After reading this post 2-3 times it started to make some sense to me, so I upvoted it. But the writing was a bit too impenetrable for me to understand it on first read, or to communicate that I would have to read it multiple times to understand it. The fact that it was written by you specifically was the thing that got me to read it multiple times, since I expected there to be hidden value in it, but I am not sure how much I should expect this of other readers.
This is currently making me hesitant to promote it to the frontpage, even though I think the core point is a valuable one.
I think at least part of this is the invisibility of hyperlinks by default. I was able to guess pretty accurately where links were from context, but without links highlighted in some way, the whole thing has a feel of “you should already know what all of this is referring to and this should be trivial to parse”, which of course will not be true for most readers, and makes me want to simultaneously rush and strain mentally. (When links are visible I’m more forgiving with myself if I don’t understand something on first glance.)
I doubt this is 100% of the problem but it seems like a factor here.
Strong agreement that hyperlinks are currently way too hard to see; compare LW2.0 with LW1.0 in this respect, or a post on SSC. I often find hyperlinks that I don’t click on quite helpful, since they often are a signal that you are referring to a not-common-knowledge concept and provide a hint as to what that concept is. At their best, the link’s url tells you a lot more, as a mini-greenlink. Others are saying ‘I have data and here’s what type of data that is, and you can look if you really want to’ which again can be a key message.
I get that we’re going for a sparse look, but I think we should at least be giving full underlines.
This was definitely an intentional choice, since one of the most common experiences we’ve encountered during the original sequences was that people seemed overwhelmed by the links, and that they couldn’t really parse the text at all because of all the links.
It might be that I erred too much on the side of subtle underlines. When I tried giving them full underlines it did break the flow of the reading experience quite a bit, but I might try it again with either an off-black underline, or with some additional subtle highlighting of the text itself when it is part of a link.
Interesting. Was this a visual thing, where so many visible links caused some sort of overload in their brains? Was it that they felt the need to click on all the links if they hadn’t already read them? Or was it something else?
This seems like an important thing to get right in deciding how to link things, not only in deciding how to display them. Displaying them ‘quietly’ solves some problems but could even make others worse, and as we write we need to decide how much to link. If I’m over-linking, even after cutting drastically down on joke-linking at Raymond’s suggestion, then I need to adjust. I feel like other people might be reading posts on the internet quite differently than I do, and as this is not frequently discussed, wide variance is likely (as per Ozy’s point recently).
Another thing to think about is whether green-linking would change this calculus when we get it. I found arbital green-links not distracting at all and very helpful, but if others pay a higher cost for normal links, it is possible they also pay a higher cost for green links than I would expect.
Hmm, this might be a contrast setting difference, but I can pretty easily identify links, though they are somewhat subtle. Do you mean that for you they are basically completely invisible, or that they are so subtle that you don’t notice them unless you are specifically looking for them?
I agree that the core point is valuable and that this post is difficult to parse. I am confident I got it, but I have a lot of unfair advantages doing that, and it still required reflection.
My diagnosis is that this post assumes a lot of knowledge, both of concepts and terms, that are not reasonable to assume. In some cases links are provided, but in some cases they aren’t, and the links at best sort of get you off the hook here. The title is a huge hint to what is going on here, but the logic and terms involved need to be much more explicit.
I saw similar problems with your first main page post; it is clear that you overestimate what concepts and ideas readers are familiar with (even those who know a lot, often are missing any given thing, myself included), and how easily people can follow your logic.
I think that the better version of this post is roughly twice as long as this one.
(Note that this is not the general mistake people make, most people need to tighten things up, but this particular author making the opposite of a typical mistake should be entirely unsurprising.)
After reading this post 2-3 times it started to make some sense to me, so I upvoted it. But the writing was a bit too impenetrable for me to understand it on first read, or to communicate that I would have to read it multiple times to understand it. The fact that it was written by you specifically was the thing that got me to read it multiple times, since I expected there to be hidden value in it, but I am not sure how much I should expect this of other readers.
This is currently making me hesitant to promote it to the frontpage, even though I think the core point is a valuable one.
I think at least part of this is the invisibility of hyperlinks by default. I was able to guess pretty accurately where links were from context, but without links highlighted in some way, the whole thing has a feel of “you should already know what all of this is referring to and this should be trivial to parse”, which of course will not be true for most readers, and makes me want to simultaneously rush and strain mentally. (When links are visible I’m more forgiving with myself if I don’t understand something on first glance.)
I doubt this is 100% of the problem but it seems like a factor here.
Strong agreement that hyperlinks are currently way too hard to see; compare LW2.0 with LW1.0 in this respect, or a post on SSC. I often find hyperlinks that I don’t click on quite helpful, since they often are a signal that you are referring to a not-common-knowledge concept and provide a hint as to what that concept is. At their best, the link’s url tells you a lot more, as a mini-greenlink. Others are saying ‘I have data and here’s what type of data that is, and you can look if you really want to’ which again can be a key message.
I get that we’re going for a sparse look, but I think we should at least be giving full underlines.
This was definitely an intentional choice, since one of the most common experiences we’ve encountered during the original sequences was that people seemed overwhelmed by the links, and that they couldn’t really parse the text at all because of all the links.
It might be that I erred too much on the side of subtle underlines. When I tried giving them full underlines it did break the flow of the reading experience quite a bit, but I might try it again with either an off-black underline, or with some additional subtle highlighting of the text itself when it is part of a link.
Interesting. Was this a visual thing, where so many visible links caused some sort of overload in their brains? Was it that they felt the need to click on all the links if they hadn’t already read them? Or was it something else?
This seems like an important thing to get right in deciding how to link things, not only in deciding how to display them. Displaying them ‘quietly’ solves some problems but could even make others worse, and as we write we need to decide how much to link. If I’m over-linking, even after cutting drastically down on joke-linking at Raymond’s suggestion, then I need to adjust. I feel like other people might be reading posts on the internet quite differently than I do, and as this is not frequently discussed, wide variance is likely (as per Ozy’s point recently).
Another thing to think about is whether green-linking would change this calculus when we get it. I found arbital green-links not distracting at all and very helpful, but if others pay a higher cost for normal links, it is possible they also pay a higher cost for green links than I would expect.
Hmm, this might be a contrast setting difference, but I can pretty easily identify links, though they are somewhat subtle. Do you mean that for you they are basically completely invisible, or that they are so subtle that you don’t notice them unless you are specifically looking for them?
Completely invisible unless I mouseover
On Windows 10 I can see underlines of the links in Firefox/Chrome and Edge.
What system are you using? If it’s not Safari, did you try deleting the cache?
Interesting, can you post a screenshot?
I’d like to try to figure out together what could make it and subsequent posts clearer.
I agree that the core point is valuable and that this post is difficult to parse. I am confident I got it, but I have a lot of unfair advantages doing that, and it still required reflection.
My diagnosis is that this post assumes a lot of knowledge, both of concepts and terms, that are not reasonable to assume. In some cases links are provided, but in some cases they aren’t, and the links at best sort of get you off the hook here. The title is a huge hint to what is going on here, but the logic and terms involved need to be much more explicit.
I saw similar problems with your first main page post; it is clear that you overestimate what concepts and ideas readers are familiar with (even those who know a lot, often are missing any given thing, myself included), and how easily people can follow your logic.
I think that the better version of this post is roughly twice as long as this one.
(Note that this is not the general mistake people make, most people need to tighten things up, but this particular author making the opposite of a typical mistake should be entirely unsurprising.)
I agree: the better version of this post is roughly twice as long as this one.