News for hackers. Stuff that matters.
Re:Breaking the trillion-rows-per-second barrier with MemSQL by thinkMOAR on Monday March 19, 2018 @06:33AM (#16617364)
"When you deliver response time that drops down to about a quarter of a second, results seem to be instantaneous to users."

I don't think everybody agrees with this statement.

Re:Breaking the trillion-rows-per-second barrier with MemSQL by WorkLifeBalance on Monday March 19, 2018 @07:12AM (#16617546)
The threshold for most people is slightly lower than a quarter second, but it's certainly the case that above a quarter-second users will notice the delay[0].

However, people may be primed for longer delays still seeming instant. e.g. Smartphones for many years had a built-in 300ms delay on any click event, and even without that event typically still have delays on many 'instant' actions.

So while the delay will be registered as being present, it may not be registered as "this site is slow" but "it's just a natural delay".

[0] https://psychology.stackexchange.com/questions/1664/what-is-...

Re:Breaking the trillion-rows-per-second barrier with MemSQL by spyder on Monday March 19, 2018 @12:11PM (#16619513)
The threshold for noticeable touch "rubberbanding" is 1 ms:

https://www.youtube.com/watch?time_continue=52&v=vOvQCPLkPt4

Re:Breaking the trillion-rows-per-second barrier with MemSQL by mickronome on Monday March 19, 2018 @09:12AM (#16618061)
I completely agree, a quarter of a second is absolutely not instantaneous.

Especially not to oddballs like me that has some kind of perceptive 'bug', it's like I lack a 'motion filter'. One effect being that movement in computer games doesn't feel completely fluid until the refresh rate is close to 200Hz.

Before internet and slow loading web pages might have lowered peoples expectations, I remember Human-computer interaction guidelines stating that < 0.1 was experienced as instantaneous in almost all cases. Above 1s without feedback started to cause measurable stress in test subjects.

Since this was before computers were ubiquitous, it's probably a good measure for how we react on a more basal, subconscious level. Anything measured today is likely to be include learned expectations, so a quarter of a second seems like reasonable learned expectation of the perception of instantaneous in that particular context.

Re:Breaking the trillion-rows-per-second barrier with MemSQL by manigandham on Monday March 19, 2018 @11:16AM (#16618989)
This isn't about raw physics and biology but rather the use of business applications by users, specifically with regards to analytics and exploratory queries.

Anything less than a second to run a query and return results is considered "instant" (also sometimes "interactive") for analysis.

Re:Breaking the trillion-rows-per-second barrier with MemSQL by thinkMOAR on Monday March 19, 2018 @11:58AM (#16619403)
Since the example data is stock data, i think it is safe to assume people don't want to be 249.9ms late with their purchase :)
Re:Breaking the trillion-rows-per-second barrier with MemSQL by manigandham on Monday March 19, 2018 @12:23PM (#16619638)
What? No human trades that quickly, and this is not about automated HFT trading systems but rather running database queries and what typical business users expect as "instant" results.
Re:Breaking the trillion-rows-per-second barrier with MemSQL by rplnt on Monday March 19, 2018 @07:15AM (#16617558)
There's context missing. But I don't think anyone would imagine an interactive application (say, a game) when the topic is some data manipulation.