Stop misuse of Markdown blockquotes for emphasis

As discussed in the comment to this answer, several users of this site use Markdown blockquotes (lines starting with >, like in an e-mail) to emphasize text. For instance:

Please stop using Markdown blockquotes like this.

That tag is converted to HTML's <blockquote>; both are semantic tag, that are supposed to signal that the text is a citation. Misusing tags can mess things up with web parsers, search engine crawlers, and screen readers for disabled people. While blockquotes may seem less dangerous to misuse that other tags such as code, I still think it's bad practice.

Moreover, If the default CSS is altered, they may end up not looking like users expect. For instance, an user here states that they look much different in the mobile layout, and they don't seem like emphasized text at all.

We have had a discussion on the same notes on meta.academia.se, and other people, including a W3C working draft, recommend against this misuse in HTML.

• Related older discussion on this meta: Is using blockquote for highlighting problematic? – Martin Sleziak Feb 25 '17 at 15:44
• Related question on Meta Stack Exchange: Do we need a new Markdown formatting for indented / boxed text (for preambles, remarks / side notes, postscripts, footnotes, …)? – user642796 Feb 26 '17 at 8:22
• Comments are not for extended discussion; this conversation has been moved to chat. – Daniel Fischer Feb 26 '17 at 19:29
• Could you give actual examples of messing things up? I tried my phones voice assistance, and screen reader (for the visually impaired) and it didn't treat blockquotes or bold any differently. – Paul Plummer Feb 26 '17 at 21:32
• @PaulPlummer An example of a mess-up is the answer in the first link in my question. Quid mentioned in the comments (now moved to chat) that he thought the first sentence was a citation, and tried looking for its source for a while. I was similarly confused. I don't have sufficient experience with screen readers to find an example using them, but the markup says "quotation", so I can easily imagine programming a screen reader to spell out "citation". Or reading it in Morgan Freeman's voice. :) – Federico Poloni Feb 26 '17 at 23:02
• Okay, although I think it was just bad formating which lead that conclusion, which I tried to express here. The last part makes me think everything should be in block quotes :) – Paul Plummer Feb 26 '17 at 23:10

I had already given my opinion on this in the Q&A Is using blockquote for highlighting problematic? Let me summarize my opinion.

• I consider it as problematic when a blockquote is used only to add emphasize.

• I consider it as tolerable, yet often not optimal, when a blockquote is used to mark a somewhat self-contained, independent part of the post.

I think I did and continue to do the latter occasionally. That is, I consider the notion of "quote" as a broad one. The borderline can anyway be hazy, especially given our context. Somebody might literally quote an exercise from some lecture notes, somebody else might have created their own version and 'quote' their private maybe even just mental notes.

• If block quotes looked like they do in this answer you would definitely have a point that it should not be used for emphasis. They way they are designed is universally recognized as quotes, and even if they were you should include information about who is being quoted. Links are universally recognized as... well links, and they are not expected to be their for emphasis (they do not emphasize anything except there is more information), and on top of that there are expectations there will be links in posts, but quotes are not all that common here – Paul Plummer Feb 26 '17 at 19:12
• since this is not a forum or a chat server etc. Chat has a quoting system which makes it apparent you are referencing say the OP or whatever, getting that implemented though to choose segments of a post may be complicated though. – Paul Plummer Feb 26 '17 at 19:13
• re "If block quotes looked like they do in this answer you would definitely have a point that it should not be used for emphasis" One of the points is they could reasonably look like this tomorrow, or when the source is reused elsewhere. re " Links are universally recognized as... well links, and they are not expected to be their for emphasis" Says who? I might think it is a convenient way to color text and I might want colored text. Clearly it is no actual link that should be followed so there is little risk of confusion. – quid Feb 26 '17 at 19:32
• re "(they do not emphasize anything except there is more information)," So you do not think the fact that the text is colored makes the text stand out a bit against the rest of the text? Really? – quid Feb 26 '17 at 19:33
• Regarding the final sentence, the average user's experience is going to be "This button does what I want to do. Someone told me not to use that button, but it's the only way I know to do what I want to do". So, you get resistance from people who decide doing what they want to do is more important than listening to a random person on the internet telling them they shouldn't do things. – user14972 Feb 26 '17 at 21:01
• @Hurkyl yes. But pretty sad. I am not sure why you defend this attitude. – quid Feb 26 '17 at 21:03
• @quid: My intention wasn't to defend it, but to emphasize it exists; to achieve change one must acknowledge how people actually act, not how you wish they acted. That said.... – user14972 Feb 26 '17 at 21:08
• It is a fairly defensible attitude; "random person on the internet" isn't a particularly reliable source of information. To give up something you value, one needs a reason to believe this particular random person is worth listening! – user14972 Feb 26 '17 at 21:09
• And on a selfish note, I resent situations where the "right" thing to do is a lost cause, so that the only effect of proscribing certain behaviors is to penalize well-meaning rule-abiding people without actually achieving any meaningful positive effect. – user14972 Feb 26 '17 at 21:13
• I put that badly -- it's that I hate when these ideas to achieve positive results are done in a way that negatively impacts the people who would cooperate without actually achieving a meaningful effect. (but this is maybe getting too chatty, so I'll stop now) – user14972 Feb 26 '17 at 21:21
• Re the 'random people' For not completely unrelated matters, such as use and misuse of MathJax, we sometimes have the lead developer of the MJ project weigh in. Still people don't care. In my opinion the main problem is actually that many dislike too much to admit to themselves they were wrong, and thus dig in to avoid this at all cost. – quid Feb 26 '17 at 21:28
• On the last two comments, I actually can connect to this. Yet, first I do not see what is the big drawback to avoid blockquote for pure emphasis. Second, if there was a clear policy on such things it would be simpler. For on thing, they would provide cover for those that edit out deviations. @Hurkyl (sorry forgot the ping in the preceding commment) – quid Feb 26 '17 at 21:32

This is not a realistic expectation.

In this setting, markdown isn't being used by web developers to generate html for accessible websites; it's being used by laypeople to format their posts.

I do not believe it is realistic to expect laypeople to think that this markdown feature is anything different from "Put this text in a box that stands out from the rest of the document" (*), nor do I believe announcements such as the one in the OP can make a significant impact on the behavior of the general population.

If it is important that <blockquote> is only used for things that truly are quotations, then one of the following needs to happen:

• This markdown feature should not be rendered via <blockquote>
• The UI needs to change so that users don't arrive at the idea that this feature is for (*)

The latter could be achieved, for example, by providing a markdown feature for (*) that is separate from blockquoting, and in such a way that users are led to use the right feature for the right purpose.

It's worth noting

• You can't use <blockquote> properly via markdown anyways, since it doesn't provide a way to add a citation or a footer
• The W3C link in the OP acknowledges

Editorial Note: We need to provide pointers to alternate ways to achieve the desired effect.

• Other features that are very clearly not meant to be quotations are also rendered using <blockquote>:

Check the source to see for yourself!

• Why would the icon for "Put this text in a box that stands out from the rest of the document" be a quotation mark? Why would it use the same syntax used in emails for quotes since a very long time? It is not that all the people don't know, it is that many people don't care. This was clearly documented in the comment discussion. To pretend otherwise is borderline dishonest. – quid Feb 26 '17 at 20:46
• @quid: Because quotations are an example of something that could be formatted that way, or maybe simply because someone sometime decided to give the name "blockquote" to this kind of formatting. Why would anyone believe that this particular formatting should only be used for that purpose? This point was also brought up in the comments; to pretend otherwise is borderline dishonest. – user14972 Feb 26 '17 at 20:50
• @quid: Note, incidentally, that the button (and hotkeys!) for emphasis and strong are very, very suggestive of "use this button for italics" and "use this button for boldface", so these already contradict that point... not that "the picture on a button is a literal description of what the feature should only be used for" argument really held water in the first place. – user14972 Feb 26 '17 at 20:52
• "Why would anyone believe that this particular formatting should only be used for that purpose?" For instance, because they are told it is like this. Their reaction is however not: "Oh I did not know that. will do so in the future." But something else entirely. Many people don't care about using features correctly. On your second comment: yes, I had mentioned this in chat earlier as somewhat problematic. – quid Feb 26 '17 at 20:55
• @quid: Like I said in my answer, I do not believe public service announcements are effective; you might reach a small segment of the population (and convince a smaller segment), but as a whole this is the sort of behavior that can only reasonably be addressed as an interface issue rather than expecting people to read the manual. – user14972 Feb 26 '17 at 20:56
• Nothing would be effective when users don't care. They'll also use MathJax in the most obscure ways possible to achieve one formatting feature or other. Especially color, many colors. – quid Feb 26 '17 at 20:58
• @quid: I'm pretty sure a feature that does what they are trying to achieve without abusing blockquote would be very effective at reducing the misuse of blockquote. (especially if the new feature is easier to discover and use than the one using blockquote) Visual cues in the formatting and/or input that scream "this is a quote" might be somewhat effective too. – user14972 Feb 26 '17 at 20:59
• I asked somebody twice what feature they want, they would not even answer the question. – quid Feb 26 '17 at 21:00
• So what if we change the rendering of > to something that stands out less instead of more? For instance, light gray text set in small font. That would solve the problem? Or do you have any objections to that? – Federico Poloni Feb 26 '17 at 22:54
• @Federico: I'd be disappointed, but at least it would make sense and have a chance of working. It's a tricky problem to do something like that, though, since (1) it still has to be effective as a citation, and (2) you run the risk of it meeting some other formatting need -- e.g. what you suggest might be just perfect for formatting a humorous aside/afterthought, or to get the effect of mumbling or muttering. – user14972 Feb 26 '17 at 23:05

This is no less a problem than using codeblock for the same thing.

Emphasis and strong emphasis (or bold) already exist. The header formatting already exists. People should use those appropriately.

If you need something highlighted that badly, there are better ways to do it. If you need something separated that badly, use paragraphs.