edu/explore/network/?country=108&pa... vs Vietnam http://atlas.cid.harvard.edu/explore/network/?country=239&partner=undefined&product=undefined&productClass=HS&startYear=undefined&target=Product&year=2016 http://atlas.cid.harvard.edu/explore/network/?country=239&pa... Each node here is a specific product being exported "significantly" for a country of that size. --- So by a similar argument you could look at how similar the population that goes in and out of unemployment (seems likely to me) is every year, and what that implies about policies and job opportunities in the US.

makmanalp 8y ago

This is a nice connection and it's a kind of what's known as a structuralist argument - what you're saying is that the approach of "uptime is uptime is uptime" is wrong, and distribution of /where/ downtime happens actually matters a lot. Economics did run into similar issues with GDP because it assumed for a long time that "money is money is money", until it found out that just because both Iraq and Vietnam have similar GDPs doesn't mean that they're at all equivalent - one is heavily dependent on oil and has few development prospects, the other has a wealth of productive know how in a ton of different areas (shameless plug, I work on this tool): Iraq http://atlas.cid.harvard.edu/explore/network/?country=108&partner=undefined&product=undefined&productClass=HS&startYear=undefined&target=Product&year=2016 http://atlas.cid.harvard.edu/explore/network/?country=108&pa... vs Vietnam http://atlas.cid.harvard.edu/explore/network/?country=239&partner=undefined&product=undefined&productClass=HS&startYear=undefined&target=Product&year=2016 http://atlas.cid.harvard.edu/explore/network/?country=239&pa... Each node here is a specific product being exported "significantly" for a country of that size. --- So by a similar argument you could look at how similar the population that goes in and out of unemployment (seems likely to me) is every year, and what that implies about policies and job opportunities in the US.