Otherwise how does a user know that the URL they went to is unknown or expired?
I agree 404 etc pages /should/ be better designed to inform the user then give them links/options to carry on their journey.
Here: http://www.bbc.co.uk/cbbc/404.shtml - we tried to explain a little about the problem, to a young audience and then offer them two ways of continuing their journey - go to homepage or use search.
(The poster doesn't mention it but I'm sure he's talking about the Hacker News pagination links- if you dig a page or two deep and get caught up in an article for a few minutes and then try to click to the next page again, you get an ugly, worthless error message about the link being expired.)
or better yet, just make it so that HN's pagination links don't expire.
I don't think there should be a redirect. That's unexpected behavior. It wouldn't be bad to have the "Unknown or expired link" actually link to the home page--or better, have all or part of the navigation header on the page as an escape hatch.
Silently ignoring errors that are require user interaction and serving the user something that is a valid response to a different action is terrible for ux.
Say I take too long entering a comment, I get distracted at work, then hit submit. I'm taken to the homepage. Did the comment go though nor not? How an I supposed to know what the hell happened?
Honestly, of all of the sites I frequent, Hacker News has the worst user experience. This is how I use the site:
1. Go to page 1
2. Read a couple of articles on page 1
3. Hit next page and receive error.
4. Go back to page 1
5. Go to page 2
6. Read a couple of articles from page 2
7. Go to next page and receive error.
8. Go to page 1
9. Go to next page
10. Go to next page
11. Read article from page 3
...
I would be curious to know what percentage of requests to HN return "Unknown or expired link". If I were in control of this website, I would consider this behavior to be unacceptable.
I never really understood why pagination links expire. Most sites don't expire, you just have to deal with knowing that content that WAS on page 3 when you loaded page 2 might have changed.
I'm kind of annoyed by it as well but I've learned to click on the top left logo every once in a while and start over from page 1.
So it seems like there's some real UX concerns expressed below re: auto redirect to homepage w/o sufficient error messaging. One solution would be to render the error message on the top of the homepage "Sorry, articles expired like milk so we brought you back to the homepage for fresh articles.".
I would actually prefer an infinite list of stories with stories moving up and down in a clean AJAX'ed way based on live votes so at any given time you are seeing the most up to date list of articles.
Oh yes, this is about the only thing that annoys me about HN. I love the clean, simple interface. But the expiring continuations are... annoying. Like some enterprise systems that log you out after a few minutes of being idle (luckily, you don't have to reenter your password in this case :-)
Would there be much overhead from increasing the expiry time to 3 hours or so? After 3 hours I wouldn't care about pagination, but it seems to be less than an hour currently.
The way I overcome this problem is by using ReadItLater.
1) Open Page 1.
2) Open new tabs of all the links I want to read in that page as a new tab in the background (I use Opera). This leaves me still in Page 1 while the other tabs open in the background
3) Repeat process for next few pages
4) Finally, go to each tab and click on the ReadItLater Add bookmarklet.
5) Close all tabs. (I don't close the tabs separately but move on the next tab as I find it more efficient that way)
11 comments
[ 4.2 ms ] story [ 25.3 ms ] threadI agree 404 etc pages /should/ be better designed to inform the user then give them links/options to carry on their journey.
Here: http://www.bbc.co.uk/cbbc/404.shtml - we tried to explain a little about the problem, to a young audience and then offer them two ways of continuing their journey - go to homepage or use search.
We also made other URLs fail gracefully. If a user was at: http://www.bbc.co.uk/cbbc/thingstodo/by/type/printgames and then 'hacked' the URL to get http://www.bbc.co.uk/cbbc/thingstodo/by/type - we still offered them a page that correctly renders and has the information they were looking for - though we'd never link to this page ourselves (the correct root here would be http://www.bbc.co.uk/cbbc/thingstodo )
or better yet, just make it so that HN's pagination links don't expire.
Say I take too long entering a comment, I get distracted at work, then hit submit. I'm taken to the homepage. Did the comment go though nor not? How an I supposed to know what the hell happened?
1. Go to page 1
2. Read a couple of articles on page 1
3. Hit next page and receive error.
4. Go back to page 1
5. Go to page 2
6. Read a couple of articles from page 2
7. Go to next page and receive error.
8. Go to page 1
9. Go to next page
10. Go to next page
11. Read article from page 3
...
I would be curious to know what percentage of requests to HN return "Unknown or expired link". If I were in control of this website, I would consider this behavior to be unacceptable.
So it seems like there's some real UX concerns expressed below re: auto redirect to homepage w/o sufficient error messaging. One solution would be to render the error message on the top of the homepage "Sorry, articles expired like milk so we brought you back to the homepage for fresh articles.".
I would actually prefer an infinite list of stories with stories moving up and down in a clean AJAX'ed way based on live votes so at any given time you are seeing the most up to date list of articles.