Ask HN: What do you think about our last HN Search update?

1. Is there a way to force off the typo correction? I tried "smartos" both with and without quotes, and either way I get results for the word "smartest", which aren't about SmartOS (an operating system derived from OpenSolaris).

2. In terms of CSS/font/layout, honestly I like hnsearch.com's results better. I think part of it is that I use hnsearch.com as an alternative HN interface, not just a search engine, and it's usable for that purpose, in part because it looks more like regular HN.

The issue for me more generally is that hnsearch.com is almost perfect as an HN search engine, as far as I'm concerned. It does what I want, does it fast, with good coverage and a usable interface. So my advice for alternatives would tend towards just "yeah, make it more like that", which is maybe not the most useful commentary.

Btw we have just put in production a new version with sort by date :)

If I understand correctly, it is for enterprise search, not for web search, correct?

The response time is really impressive, especially with the new sorting by date function. To me, using NoSQL database, it's hard to do sorting if it's not impossible. That's why it's not available in the initial current release.

This implementation reminded me about the CTF algorithm, which needs to match the input query against a file. The reason why I thought it's not for large volumes of queries is because:

1. Each keystroke is a AJAX call which could be lightening fast when the query volume is low. But JavaScript could run slow on mobile phones. In this case probably it's not too bad because the AJAX call is very simple, fetch the results from the server

2. After receiving the request, on server side, there are only three steps to go before the results are returned:

- Initialized a new index (actually it's not a new index, just a new search)

- set criteria for the order of the attributes sets

- Call search and return

Of course, there are more detailed steps under the search api. The majority of the work is done on the search engine server side to keep crawling and updating the indexes.

When the request volume becomes huge, the real time response may be slow down due to the file size.

I appreciate this tool to help me find things quickly on HN.