Hacker News Reader

Comment by davej | Hacker News Reader

I wonder how many people gave kudos to the post by mistake? I know I did and there's no way to remove the kudos you gave afterwards. User actions should be a click not a hover.

View on HN

He's already received complaints and sort of laughed them away: http://dcurt.is/unkudo

Personally I think it's bad UX, but what do I know; I'm not Dustin Curtis.

I just hovered it, damn you dcurtis!

I have a ton of respect for Mr. Curtis, but he's utterly and completely wrong on the issue. It's not a "serious" thing, just like calling someone a mean name isn't necessarily a "serious" thing, but both kind of make you a dick.

How could you possibly care enough about Curtis' made-up number to call him a dick?

Never mind. I couldn't possibly care. It's a game. Relax.

It's a dark pattern. In my opinion, at least; an ambiguous interface element with unexpected and unintended consequences. Judging by his obvious design talent, I think Mr. Curtis should know better, especially in light of the incredibly negative response. It's a dick move.

You care enough to follow through with replying to me, it seems. I'm plenty relaxed, I just thought it was worth calling a spade a spade in this case.

Except it doesn't actually have consequences.

It does have, and you're seeing them right here and now.

It's definitely bad UX. Curtis tries to brush away criticism with his claim that the "kudos" number is "an otherwise meaningless number," but this argument is fatally flawed. Plenty of things that human beings use for social signaling are, when removed from context, technically "meaningless", but within context they can hold plenty of meaning. In the case of Curtis's website, the "Kudos" counter is implied to display a count of voluntary approval. By essentially tricking users into incrementing the count, Curtis violates this expectation, which is why everyone feels "cheated" by it. Subsequent visitors to the site who don't realize the counter has been gamed may take it as an indicator of how many people enjoyed the article, when in fact the numbers are inflated by other users against their will. This is not only terrible UX, but bordering on outright deceit.

In a world dominated by social media and ubiquitous "Like" counters, for someone to brush off a Kudos count as "an otherwise meaningless number" is either an incredible display of naivety or straight-up bullshit.

While I agree, my complaint with it is even simpler: hit refresh and you can ostensibly keep adding kudos without limit. If that's not the case, then it's a UI problem because it doesn't tell you as much. If it is the case, then it certainly removes whatever meaning it has the moment someone starts inflating the count.

The problem is really a combination of the two. People expect kudos to be tracked in a reasonable fashion for the reason that, if the counter exists, it must be taken seriously. The rest of the system is stripped down, leaving people with the impression that whatever is left must be important. This is a false impression, and that's what's disappointing about the feature. Dustin Curtis has taken a serious (read: pretentious) approach to the rest of the site, but he's left the kudos system hanging by a chad.

I repeated this about a thousand times and he seems to have fixed it ;). Also, he seems to have reverted the count on his unkudos page. Not the actions I would have expected from someone who thought nothing of it.

I disagree with your outright deceit allegation, because the number of people who enjoyed the article enough for a kudos is sure to be less than the number of people who bothered with the widget in the upper-right corner of the screen. I wish I had a study handy but it's well-established that far fewer people who read an article continue interacting with the page after they're done. For me to believe that it's a serious problem I'd have to be convinced that at least 50% of the kudos were from people who didn't think it was worthy of kudos.

On the other hand, first-time visitors or those that come to look at the design (as those who get there from this discussion) are IMO likely to think "What's that? Let's hover it; maybe some tooltip will pop up"

It's a pretty obvious dark pattern.

this is a good reminder as to how different visual and user experience design are. just because you can make a pretty design does not automatically qualify you to design useable software.

This may be a UX/Visual issue, but how is it a usability issue?

It breaks discoverability by executing its action /during discovery/.

Imagine if you, using a new web browser, clicked on the "File" menu to see what was there and instead of dropping down a menu, it opened a file browser. In the end not a life-changing or permanent problem, but bad usability problem.

Also, usability is a subset of UX.

I would agree with you, but this is a blog. You're taking it too seriously.

From this very article it seems the author of the blog is taking its design pretty seriously. Why shouldn't we?

I know usability is a subset of UX. Thanks for the explanation. Breaking discoverability makes sense re: usability.

It's easy to call it a meaningless interaction, but it's far from meaningless. He's explicitly breaking an implicit trust between himself and the reader. He's saying "Hey reader, I'm watching you."

If you don't think this is a big deal, I invite you to think about a natural extension of this concept: A comment form that sends its entire contents to the server on every keypress.

I wonder how many already do that....

Stop wondering and get a browser or extension that will somehow explicitly notify you when a connection is active.

The keypresses could be stored until the form is sent, the javascript doing this could be heavily obfuscated, the message could be encrypted.

On closer inspection, the site doesn't check for duplicate kudos. Why don't we give him all the kudos he wants?

Here's a quick python script I whipped up to dole them out. Customize to the blog post of your choosing, if you like.

https://gist.github.com/2167738

Needs a session cookie now.

https://gist.github.com/2169730

I thought the same the first time I used it, I did however only do this mistake once.

Maybe it’s a case of something that is confusing at first, but people learn fairly quickly?

Reply on news.ycombinator.com