Have been meaning to polish it a bit more, but I suppose now is as good a time as any to share it. One outstanding feature that I think will be pretty useful is caching user comments--right now it re-constructs the comment feed on every page load, which isn't ideal.
The funny part is that @raquo and I posted our "follow" apps/extensions (separate projects) this week and got no love from HN. And then this thread took off for some reason. Timing? Titling? God I hate the randomness of marketing.
First, the title of your submission:
> Show HN: Search HN by keywords, user, score, Read online, on mobile or RSS
It's completely unclear what your product is. Honestly it sounds like you packaged up HN with some more CSS and an RSS reader or something? I don't really know. "Read online?" How else would I read HN? "Search HN?" I don't really have anyting to search for, and if I did I would just use Google.
The reason film42's post took off is because people were interested in following. Yours didn't even mention following.
Now let's look at film42's title:
> Show HN: HN Follow – Follow Your Friends on HN
My first reaction? "Oh cool, I have a lot of friends who use HN, I wonder what they've been saying."
Then perhaps the best part of film42's app is that he used HN's styling in a way that almost felt like you hadn't left HN. It is completely familiar, and, as others have pointed out, it feels like something that should have been part of HN already. The friction of onboarding was completely eliminated - I didn't have to sign up for or download anything, but I could use it and wanted more. The missing auth is probably a mistake in the long-run, but missing that feature probably reduced some friction getting started.
At the end of the day, you have 5 seconds to impress me or I'm gone. It's quite possible, when your app doesn't take off, that the only thing you did wrong was not use those 5 seconds.
I'd be interested why you think mine didn't work. Chrome extension link: https://chrome.google.com/webstore/detail/logdfcelflpgcbfebi...
It certainly could have been a matter of posting time - film42 posted early in the morning when most of the stuff was old. It only took a few votes to get to the front page, yet a lot of HN users were online and voting.
But I think perhaps the bigger problem is around simplicity and familiarity. A chrome extension is always a hard sell because to play with it I have to download something that I have to dig out and uninstall later. I want to play with something before I "buy" it. Now people are asking for film42's app to become a chrome extension, but I bet if he had posted a chrome extension few people would have downloaded it.
Getting people to a moment when they are pleased with your product before they give up on it is probably the most difficult aspect of marketing/growth.
Feel free to check out my follow page: https://hn-follow.desh.es/?user=film42
Another example, HN top 10: https://hn-follow.desh.es/?user=hn-top-10
I wish there was a way to do a bulk query (ala Elasticsearch), so I could get all comment ids in one request. The response time would still increase with increasing number of users, but it would be marginal compared to the current situation.
I don't think HN is doing that great a job in surfacing the best content or encouraging better discourse in comments, but making it easy to follow favorite users instead of going to the front page would make it worse, IMHO.
https://chrome.google.com/webstore/detail/hn2go/logdfcelflpg... (shameless plug)
... I definitely wondered about this. All of a sudden, it turns HN into a popularity contest where the elite get all the attention and everyone else is completely ignored.
While the effect isn't overwhelming, making the content of a limited pool of 'thought leaders' more prevalent is certainly the bias that a karma system is intended to create. Of course HN presupposes that high karma correlates to high quality content and a high quality user, which is probably why as I understand it, karma gets used as a weight for things like the effect of flagging or downvoting. But if they didn't want it to be a popularity contest, I would assume they wouldn't make user karma public. You don't give people a metric without expecting it to be gamed.
I've never heard that one, and it's not the purpose. PG added it years ago when the YC alumni community had gotten too big for everyone to know each other.
That's hardly HN's goal, though. HN itself has code and features that encourage some users to vote for content purely based on who submitted the content or what company they work for. HN's certainly not a democratic voting society where votes are treated equally and we're all encouraged to vote purely on the merits of content.
I do like HN a lot, even with that lack of fairness built in to the system. I agree that letting normal users follow each other would make the 'problem' worse (this is what brought Digg down, as a core issue), but it's not like HN sees this as a problem, exactly.
That is not true. What gave you that impression?
I understand that you may not have built those features with that favouritism consequence in mind, but it's there for sure. Also, it seems likely to me that you'd consider this favouritism to (probably) produce higher-quality content hitting the front page more frequently, which I've certainly seen as a stated goal of HN (quality is more important than fairness or user preference, hence titles changing, etc). So maybe it's not a bad thing. I dunno, really.
I'm admittedly a bit sensitive to charges of favouritism on HN because we put so much effort into trying to be even-handed, even though (a) it's impossible to balance these things perfectly, (b) it's hard to measure whether we're getting it right, and (c) some people won't believe what we say anyhow. For what it's worth, the corresponding complaint on the other side (i.e. among YC founders) is that HN is biased against them.
Hmm? What do you mean by this? If you're referring to the (YC) submission tags, that doesn't indicate that particular bias.
Note that submissions from certain people/companies add extra relevant context
> Note that submissions from certain people/companies add extra relevant context
Yes, I think I agree. I'm okay with the way things are, ish. It would be better with more transparency, though, but HN's "goal" seems to be quality, and they certainly think that quality is increased with increased visibility to YC-related submissions. My only problem would be HN posing as "fair", which I don't think they do.
If you are a very well known user like patio, it's possible that you'll get a good amount of upvotes more easily. For most users, however, it seems to me to be a matter of good title, good content and luck as to whether or not your submission does well.
Patio's comments, in my experience, are an outlier in quality/length, so it's hard to use that as a trend.
Anyway, to me it seems that, with the rare exception, HN is pretty good in each submission being judged on its own.
I'll take a look. Thanks for sharing!
There is a mirror floating around of news.arc, the original version of HN from pg, but it's pretty dated and missing tons of the secret sauce.
Lets you follow users, topics, and much more. Also provides RSS and JSON feeds. It's open-sourced too, under MIT license.