I've found SearchYC to be very helpful. I use it quite a bit to look up old information I only vaguely remember or to find people. Glad to see it's only going to get better - congrats Mike and Jerry.
I always love hearing about behind the scenes updates, thanks for posting.
As a more dev-heavy optimization, it might be possible to return cached results (user-search:PG), and then trigger a backend server to refresh the cache, possibly even using ajax to inject the new results.
Thanks for the site, it's my top choice for searching HN.
As an aside, I'd love to see a standard for integrating site-specific search into a larger search engine. Imagine if you could go to Google, and type in
"Site:News.ycombinator.com user:e1ven" and it would send those as post-requests to the search engine designated by robots.txt on the host?
That would make search dramatically "deeper", by letting google act as a front-end to each site's results.
There's the "OpenSearch" protocol, which just tells your browser "to search this site go to [someurl]?search={SearchTerms}" SearchYC could certainly add this, though I think it may be limited to 1 search only, and not something like: "I have two types of searches, one for everything, and and another for users."
But beyond that, OpenSearch it's pretty limited. I expect this area of semi-semantic searching to explode, which is why I'm investing my own time into it.
So, hang tight for a couple of weeks, then the problem will be solved :)
That's strange. We made changes to DNS a few days ago. I suppose it's possible that if you're on the very outskirts of the internet, that it hasn't completely propagated yet.
Google gives more comprehensive results too. But there is a time lag with either. I was mining my bookmarks tonight to find links worth posting, and searched to see if one had been posted before. Search YC indicated it had not been, so I posted it, only to discover to my chagrin that someone else had just submitted it just before I did. (My submission is now deleted, to avoid duplication.) I try to search before I submit, so accurate and timely search is helpful.
I haven't done a lot of submitting myself.. I usually just do "Ask HN" type stuff. Hacker News doesn't automatically tell you that the URL you're submitting has already been submitted? Or was the URL slightly different?
HN automatically upvotes the previous post if you submit an EXACT duplicate URL. But some webpages live under more than one URL, and not everyone submits the canonical URL.
Notifications was one of the original motivations for creating SearchYC, but there are a couple of fundamental flaws:
1) How would we know if you've already read a comment/submission? It would get spammy really quick unless we did some kind of periodic digest. We decided that creating RSS feeds for search terms (e.g. username) is a happy medium for this, but we're open to suggestions.
2) New items can take quite a bit of time to get indexed (usually 0-2 hours depending on the number of new items and position in the queue).
25 comments
[ 3.0 ms ] story [ 77.0 ms ] threadAs a more dev-heavy optimization, it might be possible to return cached results (user-search:PG), and then trigger a backend server to refresh the cache, possibly even using ajax to inject the new results.
Thanks for the site, it's my top choice for searching HN.
As an aside, I'd love to see a standard for integrating site-specific search into a larger search engine. Imagine if you could go to Google, and type in
"Site:News.ycombinator.com user:e1ven" and it would send those as post-requests to the search engine designated by robots.txt on the host?
That would make search dramatically "deeper", by letting google act as a front-end to each site's results.
-CPD
But beyond that, OpenSearch it's pretty limited. I expect this area of semi-semantic searching to explode, which is why I'm investing my own time into it.
So, hang tight for a couple of weeks, then the problem will be solved :)
Searchyc is a lot more useful, and you don't need to install anything to use it.
It's in the cloud of names.
edit.
thanks to the people that tested it, ok so that seems to be my problem then, here is a traceroute of the last steps before it stops responding:
11 * pos-1-4-0-0-cr01.mclean.va.ibone.comcast.net (68.86.86.29) 251.887 ms 252.502 ms
12 pos-1-11-0-0-cr01.atlanta.ga.ibone.comcast.net (68.86.85.241) 267.698 ms 268.398 ms 234.520 ms
13 pos-1-13-0-0-cr01.dallas.tx.ibone.comcast.net (68.86.85.253) 227.691 ms * 153.851 ms
14 pos-0-13-0-0-cr01.losangeles.ca.ibone.comcast.net (68.86.85.145) 202.909 ms 204.094 ms 204.818 ms
15 * pos-0-14-0-0-cr01.sacramento.ca.ibone.comcast.net (68.86.85.189) 215.876 ms 216.499 ms
16 pos-0-15-0-0-ar01.oakland.ca.sfba.comcast.net (68.86.90.138) 219.741 ms 220.384 ms *
17 po-30-ur02.santaclara.ca.sfba.comcast.net (68.87.226.222) 222.161 ms 223.248 ms 222.742 ms
18 te-8-4-ur04.santaclara.ca.sfba.comcast.net (68.87.226.85) 224.225 ms 224.801 ms 225.416 ms
Good progress, guys!
up for me.
jam@jam:~$ nslookup searchyc.com
Server: 192.168.1.254
Address: 192.168.1.254#53
Non-authoritative answer:
Name: searchyc.com
Address: 24.7.119.0
isn't a .0 a network address ?
http://www.google.com/search?q=site:news.ycombinator.com+sea...
Send a notification to my email __________ whenever:
(checkbox) Someone replies to a comment from user _________
(checkbox) User _________ makes a submission/comment
(submit button)
1) How would we know if you've already read a comment/submission? It would get spammy really quick unless we did some kind of periodic digest. We decided that creating RSS feeds for search terms (e.g. username) is a happy medium for this, but we're open to suggestions.
2) New items can take quite a bit of time to get indexed (usually 0-2 hours depending on the number of new items and position in the queue).