<form action="http://searchyc.com/search/yc">
      <input name="search[id]"> 
    </form>

http://news.ycombinator.com/item?id=1329380

- 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.

(See how complicated such matters actually are? Nothing that goes on the front page of the site is ever a trivial matter.)

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 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.

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.

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%.

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 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 not saying that features shouldn't be added period, but that those features aren't worth even the minor resources required to add them.