Hacker News new | past | comments | ask | show | jobs | submit login

At my company, we populate pre-production data with old production data. If an item number has changed, we have a script that converts the old item into a new item number or drops it altogether. If Product A is no longer sold but Product B is now being sold, the script will convert Product A sales and pricing figures into Product B's data. It's just so we have some "real" numbers to test with. Every few months, this data is refreshed with current (old) data.

Of course, we also have random number generation to populate the fields with junk if we can't get real data (personally identifiable information). In that case, it's just junk that looks like real numbers. This scenario from the article isn't that far outside of what a normal business might do.




> This scenario from the article isn't that far outside of what a normal business might do.

Sure, its something a normal business might do -- but its completely inconsistent with the facts. Neither the numbers nor the candidates matched the previous election. So whatever was released, it wasn't the previous election's data.

Had they said it was test data (whether resulting from rather substantial massaging of prior election data or something else) that would have been, if not necessarily the kind of thing anyone would be inclined to accept implicitly given the totality of the circumstances surrounding elections in the country, at least a superficially plausible explanation of the data released. But that's not what they claimed, they claimed it was the previous election's data.


Developer likely said: - Oh crap, its test data, I used the current candidates with past elections data just to make sure it was working. Result in the news after some levels of PR handling people: - No problem, its just past elections data.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: