Recent Topics

1 Jan 09, 2011 13:27    

I gave a try to the new 4.0.3. Even though I am a "set if and forget it" type of person, I went for it because of the improved spam comment handling features.

Indeed, it is a great step forward, a great deal of time can be saved. I think it can be further improved, using 2 approaches:

- allowing an action to be applied to a group of comments
- marking the GOOD comments (which are typically a small percentage of all the new comments), and then removing all the remaining unmoderated comments.

I've explained it in [url=]this video[/url].

I am currently setting up a [url=]user experience research lab[/url]; my students and I can develop a detailed sketch of what it should look like.

It would be great if someone could share their opinion on this matter.

2 Jan 21, 2011 03:06

Great! Well made presentation! I totally agree with you!

3 Jan 21, 2011 21:44

Thank you for the feedback, I am glad that you agree with the idea. I hope others will join the conversation, and hopefully we can devise a model that provides a better experience.

4 Jan 22, 2011 00:00

gr8dude wrote:

- allowing an action to be applied to a group of comments
- marking the GOOD comments (which are typically a small percentage of all the new comments), and then removing all the remaining unmoderated comments.

The first sounds reasonable.I've always shouted out loud for checkboxes to allow multiple selection for easy moderation.

The latter shouldn't be the way to approach it.Theoritacally actual comments > spam comments.Practically you may be having too much spam, making the number of spam comments > actual comments, however you are normally not supposed to.

I am pretty satisfied with my evo recently.I am nearly getting zero spam comments.Maybe my hits may not be that high (2.5k unique and 5-6k pageviews per day), but still it's a thing right.But of course "set it and forget it" does not apply to me.I have set my comments to be "moderated".Because when you "set and forget it" you end up being filled with spam no matter what precautions you take, there are examples of it even in big organizations, corporate web sites etc.If you insist on setting and forgetting it, you have to put limits (i.e. John Resig closes posts to comments after 15 days of publishing them) since you can not check everything every time.

As far as i am concerned, the ideal way would be:

Allow comments unmoderated for everything.(free as a bird)
Set comments of post/page to be either closed/moderated after a given time(say 15-30 days).
DO NOT use Captcha, it does nothing except annoying your actual visitors, and setting barriers for handicapped.
Take easy measures of spam checking:
- bait the spam bot with an input field [email or website or use your imagination and create a new field].Hide the bait, and if it gets filled, tell it to f* off.
- Now afai believe the above is a simple method to detect spam bots, however there is not much to avoid human generated spam.It can be only avoided by setting a treshold between comments (which is also kind of annoying and limits the actual user), and block certain ips, websites and emails (be careful doing that, you might block actual visitors from that server).

that is my two cents, some might make sense some might not

5 Jan 23, 2011 22:16

Thank you for your feedback.

I've thought about it and I think that checkboxes are not the best approach. Fitts' law dictates that widgets are easy to use if they are:
- close to the current location of the pointer
- big targets

Checkboxes are neither. Each comment has a checkbox, so you have to move the mouse pointer all the time. They are also quite small, so you have to do some sniping to hit one accurately. I know, I know - it is easy enough for experienced folk, but it doesn't mean we cannot make it even easier.

I guess that a friendlier approach is:
- publish the GOOD comments by pressing the 'Publish' button (it is large and easy to click)
- in the end, you are left with BAD comments - which can be removed in a single click of a hypothetical 'Remove spam' button

It is the same thing:
- N comments
- M good comments out of N
- cost of operation = M+1 clicks (M for the comments, 1 for "remove the remaining ones")

The checkboxes would also take M+1 clicks, but those M clicks would require more effort.

Now, what worries me is that you say that in your case M is very close to N (i.e. most of your comments are good). My experience, unfortunately, is the opposite. I run two sites with b2evo, a company blog and my personal site. In both cases, there is so much spam that my spirits are dampened and I don't feel like browsing through them anymore. Really, there are many, so I am forced to use a CAPTCHA. I know it is inhumane, but I don't have other choices.

Is there some tutorial that walks one through the process of adding a decoy field to a form and describing its logic?

To further extend your description of the ideal approach, I think it can be improved if there was a mechanism that removes the burden off the shoulders of returning readers who have previously left a comment that was approved.

Each time you post a comment, b2evo leaves a cookie (or something similar, I am using that term because I am not aware of a better one). When a new comment is posted, b2evo verifies if the person has been here before, and if yes - see the status of their comment. If it was approved - it will accept the new comment without further verification.

This way a person only has to 'pay' when leaving their first comment; returning users have no constraints. Sure, if you use another browser or a different computer, b2evo won't see you as the same person. That can be addressed by verifying the email address left by the reader (if they left it). Since email addresses are not public, b2evo can check the email address assigned to the comment, and see if that email address was previously used with comments that were approved.
These are little tricks, but they do provide a better experience for returning readers.

Hmmm... I wonder.. what am I doing wrong, if I keep getting all that spam. I wonder whether in this case I represent the minority, or the majority [or neither] :-)

6 Jan 28, 2011 16:09


that may be true for your situation, but we can not assume the average user gets bad comments > good comments, and take precautions according to that assumption.

Other than that, first you should detect the type of spam you are getting. is it a spambot, or actual humans spamming you ?

if you are getting most of your spams via spambots, that's fairly easy to deal are not likely to be attacked by thousands or hundreds of spambots from different origins.check for email patterns, keyword patterns, ip's, everything, they should have sth in common.And be ruthless taking action.Do not think of too much detail, do not be afraid of blocking ip's, ip ranges, even a whole hosting provider if they cause too much trouble to ya.

About actual humans submitting spam comments, ther e is not much to do there, other than using the banning methods above, but it's a harder task of course.God forbid everyone from those chinese or indian sweatshop spammers.

As far as the decoy comment field goes; check this topic

7 Jan 30, 2011 11:12

that may be true for your situation, but we can not assume the average user gets bad comments > good comments, and take precautions according to that assumption.

You are right, unless we conduct some social studies, we can't know. I think we can address this problem in such a way that both camps are satisfied.

The solution is to add 2 buttons:
- Publish all unmoderated comments
- Delete all unmoderated comments

If you are in the goodComments > badComments group, you will manually delete a few bad ones, then publish all the good ones in one click.

If you are in the goodComments < badComments group, you will manually approve the few good ones, then get rid of all the remaining comments in one click.

This is a risky operation that can lead to data loss, so "delete all unapproved comments" must request a confirmation - to reduce the possibility of accidental removal.

This could be further enhanced by minimizing the cost of a mistake and making the system forgiving; i.e. provide an Undo feature. This option, however, requires more effort, as it implies changes in other modules (ex: add a new status for comments in the database schema, test that, etc).

So, you pointed out that there are 2 scenarios and if you optimize for one, the others suffer, and vice-versa. My current suggestion takes care of that concern, because it makes the lives of both groups easier, and it is not an "us vs them" issue anymore :-)

----------, I am doing something similar on a forum I manage [it runs on VBulletin]. There is a plugin called spamBuster, which takes a list of regular expressions and checks how many of them matched the text in a new message. If the score is above a certain threshold, the post is not published.

The method works, but spammers continue to promote new products and services, so you will have to update the list of regexps with whatever is the new and improved version of Via-gra or new toys from Apple, etc. It is tiring, not to mention that regexps are not a trivial subject :-)

8 Jan 30, 2011 13:34

Checkboxes ++ "select all" ++ "delete"/"publish" selected ;)


9 Feb 08, 2011 21:56

¥åßßå wrote:

Checkboxes ++ "select all" ++ "delete"/"publish" selected ;)


yep checkboxes would prettyy much do it

10 Sep 10, 2011 22:09

gr8dude, please have a look at how you can moderate comments in the dashboard on b2evo 4.1.

It's a 1 click on big button for publish or 1 click on big button for delete. Then the comment disappears and the next one scrolls up automatically.

It's not minimizing clicks below 1 per comment but it removes scrolling and it removes going to the next page of comments or things like that. Plus it's pretty satisfying to have that big red splash on spam and see them go to the trash.

Please let me know what you think and if you believe this needs further streamlining.


Form is loading...

Content Mangement System – This forum is powered by b2evolution CMS, a complete engine for your website.