Official Less Wrong Redesign: Nearly there
Design sketches
1. Notes on differences between these images and what we’re planning on implementing:
2. Meetup widget in sidebar, listing the 5 nearest meetups occurring in the next 2 weeks, linking to:
(something like right-click & “Open Image in new tab” will help here)
3. New headers:
Things I don’t like:
1. Disappearance of the background art in the heading.
2. “Discussion” tab is right next to “Comments”. This may be confusing to visitors, who will not immediately know the difference.
3. The bolding scheme in the Recent Comments sidebar, which is
instead of
(the current) or
4. The fact that comment scores have been moved from the top of the comment boxes to the bottom. This makes visual scanning by score harder.
5. Upvote/Downvote/Agree/Disagree buttons are not clearly labeled. They should be labeled with words, as they are now, not pictures, and also put next to each other, to remind people that there are two categories of voting and prevent new users from noticing only one of them.
6. I’m not sure I like the two categories of voting—a change in this area will change the existing status dynamics, to which my brain is already accustomed.
7. There seems to be more bright whiteness in the new design, as opposed to the “gray” feeling of the current scheme.
8. What’s wrong with bullets?
Also: “Hall of Fame” is not a good substitute for “Top Contributors”, since it implies that the people in it are no longer active. (Why bother to rename the list anyway? “Top Contributors” is a perfectly fine description.)
Why oh why can’t we have a way of bringing information that would be good for people to see into their view without turning the underlying system into “status dynamics”? :( I don’t care that karma is currently descriptively status, let’s figure out what we want to do, then how to implement it so the useful behavior doesn’t get hijacked by the status parasite.
Edit: Here’s what I want, off-the-cuff: I want a classifier for posts that buckets them into “definitely not worth anyone’s time”, “friggin’ awesome and everyone on LW and probably elsewhere should read this”, and “the rest”. I would like a classifier for comments that buckets them into “troll or hopeless”, “important part of the conversation”, “practically a required addendum to the post”, and “the rest”. I would like another classifier for comments that buckets them into “hilarious or otherwise worth reading”, “may cause loss of brain cells by reading”, and “the rest”. I would like a way to tell the author of a post or a comment “you are awesome”. I would like a way to tell the potential readers of a post or a comment “you will find this an important part of the conversation” or “you do not need to read this”. I would like a way to tell the potential readers of a post or comment “this is hilarious or awesome” or “this is distasteful”. I would not like these things to be conflated.
How about typing the words “You are awesome”?
The paperclips forgone to make them.
8 separate points in this comment, and 9 upvotes. I can’t tell which of your points has community support, so I’m discounting all of them.
That’s a shame—I’m losing a lot of information. Next time, 8 separate comments?
PS: “discount” != “ignore”
I would-have-upvoted-in-isolation 3, 4, 5, 6, and 7.
I prefer the old map-territory themed header over the new design.
I didn’t realize it was map/territory until now. I think it is clever, but it always felt a little odd and hard to read and I think I prefer the functionality of the new one.
I think the old header is sorta refreshing because the rest of the page is very clean (perhaps to the point of sterility) and devoid of graphic elements. Also, I find that the header guides the eye quite effectively towards the FHI and SIAI logos on the right.
I don’t; I find it a bit too cutesy.
Looking quite good.
I agree. I didn’t realize it was Map/Territory until now!
Thank you for all the effort you’ve put into this!
If we go through with the double karma, let’s at least collect statistics: given that a user voted something up/down on one karma, with what likelihood did they also vote it in the same direction on the other karma?
Maybe it’ll be revealed that for most posts, most users either click neither or both, and in the same direction.
One important use for the “agree/disagree” mechanism would be polling, which shouldn’t be tracked. Maybe the “a/d” buttons should show up only if the commenter chooses.
“Double karma” is a weird phrasing—from what I understood, “agree/disagree” shouldn’t have any impact beyond a single post, we don’t keep a tally of how many people agreed or disagreed with each user … right?
Maybe we should. I think I’m a very agreeable fellow. But maybe that’s the cognitive biases talking...
Right.
I worry that if we decouple “I agree with this” and “I like this”, the latter will correlate less well with contributing usefully to LW than the un-decoupled “vote up” currently does.
I don’t have any particularly convincing reasons why I should worry that; just a vague feeling that it’ll have more tendency to reflect poster status, applause lights, random witticisms, etc., and less tendency to indicate insight and expertise. Then again, looking at my own comment scores, I can’t say that my highest-rated comments have been the ones with most insight or expertise behind them as things are...
Hm, I have the opposite impression, so let’s break this down.
Current reasons to vote on a post:
A: Agree/disagree with the conclusion
B: Think the post is well argued / badly argued
C: Think the post is witty / appeals to applause lights
Right now the votes are A + B + C, now it would become B + C, you’re worried about the increased weight of C, I’m happy about the increased weight of B :)
It was a pejorative pretty much.
Maybe it’s just a screenshot artifact, but did you change the calming LW green to a more blueish, unpleasant hue? If so, let me be the first to say They Changed It, Now It Sucks.
(No, I don’t have anything more substantial to comment on.)
The colors look the same to me. The first letters of the words are capitalized in the new headers, though, which means that the first L shows more of the darker, bluer green than the old header’s L did. That might be what you’re noticing.
I checked and the text definitely has a different color. It’s currently #538D4D, but something like #38A174 in the screenshots.
It’s not just the headers. Green text on the body of the page is bluer looking too. It’s brighter and uglier.
Definitely like the old green better.
In the headers, the Singinst logo appears to be positioned lower than the FHI one, which is awkward-looking.
I didn’t join in the bikeshedding on the first round of discussion, but I might as well now!
Sorting by “Top” tends to disproportionately favor old posts. Is it possible to implement Wilson scoring ala Reddit?
I liked the map/territory distinction in the old banner. The new ones look good, except that the background looks like grey fleece or lint to me.
To my eye, too much padding around elements. Everything seems like it could be tightened up a little.
The comment icons seem a little cartoony. I have no attachment on text vs icons, but not a fan of those particular ones.
Are the tags gone from the sidebar? They are ugly as currently presented, but I don’t see anything that would even take me to a list of tags.
I think #1 is what ‘popular’ does which is already implemented.
Ah, I thought ‘popular’ was a combination of ‘new’ and ‘top’, sorting by recency and net votes.
Actually, you’re right.
The “thumbs up”/”thumbs down” look too much like they are for voting up or down. I don’t have much better suggestions for meaning “agree” or “disagree” though -maybe just a an X and a check-mark, like this? (but without colors, they’d be distracting).
If the thumbs are kept, a more iconic, abstract thumb icon would be preferable.
We’ll show a total, and word choice for the total might solve that problem:
+5 agreevs-3 (disagree)If you are showing words anyway, do you need the thumbs at all? How about just making the text clickable, and when you click it it becomes bold (like the current voting system) and the count updates.
I agree with Emile that showing thumbs seems like a bad idea, because (at least for me) “thumb up”/”thumb down” means “this is good”/”this is bad”. Indeed, the original meaning of “spare him”/”feed him to the lion” seems more like up/downvoting than like agree/disagreeing.
How about showing:
With the 5 and 3 clickable.
I think I am convinced that thumbs should be used for karma, and +- for agree/disagree.
Maybe the most natural thing for agree/disagree is a green check mark and a red cross? Then karma could maybe stay +/-.
ETA: I see Emile’s example incorporates a check mark and cross also, though there the cross looks a little like it’s being used as a check mark.
I think thumbs shouldn’t be used for agreement, since this is a more specific characterization than relevance/quality/yay! of Karma (and the purpose of agree/disagree is partially to make this component explicit where it’s well-defined), but +/- for agree/disagree will just be confusing.
My suggestion for agree/disagree are ‘==’ (equals) and ‘=/=’ (not equals) symbols drawn in sufficiently clear way, perhaps in circles.
So long as mouseover text (for example) makes it clear what on earth the == stuff means.
Mind you I suspect I will opt out of the agree/disagree feature via greasemonkey anyway. I tend to find that the judgement of humans is leaky and that the underlying cause of a negative judgement (be it disagreement with the comment, disagreement with other comments that are similar, disapproval of the subject or disapproval of the author) is something that must be inferred from the context. I suspect that I would find the behavioural patterns related to the agree/disagree feature slightly irritating to watch.
I think it is the honesty or straightforwardness of karma votes that appeals to me. “Trivial positive political action; trivial negative political action”. So when I see something that is obviously correct that gets downvoted heavily it just means that some folks had personal reasons to downvote. Fair enough. But if there were lots of ‘disagree’ votes then that really would mean that the thinking processes of the community was defective.
The above isn’t meant as an argument against or a complaint about the implementation of the new feature. I’m sure it’ll be great for those who are interested in that sort of thing. I just may personally filter it out of my awareness. :)
This might be too late to the discussion, but have you considered combining the vote systems into a grid below each comment, with the Y-axis as karma and the X-axis as agree/disgree? This would halve the number of clicks needed to vote.
How about something like this? The agree / disagree info is nice and short at the bottom right, and you get an explanation in a tooltip.
I’m not thrilled with the voting etc. buttons having their text replaced with symbols. Will they at least have mouseover text explaining what they do?
Absolutely yes.
The subscribe, save and report icons may be distracting, they’re not things you need very often. They could be hidden in a submenu (that drop-down thing at the top right, in which case you wouldn’t need icons any more), or be completely hidden (not greyed) when your mouse isn’t over the comment’s div.
The permalink icon could also be removed by instead making the time-and-date clickable (or it could be left as an icon—it’s not very distracting—but put next to the time-and-date.
(Also, thank you a lot for your work! Sorry for just offering criticism, of course plenty of things are very neat, but that goes without saying)
Great work! I especially like the “new since last visit” highlight; that will help following conversations across multiple visits quite a bit. That said, as Armok_GoB observed,
I noticed that too. I’m not sure why, exactly, but I have a few hypotheses. It could be that the shade of green has changed (it’s darker and more saturated now). I’m neutral on this change. Another big difference is that the sidebar has greatly increased in contrast, turned almost entirely bold, and become double spaced; all of these changes serve to make it more prominent than it was.
There’s also something different about the font rendering (it’s less crisp), but I suspect it’s just a browser difference; when I zoomed in on a screenshot of LW as it is now I saw color bands that indicate sub-pixel smoothing, but when I zoomed in on the screenshot it was monochromatic.
And finally, a few nitpicks with which you may or may not agree. The bullets on the sidebar probably shouldn’t sometimes be green, the Hall of Fame list is too far right, the Recent Posts and Recent Comments boxes should be shifted one bullet-width right to align with the other boxes even if they don’t have bullets so that they’re aligned. The space between the article title and the header image has a bit too much padding in general. The ‘report’ exclamation point should have its top and bottom aligned with the ‘save’ icon. The unbolded-green karma scores in the Hall of Fame are a font-color-boldness combination that doesn’t appear anywhere else, and it seems weird (why would the score be a link but not the name? Or is it?).
I like the idea of nesting from the left only.
Thanks for giving meetups their own space!
This is a poll.
Upvote this if you think separate agree/disagree and upvote/downvote buttons are a good idea.
Upvote this if you think separate agree/disagree and upvote/downvote buttons are a bad idea.
I’m rather skeptical and leaning towards “bad”, but am not confident enough to vote either option. I say we give it a try.
Above, emile broke it down to:
“Current reasons to vote on a post: A: Agree/disagree with the conclusion B: Think the post is well argued / badly argued C: Think the post is witty / appeals to applause lights”
Separating into agree/disagree and upvote/downvote changes it into A and B+C, as emile pointed out.
However, none of us really want karma to be about C. Maybe we could do better with a simple rewording? “good argument/poor argument” vs “upvote/downvote”? Wording could go in place of the icons, or as hover text.
Downvote this for karma balance.
Switch the symbols, thumbs up/down for good/bad post, plus/minus for agree/disagree. Thumb pointing is rating a performance, not voting to show agreement, while the jump form “me too” to “+” is fairly small. On some boards there even seems to be a convention of replying with “+1” to signify something like “I wanted to say that, too”.
I think it’s thoroughly ambiguous either way—context will have to do most of the work.
But surely we’ll want plus/minus to be the buttons that add to/subtract from karma.
Not quite related to the redesign, but can the comment “Sort By” dropdown’s values have a #comments anchor appended to them? When I change the way comments are sorted, I want to be start reading through them straight away, without having to scroll down.
This looks awesome. There is somehting which seem more… cold and hard somewhat than the current aesthetic, which is bad, but otherwise it seems great.
The single line motto in the main section header next to the two-lines of motto in the other two is a little disconcerting. Any reason not to have it match them?
Background image of the new header is much less dorky than the map and aerial photo but it causes a What Is It? reaction in me that causes me to examine it in detail to see if it signifies anything.
I’ll probably suppress the background image with Adblock Plus like I do the current one. (Actually, there are two, one being a gradient.)
In the post editor, make plain text copy paste the default.
So, karma scores officially mean “agree/disagree” rather than “more like this/less like this”?
No. There will be two voting systems in the redesign.
Can we make karma transferrable?
Too gameable.
That’s what rules are for. Mainly well specified prediction market bets.
If you go down this road, you’ll probably end up reinventing money.
I think the header should look more like this. (The font difference is unintentional; I didn’t have the same font)
I’d rather not have the thought that an animal died to make the header.
Unless the header took less than five-ninths of a second to make, it’s a little late for that.
Agreement/disagreement distinction is too simplistic, I think it’s a bad idea to polarize along these lines. Do I agree with Luke’s post on neuroscience and values, or do I disagree with it? What would that even mean? There are many comments for which the situation is similar. Do I “disagree” with a comment that boldly asserts the existence of God?
I propose that voting should be for quality and relevance of content.
Why doesn’t this incorporate Clippony, the official mascot of LessWrong?
Clippony is not the official mascot of less wrong, because the person who has the authority to make something official about Less Wrong (Eliezer) has not identified Clippony as our official mascot. Clippony is a permissible, unofficial mascot because I have released the artwork for the purpose, but not an offical one.
Why doesn’t this incorporate Clippony, the unofficial mascot of LessWrong?
Because it would not fit into the design aesthetic.