>> I would be suprised if there was a single user who'd left HN because it lacked search.
While I agree with you, I think you're looking at it the wrong way. No user will leave HN because of search-issues, but a lot of users will be very happy if the site had good search.
imo it makes sense to do a trivial thing and make a lot of users happy. So while you're correct in saying that "quality of the submissions and comments" are more important, I see no reason why quality of content should preclude a decent search feature.
I see no reason why quality of content should preclude a decent search feature.
Because attention is a finite resource. I can only spend attention on minor problems by taking it away from major ones.
For example, right now, instead of wasting my time replying to this comment, I should be working on a new project we're launching soon. And I think I will, so forgive me if I drop out of this thread.
The link is an improvement over nothing, at a minimum it will cut down on the number of newbies posting 'how do I search' questions, at least, if they scan the footer.
Experience shows that this is probably not the case but then at least we can point them to the link without having to explain site prefixes, which is still a net gain.
If it were in the header (which is pretty crowded) it would be more useful, but I can see that that might lead to complications, if it were a miniature form in the header it would be better still, but I can see that that would really be much harder to integrate.
Searchyc.com probably can't handle the load if they are the 'sole provider of search' for HN, so it looks like that is not an option, when the HN robots.txt file locked out google because you had to crunch the numbers for the applications it was apparently touch-and-go on searchyc.
Now I'm really curious what your secret startup is that may solve the problem, and I hope they roll out soon.
I am currently part of a start-up, Favetop.com, which offers direct site searching through just the click of a site icon. I've found it to be quite convenient for searching sites directly and much better than constantly adding site prefixes or visiting the actual websites to search. You may want to give it a try and see if it helps you out.
Because attention is a finite resource. I can only spend attention on minor problems by taking it away from major ones.
I think you've just identified one of the best ways to "break out of the box", or dare I say, make a paradigm shift.
The issue isn't whether attention (or anything else) is finite; it's "how to make it infinite", at least for the time being.
Example: I need to add horizontal scrolling (a critical user need) to my Report Writer. I would also like to add F-Key View Toggling (a nice to have) to my Form Look-ups. As a single founder, I am constantly asking myself, "How can I have it all, even with limited resources?" I have found that the trick is not to focus on what's limited, but on how to embrace my constraints and use them to get clever out of necessity. In this case, I wrote a single program that used the same routines to do both.
This doesn't work all the time, but it's surprising how often it does. You never know until you try. You never try until you have to. You never have to until you realize how limited your resources really are.
This is the craigslist logic. Sticky sites can convince themselves they can do no wrong because really they can't. Users would need a really strong reason to leave in droves (friendster). Even spam comments would not lose users quickly unless they were overwhelming because hacker news has a critical mass of people who would have an impossible time moving en masse.
Just because people still use craigslist doesn't mean they are doing it right. A decent mapping ui and post reply management would make the site so much better, users happy, and destroy the competition. Instead Craig says basically what you say here. It's the complacent mediocrity of monopoly.
No offense. I don't even care about search. But I disagree that you shouldn't add features unless the lack of the feature would lose users. You should add features that make users happy.
There's an opportunity cost for every feature added, and it's not just the time lost in implementation. More features = more code, which means more things you have to consider each time you add the next feature. More features = more interface complexity, which means additional things that your users have to consider each time they use the site. More complexity also tends to scare off new users. And more features = more bugs, which is a further drain on productivity in the future.
Didn't we just have this thread yesterday when Google introduced a few new features that everyone seems to hate? ("Changing crap", I believe PG called them.)
Besides, the argument isn't that you shouldn't add features unless the lack of the feature would lose users. It's that you should be careful which features you add so that you can spend your limited resources on the features that truly delight your users. Big companies usually aren't mediocre because they're complacent (well, perhaps except GM). They become mediocre because their userbase is broad enough that every feature they introduce will piss someone off, so they end up with collections of features such that any given person will love 10% of the product and loathe the remaining 90%.
Ok, but honestly, do you think or not that 3 lines of html code should be added, submitting to searchyc.com? I think most of the users here strongly agree.
This community is already pretty positively biased towards YC, note that there is no added value provided by HN technologically, it's just a matter of community. So it would be wise to avoid pushing hard the YC search startup if users don't like it.
Also the "I'm wasting my time" replying to this comment is the wrong attitude. Paul Graham and everybody else is supposed to just reply or not if he got something better to do, but replying this way honestly is not great.
I think that there should be a link to Search, either Google or SearchYC (I'm biased towards Google because I work there, but SearchYC is a good site too). Which there is, now.
I don't think there should be a search box. The canonical position to put a search box is either top center or top right of the page, but there's no room for it there - he'd have to either remove some of the top bar links or shift the whole page down. Shifting the page down is a mistake; part of the appeal of HN is that the focus is on content. And honestly, I find that all the links on the top row are more useful than search would be.
I agree that "I'm wasting my time" is the wrong attitude. He should have said nothing. It's ironic, though, that this sort of uncommunication is exactly what people lambast corporations for. A lot of corporate anti-patterns arise because they're the best response to an impossible tradeoff, not because corporations are stupid.
He's referring to the classic problem in a startup of distinguishing between things users want (vitamins) and those they really want (painkillers). It's critical to concentrate on the latter because you have so little margin for error.
I assume it means that some features are "good for you," while others are opiates; and possibly that this is a bad thing because opiates mask pain, and graphing pain is a crucial part of iterating a design, so the more opiates a design has, the more confounding factors there will be to evolving a good core product. (That's not to say that I agree that on-site search is an opiate, but an argument could be made for it.)
I don't have a better word to describe "painkillers", as he is trying to convey the difference between want and need. I want candy, but it's also not good for me. What word conveys that, and isn't harmful in some way?
Comments
>> I would be suprised if there was a single user who'd left HN because it lacked search.
While I agree with you, I think you're looking at it the wrong way. No user will leave HN because of search-issues, but a lot of users will be very happy if the site had good search.
imo it makes sense to do a trivial thing and make a lot of users happy. So while you're correct in saying that "quality of the submissions and comments" are more important, I see no reason why quality of content should preclude a decent search feature.
I see no reason why quality of content should preclude a decent search feature.
Because attention is a finite resource. I can only spend attention on minor problems by taking it away from major ones.
For example, right now, instead of wasting my time replying to this comment, I should be working on a new project we're launching soon. And I think I will, so forgive me if I drop out of this thread.
Adding a google or searchyc search box is about the same amount of work as typing that comment, and several orders of magnitude more useful.
I had already added a link to Google site search. See the footer.
http://news.ycombinator.com/item?id=1329380
That's not very useful. It takes less time to just go to Google and type it there.
- searchyc usually returns better results
- a search box is much more useful than a link
- it should be in the header, or at least not in a place that nobody ever looks at
But by now it's obvious that that's not going to happen.
But it's what users asked for. How could that be wrong?
(See how complicated such matters actually are? Nothing that goes on the front page of the site is ever a trivial matter.)
The link is an improvement over nothing, at a minimum it will cut down on the number of newbies posting 'how do I search' questions, at least, if they scan the footer.
Experience shows that this is probably not the case but then at least we can point them to the link without having to explain site prefixes, which is still a net gain.
If it were in the header (which is pretty crowded) it would be more useful, but I can see that that might lead to complications, if it were a miniature form in the header it would be better still, but I can see that that would really be much harder to integrate.
Searchyc.com probably can't handle the load if they are the 'sole provider of search' for HN, so it looks like that is not an option, when the HN robots.txt file locked out google because you had to crunch the numbers for the applications it was apparently touch-and-go on searchyc.
Now I'm really curious what your secret startup is that may solve the problem, and I hope they roll out soon.
I am currently part of a start-up, Favetop.com, which offers direct site searching through just the click of a site icon. I've found it to be quite convenient for searching sites directly and much better than constantly adding site prefixes or visiting the actual websites to search. You may want to give it a try and see if it helps you out.
I haven't been here very long but I have never seen a post by PG with no upvotes. This thread has two - something is amiss.
Here's one where PG was actually downvoted into oblivion: http://news.ycombinator.com/item?id=190451
Because attention is a finite resource. I can only spend attention on minor problems by taking it away from major ones.
I think you've just identified one of the best ways to "break out of the box", or dare I say, make a paradigm shift.
The issue isn't whether attention (or anything else) is finite; it's "how to make it infinite", at least for the time being.
Example: I need to add horizontal scrolling (a critical user need) to my Report Writer. I would also like to add F-Key View Toggling (a nice to have) to my Form Look-ups. As a single founder, I am constantly asking myself, "How can I have it all, even with limited resources?" I have found that the trick is not to focus on what's limited, but on how to embrace my constraints and use them to get clever out of necessity. In this case, I wrote a single program that used the same routines to do both.
This doesn't work all the time, but it's surprising how often it does. You never know until you try. You never try until you have to. You never have to until you realize how limited your resources really are.
This is the craigslist logic. Sticky sites can convince themselves they can do no wrong because really they can't. Users would need a really strong reason to leave in droves (friendster). Even spam comments would not lose users quickly unless they were overwhelming because hacker news has a critical mass of people who would have an impossible time moving en masse.
Just because people still use craigslist doesn't mean they are doing it right. A decent mapping ui and post reply management would make the site so much better, users happy, and destroy the competition. Instead Craig says basically what you say here. It's the complacent mediocrity of monopoly.
No offense. I don't even care about search. But I disagree that you shouldn't add features unless the lack of the feature would lose users. You should add features that make users happy.
There's an opportunity cost for every feature added, and it's not just the time lost in implementation. More features = more code, which means more things you have to consider each time you add the next feature. More features = more interface complexity, which means additional things that your users have to consider each time they use the site. More complexity also tends to scare off new users. And more features = more bugs, which is a further drain on productivity in the future.
Didn't we just have this thread yesterday when Google introduced a few new features that everyone seems to hate? ("Changing crap", I believe PG called them.)
Besides, the argument isn't that you shouldn't add features unless the lack of the feature would lose users. It's that you should be careful which features you add so that you can spend your limited resources on the features that truly delight your users. Big companies usually aren't mediocre because they're complacent (well, perhaps except GM). They become mediocre because their userbase is broad enough that every feature they introduce will piss someone off, so they end up with collections of features such that any given person will love 10% of the product and loathe the remaining 90%.
Ok, but honestly, do you think or not that 3 lines of html code should be added, submitting to searchyc.com? I think most of the users here strongly agree.
This community is already pretty positively biased towards YC, note that there is no added value provided by HN technologically, it's just a matter of community. So it would be wise to avoid pushing hard the YC search startup if users don't like it.
Also the "I'm wasting my time" replying to this comment is the wrong attitude. Paul Graham and everybody else is supposed to just reply or not if he got something better to do, but replying this way honestly is not great.
I think that there should be a link to Search, either Google or SearchYC (I'm biased towards Google because I work there, but SearchYC is a good site too). Which there is, now.
I don't think there should be a search box. The canonical position to put a search box is either top center or top right of the page, but there's no room for it there - he'd have to either remove some of the top bar links or shift the whole page down. Shifting the page down is a mistake; part of the appeal of HN is that the focus is on content. And honestly, I find that all the links on the top row are more useful than search would be.
I agree that "I'm wasting my time" is the wrong attitude. He should have said nothing. It's ironic, though, that this sort of uncommunication is exactly what people lambast corporations for. A lot of corporate anti-patterns arise because they're the best response to an impossible tradeoff, not because corporations are stupid.
I think it's the craigslist logic, but I think you're missing the point about finite resources.
He's not saying that features shouldn't be added period, but that those features aren't worth even the minor resources required to add them.
painkillers and vitamins.
> painkillers and vitamins.
I can't make heads or tails of this comment. Could you please explain?
He's referring to the classic problem in a startup of distinguishing between things users want (vitamins) and those they really want (painkillers). It's critical to concentrate on the latter because you have so little margin for error.
I assume it means that some features are "good for you," while others are opiates; and possibly that this is a bad thing because opiates mask pain, and graphing pain is a crucial part of iterating a design, so the more opiates a design has, the more confounding factors there will be to evolving a good core product. (That's not to say that I agree that on-site search is an opiate, but an argument could be made for it.)
EDIT: Oh, wait, I screwed that up completely. It's from here: http://dondodge.typepad.com/the_next_big_thing/2008/03/does-...
Apparently, painkillers are the good things. Goes to show how weak analogies won't actually carry your message.
I could have sworn that it was posted in a wrong thread, but your explanation makes me doubt.
I don't have a better word to describe "painkillers", as he is trying to convey the difference between want and need. I want candy, but it's also not good for me. What word conveys that, and isn't harmful in some way?