search

Chicago / lead-safe-api-docs

http://dev.cityofchicago.org/docs/lead-safe/
1 stars 1 forks source link

Clarification on a few header items on POST #3

Open tomschenkjr opened 7 years ago

tomschenkjr commented 7 years ago

There are a few items in the header from POSTing data that are not well defined. In particular,

geneorama commented 7 years ago

I looked up the ETag concept earlier and I think it's pretty interesting. It says it's a digest (MD5 for example) of the data. It's meant to replace the "time to live" field that can be in the header.

Years ago Rob Lancaster gave a great talk about using survival analysis at Orbitz to calculate the "time to live" for their data because price information can get out of date very quickly (e.g. the last hotel room gets booked while you're booking your room) and caching that many data points is expensive / non-trivial. We shouldn't have the "big data" problem, but it could still be very useful for us.

I'm not clear what all needs to be included in the hash, or if it's a hard and fast standard. For example, I wonder if you could also include the input record into the hash so that the hash could communicate a unique pair of input / output.