11–20 of 39 posts
Those kind of links are a great irritation for me when browsing and reading HN. Instead of http://news.ycombinator.com/x?fnid=KnmhPyrF1t Why not for example: http://news.ycombinator.com/news?page=2 ?
http://news.ycombinator.com/item?id=3098756 > why wouldn't they use (searchable) permalinks? I'm assuming there's a reason for this, as its not the easiest way to do things. It's straight forward, but not when you are using closures and continuations.
So... hacker news chose to let their product be limited by an engineering issue?
Those kind of links are a great irritation for me when browsing and reading HN. Instead of http://news.ycombinator.com/x?fnid=KnmhPyrF1t Why not for example: http://news.ycombinator.com/news?page=2 ?
Because things on the front page might have been pushed to page 2 by then, and things on page 2 might have been promoted to page 1, or aged to page 3, etc. The way it works now, all the ordering is fixed when you load the page 1, and page 2 will still be the same page 2 whenever you get around to clicking on it.
I've wondered at times if PG isn't deliberately leaving this problem in place so that people who get annoyed by it are filtered out as users of the web site.
Since you know about the implementation(I haven't looked closely), won't it need a complete overhaul to take closures out of equation? Or implement some sort of closure serialization(on-disk maybe)? Is there a third option where pg just bangs at the keyboard for a couple of hours, and done?
I'm fairly certain HN itself isn't actively developed. If I remember correctly, there was quite a bit of teeth-gnashing over the state of password storage on HN awhile back, and it took the moving of mountains before things were changed (were they ever changed?).
How about some keyboard navigation shortcuts? (left-right arrows). It works on http://www.HnEasy.com .
emacs style editing in the text box
However, if anyone is curious about why it's happening, this is what the arc code is doing http://news.ycombinator.com/item?id=3098863
If you don't follow arc, here is a python and ruby approximation of the arc challenge http://news.ycombinator.com/item?id=3098863
https://github.com/rahulkmr/musings/blob/master/ruby/sinatra...
https://github.com/rahulkmr/musings/blob/master/python/closu...
Those kind of links are a great irritation for me when browsing and reading HN. Instead of http://news.ycombinator.com/x?fnid=KnmhPyrF1t Why not for example: http://news.ycombinator.com/news?page=2 ?
My work around for the "unknown or expired link" bug is to open page 1 of HN, then immediately open pages 2, 3, and 4, but in separate tabs.
That way, I never run afoul of that bug, and opening tabs is really easy in every modern browser.
I'd be interested to know how other people work around the bug.