yes I've been confused about the odd way HN handles pagination. Its retarded at the most basic level that hitting Next gives an error page because some code went stale.
It is in fact a simple fix. You just add timeouts. I already did it for the application form.
The real answer to the question is: because I have other things to do besides work on cosmetic bug fixes. I'll get to it at some point. In the meantime, please stop submitting "stories" about this. If you discover an important bug, send me an email. Otherwise the place for feature requests and minor bug reports is the feature request page; there is a link to it at the bottom of the front page.
I take it you know that the continuation timeout problem affects replies as well?
Because I think it sucks that when I spend more than a few minutes composing a reply I can't post it. On more than one occasion I lost my reply this way.
So it's not a cosmetic problem. We're not talking about the minor nuisance of hitting refresh, we're talking about losing carefully crafted replies.
I took the habit of always copying my text to the clipboard before submitting anything substantial because between session timeouts, expired links and other failures there's just too many opportunities for your contribution to land in the bit bucket.
If you click on the back button, the text should still be there in the form you submitted. Surely your brower doesn't wipe forms when you click on back?
You'd be surprised, that has always been my instinctual reaction to those kinds of failures but oftentimes the textarea would just come back blank for some reason or something like that. I think it had to do with the "you can't go back because the page was generated from a POST operation" kind of message.
Clicking back is not enough. You also have to refresh the page to get a new continuation.
On IE the text is gone after you hit refresh. On firefox the text stays when you hit refresh. Since I'm used to the behavior of firefox...
Printing the http-post data on the error page would also help. Because the moment I realize my post is gone I can still hit "forward" to get back to the error message. It's not a perfect solution, but since it only takes 15 seconds to implement it's a fair compromise.
I understand that you're busy, and I don't mean to discount that. I'm sure it takes lots of time to write those essays -- as I thoroughly enjoy how thoughtful and articulate they are. Also you've got your YC companies to look after, and lots of (I'm sure) exciting events and engagements to go to.
Hear me out. I love Hacker News, and I read it every day. Sometimes, though, I get frustrated...
I finish reading a fascinating story about hacking or startup's and, to continue in my adventure I hit the "more" button at the bottom of the site.
I'm expecting to be sent to the next page so I can continue scrolling through the news and stories from the site, and instead of seeing the next 20 or so news items, I see this, "Unknown or expired link" (see http://news.ycombinator.com/x?fnid=qQ88X1AyDg).
Then I leave the site, even though I'd rather keep reading.
It'd be nice to have this fixed. That's all I'm saying. Thanks.
The easiest fix would be to get rid of the continuations. It would be just as easy to encode the comment_id in a form field as millions of other webpages do.
Ok so I looked at the source and found:
(def process-comment (user parent text ip whence)...
I don't see why it would be hard to submit "parent" and "text" as POST-parameters. User seems to be stored in a cookie anyway and ip should also be easily retrievable.
