They are not going to make improvements to Hacker News. They have proven that.
They don't have to fix it or make any improvements. They can run a very poorly performing site with no features and a horrible appearance, because they are entrenched, and they literally own quite a large portion of their userbase.
Unfortunately no matter how many people invent software that does the same thing as Hacker News but works much better and offer it up for free, nothing will be done.
The owner of the site has too much invested in its technology stack (Lisp) and its particular implementation and appearance. The improvements necessary probably require going back on some of those fundamentals, which is never going to happen, because they are too proud.
I upvoted you because I like seeing people who go against the grain here, who hold combative viewpoints and are willing to be blunt about faults, regardless of whether or not I agree.
However, after seeing this:
> ... because they are entrenched, and they literally own quite a large portion of their userbase.
> Unfortunately no matter how many people invent software that does the same thing as Hacker News but works much better and offer it up for free, nothing will be done.
... I have to say, if you are disappointed, have you thought about testing this theory? This website has not been around forever. It isn't the first of its kind either. I'm probably not the only one here who came to this site after having frequented a bunch of others that were essentially the same thing. So I don't expect it to be the last one, either. If someone has the interest and the skill set, I say, please do make this. Worst case you waste a few weekends or something.
There are a number of Hacker News clones, i.e. people got fed up and even went ahead and figured out how to mirror the Hacker News content. Far as I know very few if any gained any popularity, but its not because they weren't better.
I quite enjoy lobste.rs. Despite the lack of discussion, the majority of the topics posted are interesting and turn over pretty fast (in comparison to some of the other replacements).
I even suspect Hacker News isn't that fussed if most of its community moves elsewhere. The core community (that it was built for) will stay, and the postings will become more focussed on what the core community finds, reads, and is interested in.
There's a place for those things. I also think it's important that you don't get overly consumed by some frivolous notion of those concepts, prioritizing it above honesty, authenticity, and everything else.
To that end I often feel it's refreshing when people say something that is not a majority opinion on this site, and don't hold back.
Seriously, this bug has been around for years and people have constantly complained about it. The only reason they would have for not having fixed it at this point is they simply do not care.
It's dangerous to claim anything is the only reason something happened because it increases the odds you could be wrong.
they simply do not care
If it's negligence that you're implying, consider a more valuable alternative. The expiration link bug is starting to look like the HN test of one's ability to tell what's important to care for. Any hacker would want to develop this ability.
Can they change the 'lasts' parameter so that things don't expire as quickly?
Another idea.
To be more sophisticated, instead of killing fnids, could first
; replace them with fns that tell the server it's harvesting too
; aggressively if they start to get called. But the right thing to
; do is estimate what the max no of fnids can be and set the harvest
; limit there-- beyond that the only solution is to buy more memory.
(def harvest-fnids ((o n 50000)) ; was 20000
Can they change that number? Maybe that would help.
Another idea. Could you change it so that when they are harvested the functions are serialized to disk (maybe temporarily), and make fns a function or something so if it had been removed from memory it could be deserialized? It wouldn't need to be a serialization solution that would work for arbitrary closures, just something that would work for those cases, like timed-aform or whatever.
I have barely ever tried to use Lisp and I don't know much about this system so this may not make any sense.
Another idea. Use one of the Hacker News clones that people have built over the years that doesn't have this problem.
That's debatable, but we are talking about a matter of degrees. Wage-based employment is a moderate improvement upon slavery by direct force. The relationship between investors and startup founders is a moderate improvement upon that, but still has quite a few similarities with traditional slavery.
Those are some very absolute claims. If you worded that as observations and not so absolutely (or provided some quotes to back up the claims), it would sound a lot more credible.
I never minded the appearance too much, but the usability bothered me. I month or two ago I tried some of the chrome extensions (specifically HackerNew[1]), and it's made a big difference. In-line replies, hover to see user stats, collapsible threads, etc. I'm not enamored with the infinite scroll, but it's not horrible.
It makes it look a little slicker too, but I truly couldn't care less about that.
It's never really bothered me much on mobile, and I use it extensively while mobile. It does bother me that comment indentation is not properly represented in elinks, but that is sort of an esoteric complaint that I can't really blame PG for.
> Unfortunately no matter how many people invent software that does the same thing as Hacker News but works much better and offer it up for free, nothing will be done.
We should name this - how about "The Craigslist Problem"?
Part of this is that HN is much more than the technology, it's the community. Sure there may be some scaling issues, but building a HN is something many people here could do in an afternoon. Building the community is a much more complicated problem.
They don't have to fix it or make any improvements. They can run a very poorly performing site with no features and a horrible appearance, because they are entrenched, and they literally own quite a large portion of their userbase.
Unfortunately no matter how many people invent software that does the same thing as Hacker News but works much better and offer it up for free, nothing will be done.
The owner of the site has too much invested in its technology stack (Lisp) and its particular implementation and appearance. The improvements necessary probably require going back on some of those fundamentals, which is never going to happen, because they are too proud.