OK, I looked into the Expression forums specifically. The Answers on the home page DO have the highest popularity. I wonder, though, if we should weight replies more heavily than views for the poularity vote. I'm not sure that IS the proper thing to do, since the idea is to highlight the threads that people are looking for the most, and most people who are looking for the answer to a problem would never post once they found the answer...
I think we thougt about a sliding timescale, but the question is how big to make that window. It's very possible that a post on how to do something posted now will still be completely relevant (and something that a lot of people are looking for the answer to) a year from now.
What do you think?
Matt
Matt Fraser, STO Forums Software Developer
Hmm, yes, it is possible that an answer may still be relevant after a year. No disputing that. The problem is that it may not be the only thing relevant, and that other, more recent issues never get a shot at top answer listing, since they inherently start out hundreds of views behind in popularity.
Consider the current situation; not one of those top five answers has an entry or reply subsequent to the release of EW2 Final (RTM), an event which brought a lot of new users (including myself) into the forum, with new questions, a number of which deal with, for example, the newly available PHP support. These questions will never have a shot at appearing in that list.
In fact, given their current lead in views, I find it difficult to imagine that the current top five would ever be supplanted. There's a psychological issue at play here, too. I'm a gearhead-type MCP who came here from DMX 2004 and Visual Studio/VWD. One of the first things I did after joining the forum was to check out those top answers to see if they addressed issues I might encounter, or features about which I wanted to learn.
In other words, they toted up another set of views not because they were the most relevant, or sought after, but simply because they were there and captioned "Top Answers." Had they been any other five answers, in exactly the same circumstances, I would have checked them out, too.
One thing I don't know, and can't take into account when considering how this might otherwise be implemented, is exactly how "popularity" is determined. If it is, indeed, simply views and replies, then I suggest that the selection method is flawed. You say above that the objective is to "highlight the threads that people are looking for the most." Using these criteria, that is not necessarily what you are doing, since there are several things other than search results which contribute to the views of these threads.
Perhaps a different approach is in order, one which I believe would help overcome the inherent bias in the current method. Instead of using views and replies, or in addition to those criteria, how about taking into account the actual search terms entered over the course of a period of time?
In any event, I do believe that a sliding time window would offer the best combination of relevance and opportunity for new topics to enter the list. After all, even with a sliding window, a truly sought-after (i.e. highly relevant) topic will still bubble to the top, and might even provide answers which are even more relevant (say, because they include information about EW2 which wasn't available previously).
cheers,
scott
