The following is bad stats, more gut: I would discard 50-75% uniques as random inbounds or non-repeating (habitual) users, bots, people who have switched wifi links, etc. That's going to be a lot closer to reality. Even then, at say 1.5M monthly uniques, divided by 30 = 0.15M/3 = 0.05M = 50,000 daily uniques. I am hugely skeptical there's that many daily human readers and would tend to believe the community aspect would fail at that scale. Besides, we often recognize one another posting. I'd have guessed 10-20% of that, give or take, ie. maybe 5,000-10,000 dailyish humans, max.
> That's going to be a lot closer to reality
Trying to tease "reality" out of these numbers drove me mad, so I stopped trying. Are bots "real"? Are spammers? How many "real" users does a unique IP represent? It's better, I think, to just measure a bunch of different things, note how far apart they are, and resist over-interpreting.
J/K, is it possible to share what kind of hardware is HN running on?
Pretty much all I know about the hardware is that HN runs on a single core on a capacious server. Perhaps kogir will show up to say more.
FreeBSD 10.1
Nginx 1.7.10
Racket 6.1.1 with some HN and FreeBSD specific patches.
2x 3GHz Intel Xeon-IvyBridge (E5-2690-V2-DecaCore)
8x16GB Kingston 16GB DDR3 2Rx4
SuperMicro X9DRI-LN4F+_R1.2A
Adaptec 71605
9x 1000GB Western Digital WD RE4
2x 200GB Smart XceedIOPS SSDDo you have any insight as to why? My first guess is that it's non-HNers who stumble across it because of a google result or similar.
I'm pretty sure it isn't one-time stumbling upon the site. It's that most regular HN readers don't log in.
I really, really enjoy the comments on this site.
Especially because HN shows no summary or anything of the linked article: I'm not going to click on a link if the only information about it that I have is an (often purposefully) abstruse title and a domain name. Once there are a few comments that give me an actual impression of what the article is going to be, then maybe I'll go back and actually read it.