We've had a huge spike in traffic lately, from roughly 24k daily uniques to 33k. This is a result of being mentioned on more mainstream sites. I hope this spike will subside, like past ones have. In the meantime I may temporarily hack a few things to make the site faster, like putting fewer results on threads pages.
You can help the spike subside by making HN look extra boring. For the next couple days it would be better to have posts about the innards of Erlang than women who create sites to get hired by Twitter.
Ok, ok, enough Erlang submissions. You guys are like the crowdsourced version of one of those troublesome overliteral genies. I meant more that it would be better not to submit and upvote the fluffier type of link. Without those we'll be fine.
Actually, I said I agreed to the post I was replying to, not to PG's post. Sorry you think it had zero value.
For what it's worth, I never said nor meant to imply that I thought the website of the lady who wanted a job with Twitter isn't suitable for HN, though I don't find it to be exemplary. It's a whole lot more on target than (as previously stated) articles on macroeconomics, politics, or flirting.
None of this is to say that I actually disagree with your point on the silly nature of what garners karma and what collects down-mods, but that is a topic for another time.
holy shit, i stepped away for like 30 minutes and the entire front page is Erlang. I just submitted the Innards of Erlang article as a response to PG's post/an inside lol for the community (the article is a decent primer though). Almost the entire front page is a bit much guys...
Ooooh, HackerNews meme. You realize how tempting it is to submit this to TechCrunch, etc., don't you?! TC Linkbait Headline: "Elitist Hacker News shuns the rest of us by pretending to be about Erlang! Oh, and somehow this relates to Twitter!1!!!1!!"
pg the Problem isn't that fluffier stories are up its that people are hungry for more of everything. I read Hacker News, Reddit, and even Digg in that order mostly every day and if there was another site that provided interesting stories and or decent comments I'd add it. It is the pursuit of information that's driving me mad.
pg, I have wanted to read about your thoughts on the Ruby programming language for a long time now, I have simply not wanted to impose or be a nuisance by clamoring for an essay . . . that would be absolutely great I think, I have gathered a few quotes from you about Ruby, and they are all positive (it is one of the reasons I developed an interest in the language) and I would love to read more: positive, negative, and neutral.
As much as I like being pompous and saying HN should try not to go into decline, this one I'm happy about. It was all a joke on PG, it had a terrific effect on the stories here in a way that was instantly noticeable here, and it was actually pretty funny.
Now we just make sure this fades quickly enough that a month from now nobody thinks they're "trendy" for mentioning Erlang Week.
That might work.Sure it will probably make the site hotter, but as long as potential users are evaluated, it probably won't hurt the quality of the site.
If we shut the door to new registrations, we definitely shouldn't make an invite only available from a member. I'm sure I would have had trouble getting an invite that way.
What might work is deciding on a growth rate that can be handled by the site and the community, and then limiting the number of sign-ups each month, or even day. We could just use a list of email addresses for new users and send out X number of invitations a day.
And we could also do more discourage those wacky comments. We could have new users do a short test, something trivial that requires them to have actually read the user guidelines. Maybe a multiple choice quiz where they have to distinguish between a trollish comment and an acceptable comment. Also we could start them out with a probationary account with stricter rules. Maybe if their account is newer than 5 days, and they have less than -8 karma, they get banned, or forced to read the site guidelines again before they posted more comments.
Though if the site does implement anything like an invite system it's going to be hard. Some people are going to accuse the site of being elitist. So it would have to be communicated that any rules are solely for all comments civil and professional, not because we are trying to exclude anyone. Hey, I like some of the pun threads on reddit, and trolls on yahoo answers, but I think this site would get ruined if even a few trollish comments were allowed.
Or include some fairly simple auto-generated questions.
"At what point does y = 2e^(2x) cross the line x = 2, accurate to 2 decimal points? y = __.__ x = __.__"
Extremely simple for a hacker, daunting for the general public. (answer: tinyurl.com/b95cdw rounded up with x = 2.00, of course)
That might backfire. I might think less of a site that thought doing trivial math problems was a sign of my level of intelligence.
And making it harder might also be bad. If they are too hard, you'll get people who are just signing up for the entrance test. Just think about how much press google's puzzles on their billboards got. And somebody will of course put the answers on the internet.
I would go further, and venture that 'serious hackers who will make good HN citizens' and 'people for whom the above problem is trivial' are two different sets (though the intersection of those sets might be significant).
I thought about this a while back - a big negative to it would be that if someone sees traffic coming from HN and comes to sign up and answer questions and feedback in the comments. I've seen lots of great first comments after signing up by people who had their content reached.
A similar idea would a hard limit on comments for new people for the first day? 3 or 5 perhaps?
You know I'd love to see a writeup of the issues and challenges you've faced with this persistence strategy as the site grew. You know, more details of what you decided to lazy load, why, and how, and the impact it had.
Technically speaking, I find hacker news persistent strategy one of the most interesting things about the implementation.
I use a similar strategy for my blog and just playing with larger datasets I've certainly run into hard limits on what seems to be acceptable.
PLT Scheme is not "interpreting" code. More than that, Arc
adds a bunch of huge overheads that could make things
substantially faster. I had some patches lying around that
made things around 4-5 times faster (including the news
33k users per day is piddly little bits. I think that a platform change is called for, the current recipe is clearly no longer up to the task. Which is a pity, because the lisp code really is quite elegant. It's just that the web and high performance and lisp are not usually used in a single sentence (other than this one...).
This question might sound ridiculous to majority of people commenting here, but why are you making the site boring instead of trying to improve its performance (putting additional hardware, ditching Arc, whatever)?
From what I can tell, correct me if I am wrong, part of the purpose of this site is to operate as a testing grounds for Arc. The focus is on making the language better. So rather then improving performance by providing better hardware, it becomes a matter of improving the site through code. Either by changing the code running the site, or by improving the language itself. But given that this is more of a side project, the resources needed to improve the site are limited. Given the unusual spike in traffic, rather then using limited resources to fix it, or sacrifice the motivation to improve the code/language, the boring article approach is being used.
I kind of love this, despite having little or no interest in Erlang. It's like that sequence from _The Wiz_ where Richard Pryor keeps declaring different colors to be in style and everyone shows up wearing them. pg should come along every so often - once a month maybe - and just announce that all submissions should be about an obscure language.
I wonder how hard it would be to inspect the referrer and if an abnormally large influx of users start coming from a specific referrer, start redirecting to programming.reddit.com or stackoverflow or something.
More people visiting a site does not make it bad..but more submissions and vote-ups could be the reason for the dilution. Just make that process as hard as possible. How hard can it be to make a user-unfriendly html-hiendish registration page, eh ? :)
- Content free linkbait articles that mention the words "hacker" or "hackers" in the title, like the absolute jewel we had yesterday, where one of the supposed must have skills for a hacker is a high wpm typing speed (I really do think that - in the near future - humans will be able to train monkeys to type)
- Anything that would appeal to the majority of the "social networking" crowd
- Anything that offers anything remotely related to the "program software (or any other worthwhile computer activity) with this product, no brain, education, or hard work required!" crack dream
- Did I mention MBA/Suit related stuff? :)
- Wannabe MBA fodder
- Last but not least, I mean (sigh), how can I say this without coming off negatively? well, can't be said, but I will do it. I really do think that the best Heuristic is the Room Temperature Test: anything that would appeal to people whose IQ hovers even close to room temperature ought to be kept off the site period - this includes all of the above, particularly the MBA/Suit Neanderthal stuff.