Writing for the Mundane

part of the “what I learned after four years at MIT” series, I guess?

I hated timed essays in high school. It’s pretty clear if you skim that part of the blog archives:

My [SAT] essay got a 10 out of 12. It’s an essay I’d be ashamed of posting anywhere else; it’s disgustingly traditional and formulaic. […] This was simply because I knew that using my normal essay-writing mindset, I’d get maybe a 3, because I’d spend the first twenty minutes debating myself over which side I was on and rewrite the introduction ten times. Too bad. I wasn’t there to write a good essay; I was there to get a good score on the SAT.

The worst issue is that students do not need to give in-depth explanation of anything they learned. Due to its stringent time limit, the essay portion rewards quick reckless writing much more than deep thought. […] encouraging students to practice writing 25-minute essays in order to improve their college-bound skills is like encouraging people to play Grand Theft Auto in order to improve their driving skills.

I took the Grand Theft Auto analogy pretty far. I’m not proud enough of these posts to link them, but you can find them if you try. In short, past me thought timed essays rewarded writing too quickly, with a disregard not only for facts but for the opportunity to lay out your opinions and thoughts so you could clarify and revise them, which was actually the most valuable part of writing.1 I conceded that the time limits made sense as a practical concession to allow you to test students’ writing skills fairly,2 but I felt like there was still way too much emphasis on the raw speed.

While I think these complaints are still valid, the complaints I’d make today have an orthogonal focus, because past me didn’t understand the kind of writing that I’d mostly be doing, nor the kind of mistakes I’d be seeing myself and others make in this kind of writing. There’s a lot of straightforward informative writing and day-to-day communication to be done where you don’t need to debate yourself or think deeply over which stance to take: setting up appointments, asking for logistical things from professors, answering standard factual questions about something you’re organizing, and so on. In fact, there’s so much of it that the mechanical component of writing quickly — the ability to nail a sentence’s grammar and word choice on the first try, and repeat several dozen times — is probably more useful than I would have given it credit for being, and that’s a skill that you can drill through timed essays.

Still, that doesn’t mean that those timed essays were good at teaching all the other skills to perform this kind of writing well, and I think the skill ceiling of this type of writing is much higher than I would have expected. Here is the advice I would have given to my past self about it.


  • The most important feature of writing is clarity. Get across what you mean clearly and precisely. Leave no room for misinterpretation. Your reader is no longer a teacher who will read what you wrote, try to figure out what you are trying to say, and tell you how to better say it (while taking off points). They will just read what you wrote and trust their first interpretation. Furthermore, some readers may not have English as their primary language, which means figures of speech and idioms that might be extra points in English class can backfire.3

  • The second most important feature of writing is conciseness.4 As Strunk & White famously said, “Omit needless words.” Every extra word is an additional chance to lose your audience’s attention.

    One of the sad facts of life that you will learn working in front-end software — or, really, in just about any public-facing role — is that people don’t read. It’s easy to be misled into thinking otherwise in school, where teachers are your captive audience and read everything you write as part of their job. But in real life, people will “read” your signs or emails or website and then ask you questions that you answered literally verbatim. (I’ve done this before and it was my fault.) The more words you write, the more words they won’t read. So why bother? Keep it short.

    If you can’t get around having a lot to say, say the most important things as soon as possible to maximize the chances that people will stop reading only after they get to them.

  • Follow standard English rules — spelling, grammar, punctuation — insofar as you have nothing to gain from not following them.

    This one requires some unpacking. English grammar is basically a giant Ponzi scheme. You should follow rules to achieve clarity and conciseness, but as long as you don’t confuse your man-eating shark with your man eating shark and don’t describe a panda who eats, shoots, and leaves, the main reason you should follow grammar rules is just that everybody else does. If you deviate from standard English, it will stand out, which is usually bad because it distracts from the meaning of the sentences you’re writing. For some writers in some contexts, deviating helps their goals;5 unfortunately for the grammar-anarchists among us, I don’t think the kind of day-to-day writing I’m primarily talking about is usually such a context. But it’s not impossible that it might be.

    Also note the corollary: if you use a weird construction with grammar that you had to double-check to get right, it can be just as bad as a straight-up grammatical mistake. It’s no solace that it was technically grammatically correct if it still distracted the reader. If your sentence is long and has dangling parts that are technically attached correctly, or if there’s something that looks like a subject-verb disagreement but technically isn’t because of some word ten words ago, you may still want to rewrite it until it doesn’t.

  • Beyond these rules, nobody cares about your small word choice decisions, your literature-class deliberations between words in the thesaurus over their subtly different connotations. Again, they’re not your English teacher. You’re not trying to score points for flowery language that rolls off the tongue and evokes the perfect mental imagery.


Some specific writing tips in specific scenarios, the chess tactics to the above strategy:

  • One of the most important areas you should be sure you’re writing maximally clearly is when you’re specifying a time. There are a staggeringly high number of things that can go wrong:

    • Many issues can arise when you describe a time relative to when you’re writing. The word “tomorrow”, for example, is ambiguous in messages sent between midnight and 6 AM or so, as it could mean “after the upcoming midnight” or “after the next time I go to sleep and wake up”.6 Also beware if you’re writing in a medium that doesn’t make clear when you’re writing — there are a lot of chat apps and fora where the date when something was posted is easy to overlook. What if due to client misconfiguration or internet hijinks, your email is only delivered after a few hours? I’ve even gotten emails timestamped from the future. When in doubt, also state the day of the week or of the month.

    • But phrases like “next Friday” are ambiguous too. If it’s Monday, is next Friday in four or eleven days? Reasonable people will disagree. There’s quite a linguistic rabbit hole to go down here. Some phrases are less ambiguous sometimes, perhaps “this Friday” or “the Friday of this/next week”, but even those will not save you all the time: when is the first day of the week? When in doubt, also specify the day of the month. (But make sure you get it right and that the day you named is in fact that day of the week.)

    • In addition, the words “after” and “before” when referring to a day can also be ambiguous. Does “after September 1st” include September 1st? There are many ways to patch this: use a construction like “on or after September 1st”, or add in an “inclusive” or “exclusive” in parentheses.

    • Speaking of inclusive/exclusive issues, there’s a dedicated Wikipedia section covering the confusion at noon and midnight of the 12-hour clock. The first problem, in short, is that because “a.m.” and “p.m.” basically mean “before noon” and “after noon”, it’s very unclear what “12 a.m.” and “12 p.m.” mean. It appears to be more common that the former refers to midnight and the latter refers to noon,7 but this is not something you should take for granted. It’s better to just say “noon” or “midnight”.

      The second problem is that it’s not clear which day a particular midnight belongs to. Is the midnight of September 1st right before 12:01am September 1st, or right after 11:59pm September 1st? I think the former opinion is slightly more common still, but I would equally not recommend that this be taken for granted. The simplest fix here is to let go of a single minute of precision and say either 12:01am or 11:59pm instead. (This is why so many assignments in college are due at the latter time. I consider this a good thing.) If a minute is too much, you can go to 12:00:01am or 11:59:59pm and so on, but those cases are rare.

    • There are even more pitfalls when you communicate across time zones. Whenever you’re specifying a time to somebody who might not be in your time zone, you should also mention your time zone explicitly. If you have extra time (harhar) and know the time zone of the recipient of your message, also convert your time to their time zone for them. Of course, if you’re going to do this conversion, make sure you do it correctly; WolframAlpha can help you with time zone math. If you really want to go the extra mile, create an event on timeanddate.com or something similar and give them a link; if they visit the page, it will convert its time to their local time zone on their device. (But don’t delete your textual description of the time when you add the link, in case they aren’t able to load the site for whatever reason.)

    • And then there’s also daylight savings time! It can mess up your timekeeping with time zones in lots of ways that are obvious, but usually those effects are limited to the few days before and after it starts and stops, assuming that you always consult the Internet for time zone math. However, one year-round issue is forgetting whether daylight savings time is in effect when you write the time zone — the U.S. east coast, for example, follows either Eastern Standard Time (EST) or Eastern Daylight Time (EDT). I’ve memorized “Spring forward, fall back”, but I still don’t know whether daylight savings time occurs from when we spring forward to when we fall back or vice versa. I grew up in a country without daylight savings time and used to think that that was just a disadvantage I’d have to learn to overcome, but since then I’ve watched plenty of people who have grown up in the U.S. get it wrong also. The solution is actually simple, and one of the few cases where I’d advocate for being less precise: drop the middle letter. Writing “ET” (or “PT”, or “MT” or “CT”) is no less clear and eliminates this failure mode.

  • Probably only slightly less dangerous is when you’re specifying a place. One of the worst ways to start a group meal is to agree on a restaurant, but then have members of your group arrive at different faraway restaurants with the same name. It’s always safer to give a street address they can put into a map or ridesharing app.8

    Sometimes, even this isn’t enough: if there’s a large building with multiple exits at the address, you likely also want to specify which exit using nearby landmarks. Sometimes you want to specify a location in an even larger area without addresses, say a park… or a college campus. I don’t have general advice for all scenarios here. Use your best judgment.

    I will also say that there’s a completely different trap you might fall into when specifying a location in a place like a college campus, where you may spend years navigating and familiarizing yourself with. After those years, you might tend to describe locations in ways that other people can understand only if they are similarly familiar with the region. This is great when you’re communicating with such people, because taking advantage of shared knowledge can allow you to be even more concise without losing clarity, but clarity can greatly suffer if you try to communicate a location in the same way to somebody who is new to the region. (MIT’s building and room numbers, for example, are concise and precise but legendarily confusing to people new to them.) Remember your audience and adjust.

  • Shifting to more broadly applicable tactics: You are allowed to use more formatting than you were in a formal school essay, and you should. Use bulleted or numbered lists to outline your email, give it structure, and show what’s related to what. Bold or italicize important text. Again, given that people won’t read your email, bold some words so you can control which words they won’t read. Don’t go overboard, because if everything is emphasized, then nothing is emphasized; but emphasizing a few parts helps people focus on them immensely.

    One caveat: if you’re writing on a platform that doesn’t support text formatting such as Twitter, think twice before faking it with Unicode bold or italic letters. It’s terrible for accessibility because that’s not what those characters mean. The letters may not render for some viewers and will be pronounced unintelligibly by screen readers.

  • For the love of Zarquon, if you’re writing an email, take the time to come up with a good subject. If in doubt, put more detail in the subject. Put the most important words first; grammar and sentence structure is less important. Keep it telegraphic, maybe in the style of newspaper headlines. I’ve never seen a subject line that was too long, although I might have written one with 19 words.

    Your subject should establish two things:

    1. How the email is relevant to the recipient. Is it related to a particular class? An organization the recipient is part of? A specific event? Then name it. Is it just related via a shared interest or something similar? Then name the shared interest. You may be able to skip this step if the recipient knows you. On the other hand, if the recipient doesn’t know you, including your name or their name in the subject does not establish relevance. I won’t fault anybody for including the recipient’s name — I get it, it feels more polite and personalized. But it’s the kind of politeness that’s devoid of information. Politeness is not a substitute for getting to the point. The subject “Hi Brian” provides basically the same amount of information as not having a subject at all.

    2. What category of actions you want the recipient to take in response to the email. If you want a question answered or request fulfilled, you could include the word “Question” or “Request”, but it’s often even better if you put a very abbreviated form of the question or request in the subject: “Pset 4 due date?” “Extension for pset 4?” In the case when you’re requesting an extension, it’s a safe bet that you’d like to receive an extension, but with other nouns it may not be clear and you may want to add a verb. “Website question” and “Website request” are better subjects than “Website”, but both could likely use more words. Better question subjects: “Software used to create website?” “Website color scheme?” Better request subjects: “Create website?” “Edit website?” “Republish website content?” A lot of these subjects are still ambiguous, but they do give the reader a general of the type and the magnitude of the requested actions. Republishing something on a website sounds a lot easier than creating a website from scratch, whoever that website is for.

      If you’re just informing them about something and don’t need a response, say the thing. If you’re providing a more specific kind of information, e.g. a receipt, say so. If there’s a poll or survey or application to fill out, say so. If there’s an event in the future, mention the time, as well as the place if there’s room. (You don’t have to be maximally precise in the subject. That’s for the body.)

    On the other hand, don’t put stuff like [ACTION REQUIRED] unless there will be serious negative consequences if they don’t do the action. No, it’s not a serious negative consequence if they miss their opportunity to respond to your survey or participate in your event and haven’t previously signed up for it.

  • If in doubt, start your email with “Hi,”. That’s “Hi” and then a comma. Sure, there are more formal greetings that will be more suitable in many contexts. But if you need to spend more than ten seconds deciding what greeting to use (typically, trying to decide if you’re on a first-name basis with the recipient), your time and effort are better spent making the rest of the email respectful, both in tone and in terms of how much time the other person will need to read it. See the rest of this list.

  • If in doubt, say thanks. Don’t overdo it because again, at some point too many words will either waste the other person’s time or distract them from reading the rest of what you’re writing — one thank-you at the beginning and one at the end is probably as much as you need. (Unless you’re writing a thank-you note, in which case, thank away.) The more specifically you can tell the other person why you are thankful for them or what they’ve done, the better, but this is of course also a tradeoff against conciseness.

    An even more specific consequence: If you want to say “thank you”, don’t say “sorry”. If you’re responding later than you intended but it’s not too bad of a delay, consider saying “Thank you for your patience.” instead of apologizing. (And also, don’t overestimate how responsive competent adults are to emails. People do not grow up and start magically responding to all emails within a few hours, or even one business day. It’s a fairly universal struggle.)

  • If you’re setting up an appointment with someone, like for an interview, give a few specific times if at all possible. You can say “I’m free at so-and-so time, but am flexible.” Some people are much busier than others.

    It’s much easier to look at a date and time and say “yes, that works” or “no, that doesn’t” than it is to respond to a vague request for an appointment listing times you’re available. This is especially true because if they list a lot of times, they might be committing to being available for all of those times, and might have trouble scheduling something else that comes up if you don’t get back to them immediately. Maybe you will get back to them immediately, but they don’t know that.

    On the other hand, it’s better to try to make an appointment with vague times when you’ll be available than to… not try.

That was a lot of words — I make no claim that I am good at following my own advice — but I want to conclude by saying that the biggest mistake you can make is not writing at all. Writing is hard, and nobody will write what they mean with perfect clarity every single time. But it’s far worse to assume that what you have to say isn’t worth saying — that the thing you noticed is slightly off or unclear isn’t worth bringing up, that the thing you’re asking for isn’t possible to grant, that the answer you’re trying to give isn’t complete enough. You never know. Some chance is better than no chance. Something is better than nothing.

Anyway, that’s a pretty good reason for me to publish this post now.

  1. Similar takes: Ben Kuhn, Evan Chen. I think the point has already been made well, and won’t dwell on it.

  2. On the other hand, apparently these time limits are circumventable. Here’s The Atlantic on the college admissions scandal earlier this year:

    When I began the job, the SAT and the ACT offered extended-time testing to students with learning disabilities, provided that they had been diagnosed by a professional. However, an asterisk appeared next to extended-time scores, alerting the college that the student had taken the test without the usual time limit. But during my time at the school, this asterisk was found to violate the Americans With Disabilities Act, and the testing companies dropped it. Suddenly it was possible for everyone with enough money to get a diagnosis that would grant their kid two full days—instead of four hours—to take the SAT, and the colleges would never know. By 2006, according to Slate, “in places like Greenwich, Conn., and certain zip codes of New York City and Los Angeles, the percentage of untimed test-taking is said to be close to 50 percent.” Taking a test under normal time limits in one of these neighborhoods is a sucker’s game—you’ve voluntarily handicapped yourself.

  3. Incidentally, for the cynical, consider that you may also be writing for lawyers, judges, and jury members in the future. Context or sarcasm that’s obvious to you, and that you think is obvious to your readers, likely won’t be to them.

  4. Ironic to say on this blog, I know.

  5. Some deviations make a point (e.g. satiric misspellings). Some deviations try to shape the language to improve it (the most prominent examples I can think of here are in gender-neutral language, e.g. the term Latinx or Spivak or other gender-neutral pronouns.) Some deviations augment the language to express certain concepts more concisely — this is usually what happens when people coin terms, but more interesting examples include e.g. the fandom exclamation mark and arguably all emoji. (There doesn’t seem to be an established name for this usage of the exclamation mark. TVTropes calls them characterization tags (obviously, warning: TVTropes link). To me, the essence of the construct is that it concisely emphasizes that it’s a different character or a differently-characterized character, instead of the same character in a different circumstance, which would be my interpretation if you just prepend an adjective to a character with a space in between.) Some deviations are part of an author’s style and help differentiate them or set the tone in some way; examples include E. E. Cummings, Cormac McCarthy, everybody in Homestuck (?)… probably if I were at all a real literature student I would have more examples. (Should E. E. Cummings’s name be spelled in lowercase? Many people do this, but Wikipedia thinks not. And here’s a copied citation.)

    Incidentally, while we’re talking about deviations from “standard English”, I do want to mention African-American Vernacular English (AAVE), although I don’t have any actionable advice about its use. AAVE and other English dialects are sometimes looked down on as grammatically sloppy, but in fact,

    …like all dialects, AAVE shows consistent internal logic and grammatical complexity, and is used naturally by a group of people to express thoughts and ideas. Prescriptively, attitudes about AAVE are often less positive; since AAVE deviates from the standard, its use is commonly misinterpreted as a sign of ignorance, laziness, or both.

    Just look at the table of verb tenses in the article. It may be easier to think of AAVE as a separate language that happens to share a lot of vocabulary with English. It should definitely not be considered as merely the result of taking American English and adding lots of grammar mistakes. Don’t be a prescriptivist.

  6. This is not a hypothetical, by the way — I have gotten, and misinterpreted, an email sent at 4:30 AM referencing “tomorrow”.

  7. This also happens to be consistent with the type of half-open intervals that many programmers will be familiar with; see the classic article on numbering by Dijkstra.

  8. Even this is not foolproof — some street addresses even repeat within the same city. Boston has multiple Church Streets, Tremont Streets, and Washington Streets (among other street names), and these streets even have overlapping numbers, so sometimes you even need to specify the neighborhood or zip code to unambiguously specify a location. Fortunately I’ve never personally experienced any issues arising from these ambiguities.

(note: the commenting setup here is experimental and I may not check my comments often; if you want to tell me something instead of the world, email me!)