Our article voting system has evolved progressively. From one person, one vote to a weighted system, to requiring comments when down-voting, to a system that statistically removed junk votes, and then lately to a system that recognised that voting patterns are not only bell curves, but sometimes, legitimately, bimodal.
We have, to a large degree, been successful at suppressing malicious down-voting. Too successful, it seems, and the article voting system is now massively weighted towards up-votes rather than down-votes. To up-vote you merely click the 4 or 5 rating. To down-vote you need to add a comment, and if your down-vote doesn't agree with the majority then your vote may not be counted until a sufficient number of other members have likewise voted the article down.
So while up-voting is great in that it rewards authors and gives readers a way to say thanks, up-votes are bad when the up-votes are not votes based on the technical merit of an article but instead based on being the author's friend, family or colleague. Make it 50 friends, family members or colleagues and the vote for a given article is hopelessly invalid.
Basically: you can have too much of a good thing. It's easy to up-vote, hard to down-vote, and so the average article rating goes up and the ability to sort the wheat from the chaff goes down.
Starting today we're removing a barrier on down-voting. You are no longer forced to provide a comment when down-voting. We have our historical-based expectations on what will happen but will be monitoring the results closely just in case.
The change is effective as of now. As always we're open to suggestions and ideas to make it even better.
We have an occasional issue whereby an author will get their friends, family, colleagues, and random people off the street to vote for their article. Our voting system[^] is geared towards handling a case where lots of people say "this is great" and a few downvoters say "boo, it's crap" by filtering out the outliers on the assumption that the group vote rules.
However, when you have 50 low rep voters saying "it's a 5" and 5 high rep voters saying "this is terrible (or dangerous)" then we need to adjust. The change we've made is that if a certain number of high rep members vote a certain way then no votes of that given score will be filtered out. The naysayers will be allowed to nay-say and balance will be restored.
We can now turn off the ability for an individual's signature to automatically populate a forum posting. This stops passive abuse ("it's just in my sig") of posting rules in the forums.
I've turned on the ability to lock messages or whole threads. I'm happy for threads to meander and evolve - that's the beauty of threaded discussions - but we do need the ability to put a lid on heated debates and also a means to keep discussions focussed.
We are now checking every forum message for spam, and every message that even hints at being spam will be moved to a moderation queue. We won't nuke the message: we leave that dirty work to the top-rep members and protectors. Once the message is approved then it gets posted as a regular message. Those messages that are moderated by a member as being spam will never see the light of day.
We hope this helps with the current challenges, and we'll be extending this to Quick Answers and the Article system as soon as this beta test is complete.
We've had Google, Facebook and LinkedIn login for a while but we're happy to announce that you can now login to CodeProject using your Windows Live ID. Just click the Windows icon at the bottom of the signin dropdown (top right of each page) or the Windows logo on the sign up / sign in page.
We are continuing to enhance releasing api.codeproject.com, our API service for those looking to harness the data and services of CodeProject. Please dive in, check the docs and the samples, and get back to us with any issues or suggestions.
As part of this we're deprecating APIs that haven't been used for years. The specific APIs removed in today's update will be: