Hi all,
I check the recommended 'new' questions section of RG on a regular basis. It can mean a lot of scrolling before I get to a question that catches my eye that I feel I can contribute to and that has just been posted recently. To me - 'new' should be chronological i.e. questions asked in the last few hours are the ones that appear first. If a question is more than a few days old, I have most likely made the decision that it is not a question that I could/would answer. My latest tactic is to just scroll down and look how long the question has been posted. It tells you in the top right-hand corner.
On a related matter, I've not been able to work out why quite a few new questions double (sometimes more) as one scrolls down the list?
Hi! No, I do not know either, I have guessed that the order is the one of ”latest changes” among new questions (someone answers, follows or the question itself is edited). But perhaps not. This should be an area where personal preferences could be applied.
I would also very much like a function of dismissing questions I am not at all interested in (they keep coming back and are annoying), and would like an algorithm to learn from that. And next dysfunction: the very unprecise automatic categorization that RG does of a question should be easy to adress with a better algorithm.
And next problem: sometimes an almost private question in a comment elsewhere or so is emerging as a question. Recent example: Are you still in this project? (Meaningless out of the wider context, and as difficult to answer yes or no to as the old Greek sophistic question: ”Have you stopped beating your wife?”
The question section needs a lot of work, but I like it.
Hi Anders - agreed. I like the question function - but it does get abused and it would be good to be able to 'vet' it somehow. The algorithm suggestion sounds appealing. Also - now I think about it, often questions are asked that have been asked many times before - so a feature that allowed those posing a new question as to if something similar matches i.e. through keywords or matching to the follower groups - might be useful. Certain subjects crop up all the time i.e. where can I get published fast, at no cost and/or plagiarism is another popular one. A note or caution could be matched to that fact before posting.
I was guessing that a reason for the odd chronology sequence might be to get older questions in first (that no-one or few have responded to) to try an re-invigorate them. However, the chronology seems too 'random' for it to be that i.e. one question is 5-days old, then the next one 1-day, then the next 3-day - and then back to say 5-day again.
Hi Dean Whitehead ! I agree.
A general problem with RG that shows up regularly is this with lacking transparency of how their processing of information works.
A very reasonable demand should be, for all these kinds of articcial intelligence functions, that they should be possible to access and analyze. The function of an algorithm may not be very transparent for anyone, but the information should be in the open and open to scrutiny. It should not be hidden what a web site does with the information from their users before re-presenting this information to them. R G has become better, but is not there yet.
- Badges
- Science topic
- Similar topics
- Papers