Please don't break the spacebar key acting as page down. The search text box takes all the key presses regardless of focus. I find this annoying in a 'breaks expectations' kind of way.

fixed

thanks!

I'd like a RSS for search result (want to subscribe to a topic and get a notification whenever it's mentionned here)

If its Solr or Lucene running underneath then you could use phonetic analyzer to get expected results if the spelling is not correct. I got results with'Steve Jobs'but when I tried 'Steve Jbs' I got no results. I would say getting close results is better than no results.

If there are only a few options for a variable, could you have them be expressed as part of the object returned by http://api.thriftdb.com/api.hnsearch.com/items?pretty_print=... Example: Type. Although it is a string, it isn't really a string field. It is a binary choice between two preset strings (comment or submission).

It would make using the schema for things much easier.

I'm not sure why, but I've noticed that sometimes the points field is null, even though it shouldn't be. See for example the second result here:

http://api.thriftdb.com/api.hnsearch.com/items/_search?q=sky...

I think points is null because that item is < 5 days old. The API should only return points for items > 5 days old.

why this restriction?

Looks like items marked [dead] in HN will show up as zombie search results -- i.e., they get returned in the search results for the items collection but the discussion field is null.

For example: filter[fields][type]=comment&filter[fields][username]=pg

...the second result there has a null discussion field. Getting the submission title &etc. would be nice.

The 'discussion' attribute contains info about the discussion that the comment belongs to. In this case the parent discussion is [dead] so the value of 'discussion' is null. Unfortunately, we don't have access to submission titles for [dead] submissions.

Is there a way to exclude values from search results in the ThriftDB API? i.e., say I want to,

    [filter][domain]=nytimes.com&[filter][username]!=ausername

...is there a way to do that? I tried the "exclude" operator after [username] and get an "internal server error" every time.

&filter[fields][domain]=nytimes.com&filter[queries][]=-username:ausername

Great! Thanks.

Argh. And also, is there any way to query ThriftDB for HN date ranges?

You can use strings to represent relative times:

&filter[fields][create_ts]=[NOW-5HOURS%20TO%20NOW]

or exact timestamps:

&filter[fields][create_ts]=[2011-06-01T05:00:00.00Z%20TO%202011-06-02T07:00:00.00Z]

or both:

&filter[fields][create_ts]=[2011-06-01T05:00:00.00Z%20TO%20NOW]

Sweet! That's perfect. Thank you.

I'm not sure US users have noticed, but I can't use the service from Europe, Hungary. It just keeps connecting to the api, tested with latest Chrome and FF.

Can you access ThriftDB (http://www.thriftdb.com)?

It's minor but sticking to conventional operators would be helpful. For example, I find myself typing "intitle:" rather than "title:" out of habit.

I just added support for the "intitle:" operator.

It would be really great to have the ability to sort by hotness (i.e. get the top K hottest results (for a given search or not)).

the default relevance algorithm uses hotness to sort results

it would be nice to see the next "start" point based off your limit / total hits this would make implementing pagination simple!

Search only titles, then order by points/date.

You can use the "title:" operator to search in titles.

Work with Opera?

Are you using 11.00? Other users have had problems accessing javascript-heavy sites with 11.00:

http://superuser.com/questions/234110/websites-not-working-w...