Any reason why the old domain couldn't continue to be a wrapper around the new engine/interface? hnsearch.com sounds more trustworthy to me (and has been around a lot longer) than algolia. (This is coming from a long-time lurker, never been into the YC scene much, so I've never heard of algolia).
Yeah, the API is different (but I feel it is overall better and they've been responsive to inquiries on github https://github.com/algolia/hn-search).
Technically, you likely would have had to change your code a month ago or so. From what I can tell, hnsearch hasn't had its content updated in the past month.
Edit: Fixed URL. Apparently if you include the period inside the parentheses right after a link (which is incorrect if it isn't its own sentence), Hacker News includes it as a part of the URL.
I highly doubt they are "branching out" from just HN. Obviously Algolia existed prior to replacing hnsearch (they were already in YC), but taking it over gives them more visibility within the startup community and also frees up responsibility for Octopart.
Yes, we already existed before operating the new HN search. We did it as a way to participate to the community and of course it's good for our visibility. The hnsearch.com domain is the property of Octopart, not ours.
Can someone outline the benefits of the new one over the old one? When I first tried the new one, the UI was severely lacking. I saw the fixed a few things, but I haven't evaluated it again.
I don't always use the HN search engine, but when I do, it's usually very helpful. I'd hate to lose that.
Not exactly, tptacek is right; there is no simple way to find all items containing "patio11" not being written by "patio11" (both the text and the author name are indexed). Handling the "NOT" operator is pretty tricky as soon as you want to retrieve relevant results from the first keystroke in a few milliseconds (your query can match all documents).
The internal API is presumably for YC companies/alum who arrange special access (HNsearch was powered by ThriftDB, another YC alum). Why not just use Algolia's API? http://hn.algolia.com/api
(Even before HNsearch, there was another YC company that served as the official search for HN for a short period. Can't remember the name now though.)
1 point for you :) That being said, we're all very open to any discussion about a missing feature. We've already added a few ones and have even open-sourced the code: we're definitely motivated to provide the best HN search engine to the community. Cheers!
The new search doesn't appear to present any filtering or sorting options, at least on mobile. This surely can't be right. The old HNSearch with its quick toggles for submission/comment and date/points was extremely effective.
I hope the new search engine will eventually get the same index as the old one
I was searching for some articles where the old search would get me lots of results that the new search would not return. So I'm hoping that's an indexing issue rather than a quality of the new search engine issue
The search results for stories show both the long (and useless) news.ycombinator URL and a link to the comments. Both links lead to the same location. The news.ycombinator URL displayed is redundant and adds visual noise. I've highlighted both links in the image linked to below.
I'd like it if the results showed "link", "parent", and "on $News_Story", like the old site did.
I wish the Algolia HN Search had a better coverage of older submissions and replies. Finding stuff that's older than 3 years is pretty hard on the new search.
60 comments
[ 3.2 ms ] story [ 109 ms ] threadTechnically, you likely would have had to change your code a month ago or so. From what I can tell, hnsearch hasn't had its content updated in the past month.
Edit: Fixed URL. Apparently if you include the period inside the parentheses right after a link (which is incorrect if it isn't its own sentence), Hacker News includes it as a part of the URL.
Seems like it might be something where they want to branch out from just HN.
Good:
1. The number of points and age of the stories are highlighted.
2. Thumbnails are a nice addition.
3. It looks less cluttered than the old search.
Bad:
1. Very dull font color for most things except the story title.
2. I had to enable two new domains in NoScript :P
I don't always use the HN search engine, but when I do, it's usually very helpful. I'd hate to lose that.
One nit: IIRC, the old engine supported vanity searches - you could search author:-patio11 to find comments mentioning patio11 not written by patio11.
(Even before HNsearch, there was another YC company that served as the official search for HN for a short period. Can't remember the name now though.)
As an executive of Agolia, you have just wasted a learning opportunity.
FWIW, I liked HN search because of it's sorting ability. By date, points and how it combined with the filter comment only, story and heading.
Cheers!
I personally hate that, is the same road google has taken the last months.
I was searching for some articles where the old search would get me lots of results that the new search would not return. So I'm hoping that's an indexing issue rather than a quality of the new search engine issue
http://hn.algolia.com/#!/all/forever/prefix/0/tptacek%20nego...
I'd like it if the results showed "link", "parent", and "on $News_Story", like the old site did.
http://imgur.com/asHzSb1
Only a minor inconvenience, but now I have to use several separate bookmarks to search for every keyword I am interested in.