h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

This idea would probably mean a major modification to the DNA system, but is it possible to show the number of views that a particular conversation has had, as well as the number of replies it has received?

I have seen this on many other forums on the Internet. The site would simply need to count the views based on when somebody clicks the link to the conversation. This would, of course, mean that a page dedicated to the 'most viewed' conversations on the site could be made.

Also, many other Internet forums have the facility for people to post custom 'polls', where the people reading the conversation have to select a radio button according to their view and can then see the results as a percentage/bar graph. It might even be useful to have such a poll at the bottom of every guide entry, where you click a radio button to vote on how good you think the entry is out of 10. Or you could give it a grade from A - F or something in the same way.

And that idea would mean a Hall of Fame for the best rated entries and the researchers who have the best grades.

A further idea of mine is using the categorisation system of the guide, linked to the grades a person has. For example, if I were to write a number of entries on engineering and receive high grades for them all, the website would calculate my rank for the engineering topic against everybody else who has written engineering articles. So as well as an overarching Hall of Fame we could have topical Halls of Fame too.

I realise that some of these ideas could bring a competitive edge to the community that some people may object to, but, if the ideas haven't already been suggested, I think they are worthy of consideration.

Thank you.

It sounds like you are more used to the sort of messageboards that have all those features. If all sites were the same, what would be the point?

The best rated entries are called the Edited Guide. And in some cases, the Underguide.

That's not exactly true. Some really bad entries are in the Edited Guide and they stay that way despite being "rated" as bad by people who know better. The pages rated the best by users in terms of visiting them a lot are probably various clubs and <./>Askh2g2</.>. Long active conversations might also be regarded as receiving some sort of popular vote. However, there isn't a way to distinguish ones with many people "voting" by contributing from ones with just a couple of people involved in a chat.

The closest thing to an automatic rating system is the <./>search</.> facility. It deliberately gives preference to the EG but at least it shows a vague idea of how relevant an article is to the keywords supplied. That's the thing which other people have mistaken for their "grade".

There have been ratings and polls and are again some. It is just that on h2g2 they tend to be run informally whereas on other dna sites the software enables them to be managed automatically (ie on onthefuture and ican).

I see what you mean, but if you ignore the competitive-type ideas, the 'number of views' is likely to be useful isn't it? Then we could see how many people had read our guide entries, and it might even be able to count people who read it who aren't registered on h2g2. And if we don't appear to be getting responses to conversations started on ask h2g2 or other places, at least we would know whether anybody had actually looked at it.

When we first launched, we tracked page views on articles, and the front page had the top five most popular entries. trouble is, being marked as the most popular on the front page simply meant they were *always* the most popular.

There is a voting system built in to DNA now, for iCan, which could be made into a general purpose voting mechanism. There's also some interest in using rating systems to rank content.

The trouble with any rating system is that it opens up channels of misuse. For example, some of the votes on iCan have been 'stuffed' by people voting thousands of times. Having open voting on articles or comments inevitably leads to people finding ways to 'game' the system. That's why Peer Review has remained the (mostly) manual system it is for such a long time. We could easily have added voting in there, for people to vote whether something should be selected or not. But that makes it too easy for people to simply vote things up or down without good reason (just because they happened to like or dislike the writer, or agree or disagree with the premise of the article). At least with PR the quality threshold is entirely under human control. It might not scale to ten or a hundred times the throughput, but we don't need that at the moment, so people are a better solution.

Having said all that, I think it's almost inevitable that some sort of rating system will eventually be introduced on other DNA sites. I think we'd have to wait and see if such a system can produce useful results, or whether it will just be another reason for people to argue with each other, or another method by which badly behaved users can annoy other people. Time will tell. And such a system wouldn't necessarily appear on h2g2 - that would be up to the staff.

Thank you for clarifying that. Although surely the issue of some people voting multiple times can be resolved with a system that recognises a repeat of an IP address?

IP address isn't enough of a differentiator. Most people don't have fixed IP, and users of big ISPs or in large offices will all appear to have the same IP address. There are ways to cut down on vote stuffing, but none that are foolproof. Our own voting system makes an effort to block multiple votes, but there's a limit to what can be done.

I see. Well, thanks again.

The iCan votes seem to be fairly well used, and the fact that it can distinguish (and display) the signed-in users makes it less open to abuse. At the end of the day, there's not a lot to be gained from voting for your own campaign a billion times...

I would like to see the ability to change (or remove) your vote though.

For the record, on the Book of the Future site the voting wasn't handled by DNA but by a seperate BBCi-wide voting system. It wasn't automatic and each vote had to be set-up and added to the article manually (I did tons of these). Fairly time-consuming and didn't have that much protection against abuse (used a cookie). I also managed to 'break' the system by adding so many votes that the hard disk on the server filled up, prompting a swift e-mail to the tech support people... So it's good to see the vote thing integrated into DNA...

How about letting the general public - who are not h2g2 researchers - to give their feedback on guide entries, even if it's just a rating?

This is the problem with the votes on iCan, once signed in you can only vote once, however the 'public' (a non-signed in user) could vote as often as they liked.

If we were to have voting I think it should be for signed in users only. One vote each which can be changed or removed at any time.

Saying that there will always be someone who will create a dozen user accounts to vote with

Mel

What does a vote mean? I've always been against rating on h2g2 on the grounds that, if it's not right, then people should say why. And also, if you don't sign up, I believe that there ought to be no interactivity - signup is free, doesn't bombard you with spam, and opens up the community to you. If you don't want to contribute, you shouldn't have a say. Simple as that.

Whoami?

Indeed. Well since the idea of polls is not generally thought high of, it is probably safe to say that it will never happen.

Did we come to any conclusion about whether a count of the number of views entries/forums get should be implemented as a feature?

I would imagine that the views thing is not high on the features list. For a start would you count visitors or visits? The first can be deceptive as people may navigate there by accident and the second can look high even though it may only be a few people visiting regularly.

All in all I can't see an urgent use for it so I would say no, not in the near future.

'Sign in' polls might be useful but not necessarily to rate the content. I know it would make the meet signups a lot easier to track! This has been discussed regularly and as iCan have already got a version of them I wouldn't be suprised if they did pop up at somepoint, I wouldn't hold your breath however as they probably wouldn't help the EG so aren't high up on the list.

Mel

There is an internal page that lets staff see how many 'hits' an entry has had. I'd quite like to see pages shown their hits, but I don't know if it will ever be allowed.