I feel more optimistic about this project after reading this! I like the idea of curation being a separate action and user-created sequence collections that can be voted on. I’m… surprised to learn that we had view tracking that can figure out how much Sequence I have read? I didn’t know about that at all. The thing that pushed me from “I hope this works out for them” to “I will bother with this myself” is the Medium-style individual blog page; that strikes a balance between desiderata in a good place for me, and I occasionally idly wish for a place for thoughts of the kind I would tweet and the size I would tumbl but wrongly themed for my tumblr.
I don’t like the font. Serifs on a screen are bad. I can probably fix this client side or get used to it but it stood out to me a surprising amount. But I’m excited overall.
I don’t like the font. … I can probably fix this client side or get used to it but it stood out to me a surprising amount.
My other comment aside, this is (apart from the general claim) a reasonable user concern. I would recommend (to the LW 2.0 folks) the following simple solution:
Have several pre-designed themes (one with a serif font, one with a well-chosen sans font, and then “dark theme” versions of both, at least)
Let users select between those themes via their Profile screen
This should satisfy most people, and would still preserve the site’s aesthetics.
I am slightly hesitant to force authors to think about how their posts will look like in different fonts, and different styles. While I don’t expect this to be a problem most of the time, there are posts that I write where the font choice would matter for how the content comes across.
Medium allows the writer to chose between a sans-serif and a serif font, which I like a bit more, but I would expect would not really satisfy Alicorn’s preferences.
Maintaining multiple themes also adds a lot of design constraints and complexity to updating various parts of the page. The width of a button might change with different fonts, and depending on the implementation, you might end up needing to add special cases for each theme choice, which I would really prefer to avoid.
Overall, my hypothesis is that Alicorn might not dislike serif-fonts in general, but might be unhappy about our specific choice of serif fonts, which is indeed very serify. I would be curios whether she also has a similar reaction to the default Medium font, for example displayed in this post: https://medium.com/@pshrmn/a-simple-react-router-v4-tutorial-7f23ff27adf
As with many such things, there are standard, canonical solutions to your concerns.
In this case, the answer is “select pairs/sets of fonts that are specifically designed to have the same width in both the serif and the sans variants”. There are many such “font superfamilies”. If you’d like, I can draw up a list of recommendations. (It would be helpful if you could let me know your constraints w.r.t. licensing and budget.)
Theme variants do not have to be comprehensive redesigns. It is eminently possible to design a set themes that will not lead to the content being perceived very differently depending on the active theme.
P.S.:
Overall, my hypothesis is that Alicorn might not dislike serif-fonts in general, but might be unhappy about our specific choice of serif fonts, which is indeed very serify.
I suspect the distinction you’re looking for, here, is between transitional serifs (of which Charter, the Medium font, is one, although it’s also got slab-serif elements) and the quite different old-style serifs (of which ET Book, the current LW 2.0 font, is one). (There are also other differences, orthogonal to that distinction—such as ET Book’s considerably smaller x-height—which also affect readability.)
Alicorn, if you’re reading this, I wonder what your reaction is to the font used on this website:
P.P.S.: It is also possible that the off-black text color is negatively impacting readability! (Especially since it can interact in a somewhat unfortunate manner with certain text rendering engines.)
Alicorn, what OS and browser are you viewing the LW 2.0 site on?
FYI, your assessment is in the extreme minority; most people who have seen that site have responded very positively to the font choice (and the typography in general). This suggests that your preferences are unusual, in this sphere.
I say this, not to suggest that your preference / reaction is somehow “wrong” (that would be silly!), but a) to point out the the danger in generalizing from one’s own example (typical mind blah blah), and b) to underscore the importance of user choice and customization options!
rest of this response is not specifically for Alicon but is re: this whole comment thread
This is still a gold standard of UX design: sane defaults plus good[1] customizability.
[1] “Good” here means:
comprehensive
intuitive
non-overwhelming (i.e. layered)
Note, these are ideals, not basic requirements; every step we take toward the ideal is a good step. So by no means should you (the designer/developer) ever feel like “comprehensive customizability is an unreachable goal; there’s no reason to bother, since Doing It Right™ is too much effort”! So in this case, just offering a couple of themes, which are basic variations on each other (different-but-matching font choices, a different color scheme), is already a great thing and will greatly improve the user experience.
The Merriweather and Merriweather Sans fonts (available for free via Google Fonts) are, as you can see, designed to be identical in width, line spacing, etc. They are quite interchangeable, in body text, UI, etc. Both are quite readable, and aesthetically pleasing.
(As a bonus, active on that page is a tiny bit of JavaScript trickery that sets different body text font weights depending on whether the client is running on a Mac, Windows, or Linux platform, to ensure that everyone sees basically the same thing, and enjoys equally good text readability, despite differences in text rendering engines. Take a look at the page source to see how it’s done!)
UPDATE 2: A couple of mockups (linking to individual images because Imgur’s zoom sucks otherwise). Be sure to zoom in on each image (i.e. view at full magnification):
1. OS (and version), browser (and version), device/display, etc.?
(General note: folks, please, please include this information whenever you say anything to a web designer/developer/etc. about how a website looks or works for you!!)
2. Great! If one of them feels clear, then this goes to show exactly what I was saying: user choice is good.
In the age of the Internet and in the company of nonconformists, it does get a little tiring reading the 451st public email from someone saying that the Common Project isn’t worth their resources until the website has a sans-serif font.
This is old “received wisdom”, and hasn’t been the case for quite a while.
Folks, this is what people mean when they talk about LessWrong ignoring the knowledge of experts. Here’s a piece of “knowledge” about typography and web design, that is repeated unreflectively, without any consideration of whether there exists some relevant body of domain knowledge (and people with that domain knowledge).
What do the domain experts have to say? Let’s look:
But this domain knowledge has not, apparently, reached LessWrong; here, “Serifs on a screen are bad” and “sans serif are for screens, and serif is for print” is still true.
And now we have two people agreeing with each other about it. So, what? Does that make it more true? What if 20 people upvoted those comments, and five more other LessWrongers posted in agreement? Would that make it more true? What amount of karma and local agreement does it take to get to the truth?
Here’s what I think is the conventional wisdom about serif/sans-serif; I don’t think it is in any way contradicted by the material you’ve linked to.
Text that is small when measured in display pixels is generally harder to read fluently when set in a typeface with serifs.
Only interested in readers with lovely high-DPI screens? Go ahead, use serifs everywhere; it’ll probably be fine. Writing a headline, or a splash screen with like 20 words on it? Use serifs if they create the effect you want; the text won’t be small enough, nor will there be enough of it in a block, for there to be a problem.
But if you are choosing a typeface for substantial chunks of text that might be read on a not-so-great screen, you will likely get better results with a sans-serif typeface.
So, what about those domain experts? Jakob Nielsen is only addressing how things look on “decent computer screens with pixel densities of 220 PPI or more”. Design Shack article 1 says that a blanket prohibition on serifed typefaces on screens is silly, which it is. But look at the two screenshots offered as counterexamples to “Only use serifs in print”. One has a total of seven words in it. The other has a headline in a typeface with serifs … followed by a paragraph of sanf-serif text. Design Shack article 2 says that sans-serif typefaces are better “for low-resolution displays”, though it’s not perfectly clear what they count as low-resolution. The Quora question has a bunch of answers saying different things, mostly not from “domain experts” in any strong sense.
I like seriffed typefaces. In a book, sans-serif is generally hideous and offputting to me. On my phone or my laptop, both of which have nice high-resolution displays, Lesser Wrong content with serifs looks just fine. (Better than if it were set sans-serif? Dunno.) On the desktop machine I’m using right now, though, it’s ugly and it feels more effortful to read than the corresponding thing on, say, Less Wrong. For me, that is.
now we have two people agreeing [...] Does that make it more true?
Yes. More precisely: the proposition we should actually care about here is not some broad generality about serif versus sans-serif typefaces, but something like “Users of Lesser Wrong will, on the whole, find it a bit easier on the eyes if content is generally set in sans-serif typefaces”. Consider the limiting case where every LW user looks at the site and says “ugh, don’t like that font, the serifs make it harder for me to read”. Even if all those users are shockingly ignorant of typography, this is a case where if no one likes it, then it is ipso facto bad.
Of course we don’t have (anything like) the entire LW community saying in chorus how much they dislike those serifs. But yes, when what matters is the experience of a particular group of people, each individual person who finds a thing bad does contribute to its badness, and each individual person who says it’s bad does provide evidence for its badness.
What amount of karma and local agreement does it take to get to the truth?
Karma is relevant here only as a proxy for participation. A crude answer to this question is: enough to constitute a majority of users, weighted by frequency of use.
In case I haven’t made it clear enough yet, I am not arguing that LW people are always right, or that high-karma LW people are always right. I am arguing that when the thing at issue is the experience of LW people, the experiences of LW people should not be dismissed. And I am arguing that on the more general question (are typefaces with serifs a bad idea on the web?) the simple answer “no; that’s an outdated bit of bogus conventional wisdom” is in fact just as wrong as the simple answer “yes; everyone knows that”.
And I am arguing that on the more general question (are typefaces with serifs a bad idea on the web?) the simple answer “no; that’s an outdated bit of bogus conventional wisdom” is in fact just as wrong as the simple answer “yes; everyone knows that”.
Disagree. (Keep reading for details.)
But if you are choosing a typeface for substantial chunks of text that might be read on a not-so-great screen, you will likely get better results with a sans-serif typeface.
This is still incorrect, because serif readability is superior to that of sans-serif, and see below for the matter of “not-so-great screens”.
Screen DPS
Given the pixel resolution per character you need to make serifs work, they are inferior on the screen… if you have a 72ppi (or less) display.
Now, such displays exist; here’s one. They are quite rare, though, and designed for entertainment, not work. The idea that any appreciable percentage of LW users have such hardware seems implausible.
On a ~96ppi display (such as this nearly decade-old cheap flat-panel I’m using right now, or indeed any display display made in the past 15+ years), the apparent (angular, a.k.a. “CSS reference pixel”) font size that you need to bring out the superiority of serif typefaces is no larger than the minimum size called for by other accessibility guidelines.
“The LW 2.0 font is less readable”
On the desktop machine I’m using right now, though, it’s ugly and it feels more effortful to read than the corresponding thing on, say, Less Wrong. For me, that is.
1. What OS is this on? If the answer is “Linux” or “Windows”, then part of the answer is “text rendering works very different on those operating systems, and you have a) test your site on those systems, b) make sure to make typographic choices that compensate, c) take specific actions to ensure that the user experience is adjusted for each client platform”. I of course can’t speak to (a), but (b) and (c) are not in evidence here.
2. The body text font size on LW 2.0 is too small (especially for that font), period. Again I refer you to https://www.readthesequences.com/Biases-An-Introduction; the body text is at 21px there. I consider that to be a minimum (adjusted for the particular font); whereas LW 2.0 (with a similar type of font) is at 16px. Yes, it looks tiny and hard to read. (But have you tried zooming in? What happens then?)
3. Other issues, like color (#444, in this case) affecting text rendering. I speak of this in my other comments.
“Consensus matters locally”
Consider the limiting case where every LW user looks at the site and says “ugh, don’t like that font, the serifs make it harder for me to read”. Even if all those users are shockingly ignorant of typography, this is a case where if no one likes it, then it is ipso facto bad.
If every LW user looks at the site and says that, then we can’t still conclude anything about serifs from that, because if all of those users have not the smallest ounce of typography or design expertise, then they don’t know what the heck they like or dislike, serif-wise.
Let me be clear: I’m not saying that people can’t tell whether they like or dislike a particular thing. I am saying that without domain knowledge, people can’t generalize their preferences. Ok, so some text on their screen is hard for them to read. What’s making it so? The fact that a font has serifs? Or maybe just that it’s a particular kind of serif font? Or the font weight? Or the weight grade? Or the shape of the letterforms (how open the curves are, for instance, or the weight variability, perhaps due to which “optical size” is being used)? Or the color? Or the subpixel rendering settings? Or the kerning? Or the line spacing? Or the line length? Or the text-rendering CSS property setting? If you (the hypothetical-user you) don’t know what most or all of those things are, then sure your preferences are real, but your opinion (generalized from those preferences) is worth jack squat.
In other words: “if no one likes it, then it is ipso facto bad”—yes, but what, exactly, is “it”? You’re equivocating between two meanings, in that sentence! So, this is true:
“If no one likes , then is bad.”
Yes. Granted. But you seem to want to say something like:
“If no one likes , then are bad.”
But any particular thing belongs to many different classes, which intersect at the point defined by that thing! Obviously not all those classes are ipso facto bad, so which one(s) are we talking about?? We have no idea!
I am arguing that when the thing at issue is the experience of LW people, the experiences of LW people should not be dismissed.
Dismissed? No. Taken at anything even remotely resembling face value? Also no.
Come on, folks. This is just a rehash of the “people don’t have direct access to their mental experience” debate. You know all of this already. Why suddenly forget it when it comes up in a new domain?
serif readability is superior to that of sans-serif
Do you have actual solid evidence for that? I’m guessing that if you did you’d have given it already in your earlier comments, and you haven’t; but who knows? (One of the answers to that Quora question mentions a study that found a small advantage for serifs. It also remarks that the difference was not statistically significant, and calls into question the choice of typefaces used, and says it’s not a very solid study. So I hope you have something better than that.)
On a ~96ppi display [...] the apparent [...] font size that you need to bring out the superiority of serif typefaces is no larger than the minimum size called for by other accessibility guidelines.
Again, I would be interested in more information about what evidence you have about the font size required “to bring out the superiority of serif typefaces”. For the avoidance of doubt, that isn’t a coded way of saying “I bet you’re wrong”; I would just like to know what’s known about this and how solidly. I do not have the impression that these issues are as settled as you are making them sound; but I may just be unaware of the relevant work.
What OS is this on?
One instance is Firefox on Windows; the other is Firefox on FreeBSD (which I expect is largely indistinguishable in this context from Firefox on Linux). I concur with your guess that the people responsible for LesserWrong have not done thorough testing of their site on a wide variety of platforms, though I would be surprised if no one involved uses either Windows or Linux.
Yes, it looks tiny and hard to read.
LesserWrong has what looks to me like a weird multiplicity of different text sizes. Some of the text is clearly too small (personally I like small text, but I am aware that my taste is not universally shared). However—and I must stress again that here I am merely describing my own experience of the site—if I go to, say, this post on the Unix box at my desk right now then (1) the size of the type at my typical viewing distance is about the same as that of a decently typeset paperback book at its typical viewing distance, and (2) I find the text ugly and harder to read than it should be because various features of the typeface (not only the serifs) are poorly represented—for me, on that monitor, after rendering by my particular machine—at the available resolution. (The text is very similar in size and appearance to that on readthesequences.com; LW2.0 appears to be using—for me, etc., etc. -- ETBembo Roman LF at 19.2px actual size, whereas RTS is using GaramondPrmrPro at 21px actual size. ETBembo has a bigger x-height relative to its nominal size and most lowercase letters are almost exactly the same size in each.)
Other issues, like color
Yup, agreed. But I would say the same about readthesequences.com even though its body text is black.
If every LW user looks at the site and says that, then we can’t still conclude anything about serifs from that,
I agree. (Though it would, despite their hypothetical ignorance, be evidence. Someone who says “this text is hard to read because of the serifs” may be wrong, but I claim they are more likely to say it in the face of text that’s hard to read because of its serifs than of text that’s hard to read for some other reason.)
Perhaps I left too much implicit in my earlier comment, so let me try to remedy that. I firmly agree that the mere fact that some LW users believe some proposition about serifs in webpage text is perfectly compatible with the falsehood of that proposition. Even if it’s quite a lot of LW users. Even if they have a lot of karma.
But the thing that actually matters here is not the general proposition about serifs, but a more specific question about the type used on LesserWrong. I wasn’t equivocating between this and the general claim about serifs, nor was I unaware of the difference; I was deliberately attempting to redirect discussion to the more relevant point.
(Not that the general question isn’t interesting; it is.)
[EDITED to add:] Of course much of what I wrote before was about the general proposition. Whether I agree with you about that depends on exactly what version of the general proposition we’re discussing—I take it you would agree with me that many are possible, and some might be true while others are false. In particular, I am somewhat willing to defend the claim that there are otherwise reasonable choices of text size for which typical seriffed typefaces make for a worse reading experience than typical sans-serif typefaces for people using 100ish-ppi displays, and that while this can be mitigated somewhat by very careful choice of serif typefaces and careful working around the quirks of the different text rendering users on different platforms will experience, selecting sans-serif typefaces instead may well be the better option. I am also willing to be convinced to stop defending that claim, if there is really good evidence against it.
Not close at hand. You may reasonably consider my claim to be undefended for now. When I have the time, I’ll try to put together a bit of a lit survey on this topic.
LesserWrong has what looks to me like a weird multiplicity of different text sizes. Some of the text is clearly too small (personally I like small text, but I am aware that my taste is not universally shared). However—and I must stress again that here I am merely describing my own experience of the site—if I go to, say, this post on the Unix box at my desk right now then (1) the size of the type at my typical viewing distance is about the same as that of a decently typeset paperback book at its typical viewing distance, and (2) I find the text ugly and harder to read than it should be because various features of the typeface (not only the serifs) are poorly represented—for me, on that monitor, after rendering by my particular machine—at the available resolution. (The text is very similar in size and appearance to that on readthesequences.com; LW2.0 appears to be using—for me, etc., etc. -- ETBembo Roman LF at 19.2px actual size, whereas RTS is using GaramondPrmrPro at 21px actual size. ETBembo has a bigger x-height relative to its nominal size and most lowercase letters are almost exactly the same size in each.)
Right you are. The 16px size is what I saw on the front page.
Even on my machines, ET Book (source) does not seem to render as well as Garamond Premier Pro (in a browser).
Though it would, despite their hypothetical ignorance, be evidence. Someone who says “this text is hard to read because of the serifs” may be wrong, but I claim they are more likely to say it in the face of text that’s hard to read because of its serifs than of text that’s hard to read for some other reason.
I think this is literally true but relevantly false; specifically, I think this is false once you condition on the cause of the text’s unreadability not being some gross and obvious circumstance (like, it’s neon purple on a fuchsia background, or it’s set at 2px size, etc.)
I think that someone who is ignorant of typography is no more likely to blame serifs in the case of the serifs being to blame than in the case of the text rendering or line length being to blame.
But the thing that actually matters here is not the general proposition about serifs, but a more specific question about the type used on LesserWrong. I wasn’t equivocating between this and the general claim about serifs, nor was I unaware of the difference; I was deliberately attempting to redirect discussion to the more relevant point.
Noted. I was responding to the general claim.
As to the specific question, the matter of serifs is moot, because (as with all specific design decisions), each designer decision should be comprehensively user-tested and environment-tested, and as much user choice should be offered as possible.
Of course much of what I wrote before was about the general proposition. Whether I agree with you about that depends on exactly what version of the general proposition we’re discussing—I take it you would agree with me that many are possible, and some might be true while others are false.
Indeed.
In particular, I am somewhat willing to defend the claim that there are otherwise reasonable choices of text size for which typical seriffed typefaces make for a worse reading experience than typical sans-serif typefaces for people using 100ish-ppi displays … I am also willing to be convinced to stop defending that claim, if there is really good evidence against it.
Nope, the claim is reasonable. Websites where information density is more important than long-form readability, or where text comes in small chunks and a user is expected not to read straight through but to extract those chunks, may be like this. For that use case, a smaller point size of “body” text may be called for, and a well-chosen sans font may be a better fit.
LessWrong is not such a website, though a hypothetical LessWrong community wiki may be (or it may not be; it depends on what sort of content it mostly contains).
(Aside: I somewhat object to speaking of “typical” serif typefaces, because that’s hard to resolve nowadays. I suspect that you know that, and I know that, but in a public discussion it pays to be careful with language like this.)
However:
very careful choice of […] typefaces and careful working around the quirks of the different text rendering users on different platforms will experience
… is always advisable, regardless of typographic or other design choices.
I have no knowledge of typography, but was thought in University that serif fonts should be used for screens, and sans serif for print; it is very possible, that my lecturers were wrong.
Would that make it more true?
No.
What amount of karma and local agreement does it take to get to the truth?
None. The truth is orthogonal to the level of local agreement. That said, local agreement is Bayesian evidence for the veracity of a proposition.
Mathematically, if the truth is orthogonal to the level of local agreement, local agreement cannot constitute Bayesian evidence for the veracity of the proposition. If we’re taking local agreement as Bayesian evidence for the veracity of the proposition, we’re assuming the veracity of the proposition and local agreement are not linearly independent, which would violate orthogonality.
An outcome is Bayesian evidence for a proposition, if the outcome is more likely to occur if the proposition is true, than vice versa.
Based on that understanding of Bayesian evidence, I argue that Lesswrong consensus on a proposition is Bayesian evidence for that proposition. Lesswrongers have better than average epistemic hygiene, and pursue true beliefs. You expect the average lesswronger to have a higher percentage of true beliefs than a lay person. Furthermore if a belief is consensus among the Lesswrong community, then it is more likely to be true. (A single Lesswronger may have some false beliefs), but the set of false beliefs that would be shared by the overwhelming majority of Lesswrongers would be very small.
An outcome is Bayesian evidence for a proposition, if the outcome is more likely to occur if the proposition is true, than vice versa.
That assumes that there is a statistical correlation between the two, no? If the two are orthogonal to each other, they’re statistically uncorrelated, by definition.
The local agreement (on Lesswrong) on a proposition is not independent of the veracity of the proposition. To claim otherwise is to claim that Lesswrongers form their beliefs through a process that is no better than random guessing. That’s a very strong claim to make, and extraordinary claims require extraordinary evidence.
“The local agreement (on Lesswrong) on a proposition is not independent of the veracity of the proposition.”
Sure, and that is equally true of indefinitely many other populations in the world and the whole population as well. It would take an argument to establish that LW local agreement is better than any particular one of those populations.
Well, in that case, what I conjecture is simply that either this (your university classes) took place a while ago, or your lecturers formed their opinions a while ago and didn’t keep up with developments, or both.
“Use sans-serif fonts for screen” made sense. Once. When most people had 72ppi displays (if not lower), and no anti-aliasing, or subpixel rendering.
… local agreement is Bayesian evidence for the veracity of a proposition.
Why? Are people around here more likely to agree with true propositions than false ones? This might be true in general, but is it true in domains where there exists non-trivial expertise? That’s not obvious to me at all. What makes you think so?
Are people around here more likely to agree with true propositions than false ones? This might be true in general,
I was generalising from the above. I expect the epistemic hygiene on LW to be significantly higher than the norm.
For any belief b, let Pr(b) be the probability that b is true. Forall b such that b is a consensus on Lesswrong (greater than some k% of Lesswrongers believe b), then Pr(b) > 0.50 is a belief I hold.
But this is an entirely unwarranted generalization!
Broad concepts like “the epistemic hygiene on LW [is] significantly higher than the norm” simply don’t suffice to conclude that LessWrongers are likely to have a finger on the pulse of arbitrary domains of knowledge/expertise, nor that LessWrongers have any kind of healthy respect for expertise—especially since, in the latter case, we know that they in fact do not.
simply don’t suffice to conclude that LessWrongers are likely to have a finger on the pulse of arbitrary domains of knowledge/expertise
Do you suggest that the consensus on Lesswrong about arbitrary domains is likely to be true with P ⇐ 0.5? As long as Pr(B|lesswrong consensus) is > 0.5, then Lesswrong consensus remains Bayesian evidence for truth.
I expect that for most domains (possibly all), Lesswrong consensus is more likely to be right than wrong. I haven’t yet seen reason to believe otherwise; (it seems you have?).
Again, there is nothing special about this. Given that I believe something, even without any consensus at all, I think my belief is more likely to be true than false. I expect this to apply to all domains, even ones that I have not studied. If I thought it did not apply to some domains, then I should reverse all of my beliefs about that domain, and then I would expect it to apply.
The LessWrong consensus is massively overweighted in one particular field of expertise (computing) with some marginal commentators who happen to do other things.
As for evidence to believe otherwise, how about all of recorded human history? When has there ever been a group whose consensus was more likely to be right than wrong in all domains of human endeavor? What a ludicrous hubris, the sheer arrogance on display in this comment cowed me, I briefly considered whether I’m hanging out in the right place by posting here.
Let B be the set of beliefs that are consensus among the LW community. Let b be any arbitrary belief. Let Pr(b) be the probability that b is true. Let (b|B) denote the event that b is a member of B.
I argue that Pr(b|B) (Probability that b is true given that b is a member of B) is greater than 0.5; how is that hubris?
If Lesswrongers are ignorant on a particular field, then I don’t expect a consensus to form. Sure, we may have some wrong beliefs that are consensus, but the fraction of right beliefs that are consensus is greater than 1⁄2 (of total beliefs that are consensus).
This entire thread is reason to believe otherwise. We have the LessWrong consensus (sans-serif fonts are easier to read than serif fonts). We have a domain expert posting evidence to the contrary. And we have LessWrong continuing with its priors, because consensus trumps expertise.
I’m not continuing with my priors for one (where do you get that Lesswrong is continuing with its priors?).
It is not clear to me that “serif fonts are easier to read than sans-serif fonts” was ever/is a consensus here. As far as I can tell, fewer than ten people expressed that opinion (and 10 is a very small sample).
1 example (if this was that) wouldn’t detract from my point though. My point is that lesswrong consensus is better than random guessing.
Now that I’m looking at it more closely: Quoted text in comments does not seem sufficiently set off. It’s slightly bigger and indented but it would be easy on casual inspection to mistake it for part of the same comment.
I feel more optimistic about this project after reading this! I like the idea of curation being a separate action and user-created sequence collections that can be voted on. I’m… surprised to learn that we had view tracking that can figure out how much Sequence I have read? I didn’t know about that at all. The thing that pushed me from “I hope this works out for them” to “I will bother with this myself” is the Medium-style individual blog page; that strikes a balance between desiderata in a good place for me, and I occasionally idly wish for a place for thoughts of the kind I would tweet and the size I would tumbl but wrongly themed for my tumblr.
I don’t like the font. Serifs on a screen are bad. I can probably fix this client side or get used to it but it stood out to me a surprising amount. But I’m excited overall.
My other comment aside, this is (apart from the general claim) a reasonable user concern. I would recommend (to the LW 2.0 folks) the following simple solution:
Have several pre-designed themes (one with a serif font, one with a well-chosen sans font, and then “dark theme” versions of both, at least)
Let users select between those themes via their Profile screen
This should satisfy most people, and would still preserve the site’s aesthetics.
I am slightly hesitant to force authors to think about how their posts will look like in different fonts, and different styles. While I don’t expect this to be a problem most of the time, there are posts that I write where the font choice would matter for how the content comes across.
Medium allows the writer to chose between a sans-serif and a serif font, which I like a bit more, but I would expect would not really satisfy Alicorn’s preferences.
Maintaining multiple themes also adds a lot of design constraints and complexity to updating various parts of the page. The width of a button might change with different fonts, and depending on the implementation, you might end up needing to add special cases for each theme choice, which I would really prefer to avoid.
Overall, my hypothesis is that Alicorn might not dislike serif-fonts in general, but might be unhappy about our specific choice of serif fonts, which is indeed very serify. I would be curios whether she also has a similar reaction to the default Medium font, for example displayed in this post: https://medium.com/@pshrmn/a-simple-react-router-v4-tutorial-7f23ff27adf
As with many such things, there are standard, canonical solutions to your concerns.
In this case, the answer is “select pairs/sets of fonts that are specifically designed to have the same width in both the serif and the sans variants”. There are many such “font superfamilies”. If you’d like, I can draw up a list of recommendations. (It would be helpful if you could let me know your constraints w.r.t. licensing and budget.)
Theme variants do not have to be comprehensive redesigns. It is eminently possible to design a set themes that will not lead to the content being perceived very differently depending on the active theme.
P.S.:
I suspect the distinction you’re looking for, here, is between transitional serifs (of which Charter, the Medium font, is one, although it’s also got slab-serif elements) and the quite different old-style serifs (of which ET Book, the current LW 2.0 font, is one). (There are also other differences, orthogonal to that distinction—such as ET Book’s considerably smaller x-height—which also affect readability.)
Alicorn, if you’re reading this, I wonder what your reaction is to the font used on this website:
https://www.readthesequences.com
P.P.S.: It is also possible that the off-black text color is negatively impacting readability! (Especially since it can interact in a somewhat unfortunate manner with certain text rendering engines.)
Alicorn, what OS and browser are you viewing the LW 2.0 site on?
I do not like the readthesequences font. It feels like I’m back in grad school and also reading is suddenly harder.
I’m on a Mac ’fox.
Ok, thanks!
FYI, your assessment is in the extreme minority; most people who have seen that site have responded very positively to the font choice (and the typography in general). This suggests that your preferences are unusual, in this sphere.
I say this, not to suggest that your preference / reaction is somehow “wrong” (that would be silly!), but a) to point out the the danger in generalizing from one’s own example (typical mind blah blah), and b) to underscore the importance of user choice and customization options!
rest of this response is not specifically for Alicon but is re: this whole comment thread
This is still a gold standard of UX design: sane defaults plus good[1] customizability.
[1] “Good” here means:
comprehensive
intuitive
non-overwhelming (i.e. layered)
Note, these are ideals, not basic requirements; every step we take toward the ideal is a good step. So by no means should you (the designer/developer) ever feel like “comprehensive customizability is an unreachable goal; there’s no reason to bother, since Doing It Right™ is too much effort”! So in this case, just offering a couple of themes, which are basic variations on each other (different-but-matching font choices, a different color scheme), is already a great thing and will greatly improve the user experience.
The Medium font is much less bad but still not great.
Update: I have a recommendation for you!
Take a look at this page: https://wiki.obormot.net/Reference/MerriweatherFontsDemo
The Merriweather and Merriweather Sans fonts (available for free via Google Fonts) are, as you can see, designed to be identical in width, line spacing, etc. They are quite interchangeable, in body text, UI, etc. Both are quite readable, and aesthetically pleasing.
(As a bonus, active on that page is a tiny bit of JavaScript trickery that sets different body text font weights depending on whether the client is running on a Mac, Windows, or Linux platform, to ensure that everyone sees basically the same thing, and enjoys equally good text readability, despite differences in text rendering engines. Take a look at the page source to see how it’s done!)
UPDATE 2: A couple of mockups (linking to individual images because Imgur’s zoom sucks otherwise). Be sure to zoom in on each image (i.e. view at full magnification):
LW 2.0 with Merriweather:
One
Two
Three
Four
Five
Six
Seven
LW 2.0 with Merriweather Sans:
One
Two
Three
Four
Five
Six
Seven
Eight
Nine
When I compare the two examples, the second one feels “clear”, while the first one feels “smudgy”. I have to focus more to read the first one.
EDIT: Windows 10, Firefox 55.0.3, monitor 1920x1080 px
1. OS (and version), browser (and version), device/display, etc.?
(General note: folks, please, please include this information whenever you say anything to a web designer/developer/etc. about how a website looks or works for you!!)
2. Great! If one of them feels clear, then this goes to show exactly what I was saying: user choice is good.
Eliezer Yudkowsky
It may be worth saying explicitly that this is from 2009 and therefore can’t be talking about responses to “LW 2.0”.
Agreed, generally, it seems that sans serif are for screens, and serif is for print.
This is old “received wisdom”, and hasn’t been the case for quite a while.
Folks, this is what people mean when they talk about LessWrong ignoring the knowledge of experts. Here’s a piece of “knowledge” about typography and web design, that is repeated unreflectively, without any consideration of whether there exists some relevant body of domain knowledge (and people with that domain knowledge).
What do the domain experts have to say? Let’s look:
https://www.nngroup.com/articles/serif-vs-sans-serif-fonts-hd-screens/ (Note that this is from the Nielsen Norman group, making this almost literally the exact scenario I outlined in my comment on expertise.)
https://designshack.net/articles/typography/serif-vs-sans-serif-fonts-is-one-really-better-than-the-other/
https://designshack.net/articles/typography/web-design-debate-do-i-really-need-to-use-sans-serif-fonts/
https://www.quora.com/Why-do-most-websites-and-mobile-apps-use-sans-serif-fonts
But this domain knowledge has not, apparently, reached LessWrong; here, “Serifs on a screen are bad” and “sans serif are for screens, and serif is for print” is still true.
And now we have two people agreeing with each other about it. So, what? Does that make it more true? What if 20 people upvoted those comments, and five more other LessWrongers posted in agreement? Would that make it more true? What amount of karma and local agreement does it take to get to the truth?
Here’s what I think is the conventional wisdom about serif/sans-serif; I don’t think it is in any way contradicted by the material you’ve linked to.
Text that is small when measured in display pixels is generally harder to read fluently when set in a typeface with serifs.
Only interested in readers with lovely high-DPI screens? Go ahead, use serifs everywhere; it’ll probably be fine. Writing a headline, or a splash screen with like 20 words on it? Use serifs if they create the effect you want; the text won’t be small enough, nor will there be enough of it in a block, for there to be a problem.
But if you are choosing a typeface for substantial chunks of text that might be read on a not-so-great screen, you will likely get better results with a sans-serif typeface.
So, what about those domain experts? Jakob Nielsen is only addressing how things look on “decent computer screens with pixel densities of 220 PPI or more”. Design Shack article 1 says that a blanket prohibition on serifed typefaces on screens is silly, which it is. But look at the two screenshots offered as counterexamples to “Only use serifs in print”. One has a total of seven words in it. The other has a headline in a typeface with serifs … followed by a paragraph of sanf-serif text. Design Shack article 2 says that sans-serif typefaces are better “for low-resolution displays”, though it’s not perfectly clear what they count as low-resolution. The Quora question has a bunch of answers saying different things, mostly not from “domain experts” in any strong sense.
I like seriffed typefaces. In a book, sans-serif is generally hideous and offputting to me. On my phone or my laptop, both of which have nice high-resolution displays, Lesser Wrong content with serifs looks just fine. (Better than if it were set sans-serif? Dunno.) On the desktop machine I’m using right now, though, it’s ugly and it feels more effortful to read than the corresponding thing on, say, Less Wrong. For me, that is.
Yes. More precisely: the proposition we should actually care about here is not some broad generality about serif versus sans-serif typefaces, but something like “Users of Lesser Wrong will, on the whole, find it a bit easier on the eyes if content is generally set in sans-serif typefaces”. Consider the limiting case where every LW user looks at the site and says “ugh, don’t like that font, the serifs make it harder for me to read”. Even if all those users are shockingly ignorant of typography, this is a case where if no one likes it, then it is ipso facto bad.
Of course we don’t have (anything like) the entire LW community saying in chorus how much they dislike those serifs. But yes, when what matters is the experience of a particular group of people, each individual person who finds a thing bad does contribute to its badness, and each individual person who says it’s bad does provide evidence for its badness.
Karma is relevant here only as a proxy for participation. A crude answer to this question is: enough to constitute a majority of users, weighted by frequency of use.
In case I haven’t made it clear enough yet, I am not arguing that LW people are always right, or that high-karma LW people are always right. I am arguing that when the thing at issue is the experience of LW people, the experiences of LW people should not be dismissed. And I am arguing that on the more general question (are typefaces with serifs a bad idea on the web?) the simple answer “no; that’s an outdated bit of bogus conventional wisdom” is in fact just as wrong as the simple answer “yes; everyone knows that”.
Disagree. (Keep reading for details.)
This is still incorrect, because serif readability is superior to that of sans-serif, and see below for the matter of “not-so-great screens”.
Screen DPS
Given the pixel resolution per character you need to make serifs work, they are inferior on the screen… if you have a 72ppi (or less) display.
Now, such displays exist; here’s one. They are quite rare, though, and designed for entertainment, not work. The idea that any appreciable percentage of LW users have such hardware seems implausible.
On a ~96ppi display (such as this nearly decade-old cheap flat-panel I’m using right now, or indeed any display display made in the past 15+ years), the apparent (angular, a.k.a. “CSS reference pixel”) font size that you need to bring out the superiority of serif typefaces is no larger than the minimum size called for by other accessibility guidelines.
“The LW 2.0 font is less readable”
1. What OS is this on? If the answer is “Linux” or “Windows”, then part of the answer is “text rendering works very different on those operating systems, and you have a) test your site on those systems, b) make sure to make typographic choices that compensate, c) take specific actions to ensure that the user experience is adjusted for each client platform”. I of course can’t speak to (a), but (b) and (c) are not in evidence here.
2. The body text font size on LW 2.0 is too small (especially for that font), period. Again I refer you to https://www.readthesequences.com/Biases-An-Introduction; the body text is at 21px there. I consider that to be a minimum (adjusted for the particular font); whereas LW 2.0 (with a similar type of font) is at 16px. Yes, it looks tiny and hard to read. (But have you tried zooming in? What happens then?)
3. Other issues, like color (#444, in this case) affecting text rendering. I speak of this in my other comments.
“Consensus matters locally”
If every LW user looks at the site and says that, then we can’t still conclude anything about serifs from that, because if all of those users have not the smallest ounce of typography or design expertise, then they don’t know what the heck they like or dislike, serif-wise.
Let me be clear: I’m not saying that people can’t tell whether they like or dislike a particular thing. I am saying that without domain knowledge, people can’t generalize their preferences. Ok, so some text on their screen is hard for them to read. What’s making it so? The fact that a font has serifs? Or maybe just that it’s a particular kind of serif font? Or the font weight? Or the weight grade? Or the shape of the letterforms (how open the curves are, for instance, or the weight variability, perhaps due to which “optical size” is being used)? Or the color? Or the subpixel rendering settings? Or the kerning? Or the line spacing? Or the line length? Or the text-rendering CSS property setting? If you (the hypothetical-user you) don’t know what most or all of those things are, then sure your preferences are real, but your opinion (generalized from those preferences) is worth jack squat.
In other words: “if no one likes it, then it is ipso facto bad”—yes, but what, exactly, is “it”? You’re equivocating between two meanings, in that sentence! So, this is true:
“If no one likes , then is bad.”
Yes. Granted. But you seem to want to say something like:
“If no one likes , then are bad.”
But any particular thing belongs to many different classes, which intersect at the point defined by that thing! Obviously not all those classes are ipso facto bad, so which one(s) are we talking about?? We have no idea!
Dismissed? No. Taken at anything even remotely resembling face value? Also no.
Come on, folks. This is just a rehash of the “people don’t have direct access to their mental experience” debate. You know all of this already. Why suddenly forget it when it comes up in a new domain?
Do you have actual solid evidence for that? I’m guessing that if you did you’d have given it already in your earlier comments, and you haven’t; but who knows? (One of the answers to that Quora question mentions a study that found a small advantage for serifs. It also remarks that the difference was not statistically significant, and calls into question the choice of typefaces used, and says it’s not a very solid study. So I hope you have something better than that.)
Again, I would be interested in more information about what evidence you have about the font size required “to bring out the superiority of serif typefaces”. For the avoidance of doubt, that isn’t a coded way of saying “I bet you’re wrong”; I would just like to know what’s known about this and how solidly. I do not have the impression that these issues are as settled as you are making them sound; but I may just be unaware of the relevant work.
One instance is Firefox on Windows; the other is Firefox on FreeBSD (which I expect is largely indistinguishable in this context from Firefox on Linux). I concur with your guess that the people responsible for LesserWrong have not done thorough testing of their site on a wide variety of platforms, though I would be surprised if no one involved uses either Windows or Linux.
LesserWrong has what looks to me like a weird multiplicity of different text sizes. Some of the text is clearly too small (personally I like small text, but I am aware that my taste is not universally shared). However—and I must stress again that here I am merely describing my own experience of the site—if I go to, say, this post on the Unix box at my desk right now then (1) the size of the type at my typical viewing distance is about the same as that of a decently typeset paperback book at its typical viewing distance, and (2) I find the text ugly and harder to read than it should be because various features of the typeface (not only the serifs) are poorly represented—for me, on that monitor, after rendering by my particular machine—at the available resolution. (The text is very similar in size and appearance to that on readthesequences.com; LW2.0 appears to be using—for me, etc., etc. -- ETBembo Roman LF at 19.2px actual size, whereas RTS is using GaramondPrmrPro at 21px actual size. ETBembo has a bigger x-height relative to its nominal size and most lowercase letters are almost exactly the same size in each.)
Yup, agreed. But I would say the same about readthesequences.com even though its body text is black.
I agree. (Though it would, despite their hypothetical ignorance, be evidence. Someone who says “this text is hard to read because of the serifs” may be wrong, but I claim they are more likely to say it in the face of text that’s hard to read because of its serifs than of text that’s hard to read for some other reason.)
Perhaps I left too much implicit in my earlier comment, so let me try to remedy that. I firmly agree that the mere fact that some LW users believe some proposition about serifs in webpage text is perfectly compatible with the falsehood of that proposition. Even if it’s quite a lot of LW users. Even if they have a lot of karma.
But the thing that actually matters here is not the general proposition about serifs, but a more specific question about the type used on LesserWrong. I wasn’t equivocating between this and the general claim about serifs, nor was I unaware of the difference; I was deliberately attempting to redirect discussion to the more relevant point.
(Not that the general question isn’t interesting; it is.)
[EDITED to add:] Of course much of what I wrote before was about the general proposition. Whether I agree with you about that depends on exactly what version of the general proposition we’re discussing—I take it you would agree with me that many are possible, and some might be true while others are false. In particular, I am somewhat willing to defend the claim that there are otherwise reasonable choices of text size for which typical seriffed typefaces make for a worse reading experience than typical sans-serif typefaces for people using 100ish-ppi displays, and that while this can be mitigated somewhat by very careful choice of serif typefaces and careful working around the quirks of the different text rendering users on different platforms will experience, selecting sans-serif typefaces instead may well be the better option. I am also willing to be convinced to stop defending that claim, if there is really good evidence against it.
Not close at hand. You may reasonably consider my claim to be undefended for now. When I have the time, I’ll try to put together a bit of a lit survey on this topic.
Right you are. The 16px size is what I saw on the front page.
Even on my machines, ET Book (source) does not seem to render as well as Garamond Premier Pro (in a browser).
I think this is literally true but relevantly false; specifically, I think this is false once you condition on the cause of the text’s unreadability not being some gross and obvious circumstance (like, it’s neon purple on a fuchsia background, or it’s set at 2px size, etc.)
I think that someone who is ignorant of typography is no more likely to blame serifs in the case of the serifs being to blame than in the case of the text rendering or line length being to blame.
Noted. I was responding to the general claim.
As to the specific question, the matter of serifs is moot, because (as with all specific design decisions), each designer decision should be comprehensively user-tested and environment-tested, and as much user choice should be offered as possible.
Indeed.
Nope, the claim is reasonable. Websites where information density is more important than long-form readability, or where text comes in small chunks and a user is expected not to read straight through but to extract those chunks, may be like this. For that use case, a smaller point size of “body” text may be called for, and a well-chosen sans font may be a better fit.
LessWrong is not such a website, though a hypothetical LessWrong community wiki may be (or it may not be; it depends on what sort of content it mostly contains).
(Aside: I somewhat object to speaking of “typical” serif typefaces, because that’s hard to resolve nowadays. I suspect that you know that, and I know that, but in a public discussion it pays to be careful with language like this.)
However:
… is always advisable, regardless of typographic or other design choices.
I have no knowledge of typography, but was thought in University that serif fonts should be used for screens, and sans serif for print; it is very possible, that my lecturers were wrong.
No.
None. The truth is orthogonal to the level of local agreement. That said, local agreement is Bayesian evidence for the veracity of a proposition.
Mathematically, if the truth is orthogonal to the level of local agreement, local agreement cannot constitute Bayesian evidence for the veracity of the proposition. If we’re taking local agreement as Bayesian evidence for the veracity of the proposition, we’re assuming the veracity of the proposition and local agreement are not linearly independent, which would violate orthogonality.
Either I don’t know what Bayesian evidence is, or you don’t.
My understanding is:
Based on that understanding of Bayesian evidence, I argue that Lesswrong consensus on a proposition is Bayesian evidence for that proposition. Lesswrongers have better than average epistemic hygiene, and pursue true beliefs. You expect the average lesswronger to have a higher percentage of true beliefs than a lay person. Furthermore if a belief is consensus among the Lesswrong community, then it is more likely to be true. (A single Lesswronger may have some false beliefs), but the set of false beliefs that would be shared by the overwhelming majority of Lesswrongers would be very small.
That assumes that there is a statistical correlation between the two, no? If the two are orthogonal to each other, they’re statistically uncorrelated, by definition.
http://lesswrong.com/lw/nz/arguing_by_definition/
The local agreement (on Lesswrong) on a proposition is not independent of the veracity of the proposition. To claim otherwise is to claim that Lesswrongers form their beliefs through a process that is no better than random guessing. That’s a very strong claim to make, and extraordinary claims require extraordinary evidence.
“The local agreement (on Lesswrong) on a proposition is not independent of the veracity of the proposition.”
Sure, and that is equally true of indefinitely many other populations in the world and the whole population as well. It would take an argument to establish that LW local agreement is better than any particular one of those populations.
Then we are in agreement.
As for Lesswrong vs the general population, I point to the difference in epistemic hygiene between the two groups.
They were lecturers in what subject? Design / typography / etc.? Or, some unrelated subject?
Unrelated subjects (insofar as webdesign is classified as unrelated).
Well, in that case, what I conjecture is simply that either this (your university classes) took place a while ago, or your lecturers formed their opinions a while ago and didn’t keep up with developments, or both.
“Use sans-serif fonts for screen” made sense. Once. When most people had 72ppi displays (if not lower), and no anti-aliasing, or subpixel rendering.
None of that has been true for many, many years.
I am currently in my fourth year.
I have expressed this sentiment myself, so it is plausible.
Why? Are people around here more likely to agree with true propositions than false ones? This might be true in general, but is it true in domains where there exists non-trivial expertise? That’s not obvious to me at all. What makes you think so?
I was generalising from the above. I expect the epistemic hygiene on LW to be significantly higher than the norm.
For any belief b, let Pr(b) be the probability that b is true. Forall b such that b is a consensus on Lesswrong (greater than some k% of Lesswrongers believe b), then Pr(b) > 0.50 is a belief I hold.
But this is an entirely unwarranted generalization!
Broad concepts like “the epistemic hygiene on LW [is] significantly higher than the norm” simply don’t suffice to conclude that LessWrongers are likely to have a finger on the pulse of arbitrary domains of knowledge/expertise, nor that LessWrongers have any kind of healthy respect for expertise—especially since, in the latter case, we know that they in fact do not.
Do you suggest that the consensus on Lesswrong about arbitrary domains is likely to be true with P ⇐ 0.5?
As long as Pr(B|lesswrong consensus) is > 0.5, then Lesswrong consensus remains Bayesian evidence for truth.
For some domains, sure. For others, not.
We have no real reason to expect any particular likelihood ratio here, so should probably default to P = 0.5.
I expect that for most domains (possibly all), Lesswrong consensus is more likely to be right than wrong. I haven’t yet seen reason to believe otherwise; (it seems you have?).
Again, there is nothing special about this. Given that I believe something, even without any consensus at all, I think my belief is more likely to be true than false. I expect this to apply to all domains, even ones that I have not studied. If I thought it did not apply to some domains, then I should reverse all of my beliefs about that domain, and then I would expect it to apply.
I never suggested that there was anything extraordinary about my claim (au contraire, it was quite intuitive) I do not think we disagree.
Just so we’re clear here:
Profession (Results from 2016 LessWrong Survey)
Art: +0.800% 51 2.300%
Biology: +0.300% 49 2.200%
Business: −0.800% 72 3.200%
Computers (AI): +0.700% 79 3.500%
Computers (other academic, computer science): −0.100% 156 7.000%
Computers (practical): −1.200% 681 30.500%
Engineering: +0.600% 150 6.700%
Finance / Economics: +0.500% 116 5.200%
Law: −0.300% 50 2.200%
Mathematics: −1.500% 147 6.600%
Medicine: +0.100% 49 2.200%
Neuroscience: +0.100% 28 1.300%
Philosophy: 0.000% 54 2.400%
Physics: −0.200% 91 4.100%
Psychology: 0.000% 48 2.100%
Other: +2.199% 277 12.399%
Other “hard science”: −0.500% 26 1.200%
Other “social science”: −0.200% 48 2.100%
The LessWrong consensus is massively overweighted in one particular field of expertise (computing) with some marginal commentators who happen to do other things.
As for evidence to believe otherwise, how about all of recorded human history? When has there ever been a group whose consensus was more likely to be right than wrong in all domains of human endeavor? What a ludicrous hubris, the sheer arrogance on display in this comment cowed me, I briefly considered whether I’m hanging out in the right place by posting here.
Let B be the set of beliefs that are consensus among the LW community. Let b be any arbitrary belief. Let Pr(b) be the probability that b is true. Let (b|B) denote the event that b is a member of B.
I argue that Pr(b|B) (Probability that b is true given that b is a member of B) is greater than 0.5; how is that hubris?
If Lesswrongers are ignorant on a particular field, then I don’t expect a consensus to form. Sure, we may have some wrong beliefs that are consensus, but the fraction of right beliefs that are consensus is greater than 1⁄2 (of total beliefs that are consensus).
This entire thread is reason to believe otherwise. We have the LessWrong consensus (sans-serif fonts are easier to read than serif fonts). We have a domain expert posting evidence to the contrary. And we have LessWrong continuing with its priors, because consensus trumps expertise.
I’m not continuing with my priors for one (where do you get that Lesswrong is continuing with its priors?).
It is not clear to me that “serif fonts are easier to read than sans-serif fonts” was ever/is a consensus here. As far as I can tell, fewer than ten people expressed that opinion (and 10 is a very small sample).
1 example (if this was that) wouldn’t detract from my point though. My point is that lesswrong consensus is better than random guessing.
Now that I’m looking at it more closely: Quoted text in comments does not seem sufficiently set off. It’s slightly bigger and indented but it would be easy on casual inspection to mistake it for part of the same comment.
I think the font feels okay (though not great) when it’s “normal” writing, but text in italics gets hard to read.