Please tell us what features you'd like in news.ycombinator

Reason being that I like to use live bookmarks in Firefox, and the RSS feed which goes directly to the submitted link means if I want to read the comments left by HN users, I must go to HN's homepage to look for the article, and then click on comments.

Which is a lot of work compared to clicking on the live bookmark entry, then opening up the submitted webpage (in a new tab) from the submission's comment page.

I am aware that occasionally there are great items that don't get a few upvotes quickly, and then get lost in the flurry of other submissions. Submitting something at a popular time is a lottery, and I think many useful items get lost, or go unnoticed.

I'd like to see an alternate ranking system based on votes, replies and page views. Let R be the current score as determined by votes, C be the accumulated votes on the comments, and V be the number of views of the page. Then let the ranking score be something like (R+C)/V. The idea is that pages without views will stay close to the top, encouraging them to be viewed and hence ranked. Pages that elicit no upvotes will then drop quickly, but at least they've been seen.

There probably needs to be a time component in there as well so that items slowly "fade" with time.

I can expand and refine this for anyone interested in seeing waht happens, but I can't produce a mock-up because I don't have access to the "page views" statistic.

Second thing, and much less of a priority, I'd like is the ability to retrieve a single item with its threading information, but not its threading content. My interests aren't entirely aligned with the majority, so I'd like to retrieve every item and then read them in thread for myself. Currently if I pull a given item I get all its sub-comments as well, which I then have to unpick. It's tedious, and I haven't bothered yet, but I can do it. It would be more value to me if I could just pull the text and ID of the parent.

Question: I'm interested to see if this comment gets read. This thread is now 2 1/2 years old. How many people read it, and what path to they take to do so?

Finally, the comment. Thank you for HN. I think it's a fantastic resource, and I look forward to contributing to it for some time to come. I hope I add value.

I know that we have to be careful how much we tax the servers for HN, but I thought of a simplistic and low-processing overhead for allowing users to block certain domains.

1) Allow us to input a list of domains we don't want to see links to.

Ex: csmonitor.com, godaddy.com, codinghorror.com etc.

2) Modify the item listings on the various listing pages to include a class that is the url's domain.

Ex: So if the link's url is http://x.com/asdfasdf the link listing would be class 'x.com'

3) Implement a basic css command generated on page load that adds our domains listed in part 1 and adds a 'display: none;' for them.

4) Win!

Obviously it's imperfect, you'll get less than the max items on a page, and users will notice numbers missing for the hidden links.