Hacker News Instant | Hacker News Reader

Hacker News Instant

hninstant.org

105 pointsbrudolph59 comments

Open articleView on HN

I'm afraid I really can't get past the orange. I think I literally felt my eyes relax when I pressed the back button.

I checked the comments first and thought you were kidding. Jesus.

Same. It made me shut my eyes instantly even if I was prepared for the worst...

It burns! Very orange :-)

My eyes are bleeding. Give me some contrast on the text also. It's like trying to read something written on the sun!

Feels very quick and snappy. Could you comment on how you did it?

BTW, the "full orange" strains my eye. Maybe it's just me, but it would be nice if you could have softer colours!

It strains my eyes as well. The reverse (darkorange text on a white background) would be better. Maybe try that?

Sure, I'm going to be writing a longer blog post on how I made it, but for now here's a short summary:

I made a script that scrapes all the links from hacker news every 15 minutes. I then open the links and process the text using python's nltk package (deciding what words are important and useful). Then I used a suffix tree in a mongodb backend to store the important words in such a way that once it looks up a word you can get the set of documents pertaining to the word. This way the search is linear in the length of the query and not the number of documents. The rest was just some jquery ajax calls and parsing of the search query.

I'll look into a new design, maybe make the orange's white and the white's orange.

Great and snappy! I would really love the longer blog post...! Two questions if you have a minute: 1) Why suffix trees and not suffix arrays? 2) How are you implementing them? Did you do the tree building yourself or is there a good library that you recommend? Thanks.

I used a suffix tree over a suffix array because I hadn't heard of suffix arrays, but after glancing at the wikipedia page for suffix arrays it seems those might have been a good choice too. I'll look more into it. I did all the tree building myself, and I'll explain that in my post. The post should be ready by tomorrow.

Awesome! Thank you so much...

From how many days have I been procrastinating doing exactly this :(. Now here it's already here! congrats :)

For this type of thing I've had much better results with LDA in the python package gensim. It is less prone to mis-matches based on similar keywords (since it is context based) the problem with LDA is that for it to be most effective you need to have a taxonomy available for the documents, but you might be able to build a corpus or two out of sites like stack overflow.

http://pennyhacks.com/2011/11/09/introducing-hninstant/

Here you go

Great! Thanks!

BTW, it's not just "British people" who write "colour"; the people can also be from British colonies. ;)

Not to thread hijack but i made instantise.com forever ago and it got on HN, it also has an HN instant.

http://news.ycombinator.com/item?id=1683584

I get error on the homepage as there is google embedded in an iframe which, as my browser says, is not ok. If I switch to other site it's ok.

Very nice. Do the search results link to the comments pages anywhere? It looks like there are only links directly to the original articles.

I too would like to see a comments link, because without it this is useless to me, which is a shame, as it's otherwise very nice.

Thanks, currently it doesn't link to any comment pages, but hopefully in the future I'll add it

Is it just me, or does this not work at all? I'm on the latest version of Chrome for Windows.

Also did you create two accounts right when you posted this to comment on it and say it's awesome?

EDIT: I figured out why it seemed like it didn't work at all:

http://i.imgur.com/CmsZG.png

Is that what the green colored usernames mean?

Yeah, I guess green usernames indicate new (green) users, because konakid and cal8food were both created about the same time this was posted.

haha no, just my friends in my dorm watching me launch it

same here it's not working for me either. Same effect browser/OS - Firefox 6.0 windows XP.

As long as you were "touching up" the photo, you might have wanted to hide the bookmark tab titled "Piracy"

Always 'peel' away the tab to its own window before taking a screenshot. And, uncheck "Show bookmarks bar" ;)

Eh, whatever. I was mostly just trying to cover up any personally identifiable information.

So the guy researches acts of seafaring robbery and violence. What of it?

Does anyone knows what is the stance of Y combinator with HN scrapping?

http://news.ycombinator.com/robots.txt

Some of the searching seems off. Searching "bill" gets me headlines with "Bill Gates" in them, but searching "Bill Gates" gets me no results.

Working on that as well, it's the question mark after gates that is throwing it off.

Nice. All it needs now is a Hacker News theme. I'm thinking small black text on light grey... ;-)

For me, the only thing missing is the link for the discussion. But its a nice service, congrats!

Awesome. Love it. Suggestion as many has pointed to swap the Orange and White color. And one request, can you put HN link instead of direct article link as i want to view the discussion on that article too.

Cool, thanks, I'll put it in for the next version

I like it! Clean interface and quick search. The top articles feature is handy. Maybe make the search options more customizable though? Being able to tweak the parameters of the search would be awesome

This makes me wish there was a top articles on HN proper, of all the top articles ever.

Looks good, and quite responsive.

I noticed that the more links are in there also. You can find it by searching for 'analytics' then click on 'more' button. you'll notice a search result that links to "/news2" :)

Ah good catch!

My #1 use of hnsearch.com (and before it, searchyc.com) has been to find my own old comments. This has no results for my username, so it can't yet fill that role.

Didn't realize this was such a huge demand, I was always more interested in the content, but I'm definitely going to add the comments link

Doesn't seem to handle Japanese. I'm having the same issue with an instant search I've been developing for another purpose.

Yeah I didn't handle any cases for other languages since most of the articles are in english. But I feel ya on making search engines for multiple languages, had to do that this summer and it was a nightmare

The problem's not Japanese-specific. Try any accented European letters for the same problem.

Result set is quite poor. Some results are missing completely, other results aren't even related to the search term.

As much as I like orange, it's not meant to be used in such quantities in screens...

No workie on my iPhone (also has no mobile specific layout)

Looks like it only searches articles < 1 week old.

Yeah it'll grow as time goes on. Once week ago was when I was happy with my db structure and stopped dropping all the contents in it.

Not working, when the browser is resized. Chrome 15.

Hey this is really cool man! Great work. I love it.

Thanks for the feedback! As time goes on its article database will grow and it'll be more useful

Isn't http://hckrnews.com/ a much better effort than this?

I, for one, welcome our new orange overlords.

add keyboard shortcuts, otherwise the 'instant' becomes kind of pointless.

i am hungover and that orange is making my eyes bleed, SERIOUSLY

Reply on news.ycombinator.com