Hacker News new | comments | show | ask | jobs | submit login
Textbook manifesto (2016) (greenteapress.com)
369 points by Tomte on Apr 9, 2017 | hide | past | web | favorite | 156 comments

In many European countries, this problem was resolved by what I feel is mostly student pressure. Our student union (for lack of better word) owned printing equipment and worked with most professors to do exactly what's suggested in this article: most professors wrote their own books (not 140 pages, though). Most of my textbooks were between 2 and 7 EUR, which I'm led to believe is approximately at cost. Occasionally, a particular textbook was "recommended", but there would always be ample library copies available, and often you wouldn't _really_ need it. I'd have about 4-6 courses per semester, so I'd spend maybe 25-30 EUR on our own textbooks. Occasionally I'd have to shell out for a traditional textbook, and that would utterly dominate that semester's materials budget.

The future's already here, it's just not evenly distributed.

Italian, and fellow european here. When I started university (1999, I think) there were still plenty of Big Expensive Books (I recall my Chemistry one was like that).

The "student pressure" materialized in the form of doing illegal copies of the books. This used to be totally illegal, but incredibly easy.

At some point the legislative environment changed so that photocopying up to a decent percentage of the books was considered legit. Suddenly, low price books started to bubble up everywhere.

So, if you don't have this yet, start lobbying your parliament for a "fair use" textbook copying policy, publishers will follow.

In the US it feels to me like it's gone the opposite way. The textbook publishers are increasingly able to block people from using pirated books, or even buying used books, by including one-use online codes in the books. Then they can raise the prices because there's no other market.

Don't ask me what I think the punishment should be, for devising this system, or for a professor requiring the use of such a thing in a course.

My wife works for a major textbook publisher. Not all those free textbooks or cheap textbooks are the same quality as the ones produced by big evil corporations. Frequently, when looking at competitors books, she will find more typos, fewer citations, out of date studies with small sample sizes, poor quality graphics, and sometimes just wrong information. Big books are expensive because the companies frequently worry a lot about these things and also spend lots of effort on the pedagogy of the text.

However, management is not good, the commercial model frequently prioritizes things which are not good for students, and the industry is still shedding revenue. In an effort to change that, they seems to be doubling down on digital products, which are still less popular than print by several orders of magnitude.

Looking back, I can think of one or two professionally-done textbooks I used that were really good. (The Campbell biology books in particular were stellar.) The amount of work that must have gone in to their creation was enormous, and I am thankful for them.

But there were not enough of them. For every worth-the-purchase, there were a dozen of middling to genuine crap quality, and every single one was painfully expensive.

I genuinely feel for the good folks that I know are stuck in the system. But the current model can't die fast enough.

Thank you.

Was in the students union for my faculty. We pulled out all stops to make things as efficient and cheap as possible(we actually sold them below cost because we found other ways to subsidise them). 10ish euro maximum performance semester for all the textbooks that were used in the classes, and printing thesis was 50 % of the next cheapest option... if we had doubled our prices.

We also did "university politics", basically creating pressure for professors to make good textbooks. I typeset one myself for a professor who wanted to "beat" another one.

Join your students union, union or local cooperatives and cooperative oriented parties people. Infrastructure and necessities should be run like that or by government institutions.

>We also did "university politics", basically creating pressure for professors to make good textbooks. I typeset one myself for a professor who wanted to "beat" another one.

Out of curiosity, how would his textbook beat another colleague's?

Popularity, I assume; which is probably why he put "beat" in scare quotes.

Nicer diagrams, better structure, better writing, students being happy with the script, students grades

Whenever I was in an international class, the study language was english, the textbooks were American. Extremely dense and horrible to read tombs. In my early courses most of the material was in swedish. Textbooks were a tenth of the size and had at least some personality. Some of them were quite enjoyable. Those were readable.

I think the meaning is more accurate with tombs.

Thanks for the correction. Seems it's too late to edit though. Guess the somewhat suitable typo will have to stand. :)

As a fellow European, my experience is similar (though our prices were higher but still substantially lower than the big-name commercial publishers.), but I suspect it would never fly in the US (communism, boo!).

Having been a US citizen for all of my 60 years, and often wondering "why things so fked up?", I've come to understand or believe this about the US:

If it makes money, it's good, regardless of whether it's good or bad for people or the environment. Once something is making money, it becomes a god-given right (cheap labor, insurance companies as health-care middlemen, etc).

If you propose something that inhibits the above, you will be opposed with all available resources by elected members of government at every level. That people are the ones suffering and electing is a mystery that I don't think I have enough remaining years to understand.

Being opposed by elected officials is one thing. What truly baffles me is being opposed by the very people that are adversely affected.

Many policies are literally convincing poor families that it is in their best interests to help support the frail legacy of billionaires. How!?

Many policies are literally convincing poor families that it is in their best interests to help support the frail legacy of billionaires. How!?

In the US, poor people don't see themselves as poor. They see themselves as not-yet-rich (or, at least, not-yet-middle-class). Therefore people oppose taxes on "the rich" because they think that those taxes might one-day affect them, even if that expectation is unrealistic in the extreme.

> In the US, poor people don't see themselves as poor. They see themselves as not-yet-rich

This is repeated so much that I'm starting to distrust it...

Maybe the people don't agree with you about what's in their best interests. You may believe that getting money for yourself regardless of any other consideration is the "correct" way to vote, others may disagree.

Poor people are lifted out of poverty when they get jobs. Jobs are created by incentivizing job creators.

> Poor people are lifted out of poverty when they get jobs.

Poor people are lifted out of poverty when they get the tools necessary to secure and maintain sufficient initial and recurring quantities of material assets; jobs with sufficiently high pay do that (but not all jobs have sufficiently high pay), though they aren't the only thing that could.

> Jobs are created by incentivizing job creators.

No, jobs are created by paying people to do work. There are several ways government can cause this to happen: the simplest and least prone to be misdirected is to simply pay people to do work. Among the more complex, less reliable, and more likely to be manipulated to maximize profits to other people and minimize jobs to the people for whom the goal is to secure jobs is creating financial incentives to capitalists in the hopes that it will result in them creating jobs that lift people out of poverty.

Different jobs have different levels of net benefit to society.

For example, paying someone to "dig a ditch and fill it in" gives that person work, but is a zero-sum game - transferring wealth between people, but not increasing the total pie (total well-being), at least without considering the psychological benefits of work (and exercise!).

Similarly, breaking windows to create jobs for window repairmen creates jobs, but is negative-sum - it benefits window repairmen, but reduces the overall pie. You're better off just giving money directly to window repairmen.

In contrast, some jobs increase the total wealth, and can benefit both the worker and the people willing to pay for the work to be done. (positive sum).

Some positive sum jobs increase the total wealth more or less than others, for the amount of money invested (spend $, increase total wealth .01% or 1%?). Similarly, the net gain of the work may be distributed differently (does the purchaser receive more of the surplus or the provider?)

Focussing purely on jobs ignores the cost - is society better off if I employ 100 people to do the work that 1 person could do?

The best way we've found to increase total well-being is to allow people to pay for what they want, rather than having government mandate jobs to be done.

> The best way we've found to increase total well-being is to allow people to pay for what they want, rather than having government mandate jobs to be done.

The widespread unrest with classical, 19th Century, capitalism​ and it's displacement throughout the developed world with modern mixed economies is a direct result of the untruth of this claim.

A lot of evidence seems to point towards the best way being a combination of the two.

No. Jobs are created from demand. There will be no cars sold in a neighborhood, and therefore no car dealerships, unless the people there have money to buy a car.

If there are people there to buy a car, or a gallon of milk, then "job creators" will compete against each other to hire enough people to satisfy that demand, and take a cut of the revenue.

No one creates a job. They satisfy a demand. Steve Jobs supposedly said something like we didn't know we wanted iPhones until we saw them. That may be true, but lucky for Apple that there are enough people with disposable income to buy those iPhones.

Demand comes first, then jobs to satisfy the demand. And if a "job creator" could satisfy that demand without creating jobs, he'd do it. "Job creators" are not in the business of creating jobs, they're in the business of making money, however the current economic environment allows and demands.

The only person who creates a job is the guy buying a pack of cigarettes while filling up his car.

> No. Jobs are created from demand. There will be no cars sold in a neighborhood, and therefore no car dealerships, unless the people there have money to buy a car.

True. But when people want a car, they do not automatically fall from the sky.

> If there are people there to buy a car, or a gallon of milk, then "job creators" will compete against each other to hire enough people to satisfy that demand, and take a cut of the revenue.

And that is how they become rich. People get a car, or a gallon of milk. A few people get jobs selling cars or gallons of milk. Someone gets rich by opening a store or a car delership.

> Demand comes first, then jobs to satisfy the demand. And if a "job creator" could satisfy that demand without creating jobs, he'd do it. "Job creators" are not in the business of creating jobs, they're in the business of making money, however the current economic environment allows and demands.

And when we have all our needs met without people having to work, with machines taking care of all our necessaities, we can think about Universal Basic Income. Until then, enterpreneurs will have to hire people to satisfy demand.

One of the biggest job creators is the government, with their massive budget and public needs. Yet we are often lead to believe that we need to coddle billionaires so that they feel comfortable enough to lay a few more jobs, and starve the federal government of revenue and responsibility.

Indeed. The government, like poor people, spend every dollar they take in, which allows that money to circulate. To use a cliche, a billionaire can only buy so many yachts, pieces of art or gold bricks; at some point, the excess is going to land in an offshore account or otherwise be underutilized.

Billionaires grudgingly create jobs.

Every dollar given to the billionaire is willingly given because the product benefits the person at least a dollar (net benefit to both parties). "The excess" is invested in other companies that produce other things people are willing to pay for.

Every dollar the government spends was taken from a taxpayer (cost one dollar), to provide a product people are not willing to pay a dollar for. This yields a net loss to society, unless there are positive externalities.

This does not mean that there is not a role for government spending (a "common good" in the economic sense is undervalued in the market).

Voluntary exchange produces an economic surplus, increasing net well-being. Government (compulsory) spending hopes to create an economic surplus, but is largely not measurable.

Sometimes the aim of government spending is simply transfer of wealth, which is zero-sum (or maybe negative sum, if it damages incentives to produce, or maybe positive sum if the recipients can use it to greater net benefit). Again, hard to measure.

>Every dollar the government spends was taken from a taxpayer (cost one dollar), to provide a product people are not willing to pay a dollar for. This yields a net loss to society, unless there are positive externalities.

Except this is not entirely true. Much of modern income is rent, and people are willing to buy things like defense, and much of the "dead weight loss" you are talking about is offset by collective barganing power.

Edit: Like you say, it's complicated and I'll be the first to admit don't understand how it all works but I don't think soaking the rich works well either, in theory or in practice, and government is often unfairly maligned by people who have something to gain by less government involvement.

They're also created by removing roadblocks for job creators to create jobs. Think about all the people who were locked into jobs because switching or going out on their own would require unaffordable (or perhaps unobtainable) health insurace for themselves or family members with preexisting conditions.

Poor people are lifted out of poverty when they get jobs.

1) A bit over a third of poor people (in the US) do work: https://www.bls.gov/opub/reports/working-poor/2014/home.htm

2) Reducing poverty can work by either raising the baseline (stronger social safety net, maybe higher minimum wage, better overall vocational education, etc) or by moving individuals into other economic classes (getting people (better) jobs). I think raising the baseline is probably a better approach, since it works by not leaving people behind.

"A bit over a third of poor people (in the US) do work"

This sentence is not wrong but a casual reading of it would nearly always give the wrong impression. Herein "poor people" is defined as the people below the federal poverty limit. For one person this is around 12000 dollars.

A more common sense colloquial definition of poor is those barely scraping by based on the cost of living in their area of residence. In most of those considered poor by that broader definition someone in the household is working. Remember that not all families have both partners working.

Your perhaps too brief post would inspire people to think that most people by the common sense definition of poor are in fact unemployed which is not correct.

> Jobs are created by incentivizing job creators.

Which of course leads to the question: Which job creators are most effective at providing sustainable (and, ideally, scalable) employment for the available talent pool?

They're not called 'job creators'. Jobs don't need creating. They're called self-interested capitalists. They aren't 'creating jobs' out of the goodness of their hearts. They sometimes, to a certain limited extent, require labour in order to maximise their profits.

> If it makes money, it's good

Pretty much the broken window fallacy on a societal level, perpetuated by a system that commonly grants political influence on a per dollar, rather than per capita, basis.

The way I understand it: Making money means a market, market mean consumer choice rather then government mandate, which means freedom.

The fact that there is no actual consumer freedom in many cases like this one never manages to register; the idealism of a free market overrides all the messy practicalities of reality.

Textbooks which are essentially the same are available in europe for like 1/3 of the price of the american versions. I bought $70 books for my master's courses in Sweden that would have been $300 in the US (where I did my undergrad). The "invisible hand" of capitalism at work!

> I bought $70 books for my master's courses in Sweden that would have been $300 in the US (where I did my undergrad). The "invisible hand" of capitalism at work!

I personally believe free markets ("invisible hand" of capitalism) work well in nearly all cases (yes, there exist cases of market failure, but these are IMHO rare). The problem rather is that copyright laws prevent a free market, since the copyright holder has a monopoly (i.e. opposite of free market) on the work.

By that logic we should get rid of property, since the property holder has a monopoly on the property. I would put "(i.e. opposite of a free market)" sarcastically, but I'm being serious. Go read an economics text book on the definition of a free market, and while you are at it, what properties they have (muh freedom is not even a primary one) that make us want them so much, and what the requirements are (natural monopolies are a problem, but copyright and property are not monopolies; run-away patents are though) to have them.

The person you are replying to is wrong, the free market is what allows for corrections that are starting to happen (e.g. cheaper better books, student groups publishing their own books, etc.) it's just that Sweden has a free market with better conditions than many free markets in the U.S. so they get faster reactions in their prices.

But you are wrong in a worse way, you don't even understand what you are advocating for.

Edit: To expand on the copyright is property argument. Copyright allows you to own the result of your work. If someone happens to make something very very similar without ever seeing your copyrighted work, they have the copyright to their work, and you to yours. Like property you only have that which is yours (e.g. if I built a house on my land that looks exactly like my neighbors, I still own my property and the house; but if I took their house and put it on my property I wouldn't); this is why certain techniques like Clean Room development allow companies to build very similar products that are all protected by copyrights assigned to each company. The problem you have is with patents which allow the patent holder to claim all work done that matches the patent (even if the patent holder never actually made something that matched the patent; which is especially absurd).

An important distinction between physical and intellectual property is that physical is naturally exclusive, while intellectual is not. If I use a house for something, you cannot use it for something else without disturbing me. On the other hand my using some technology does not prevent anyone else from doing the same. (Although I may prefer to have an exclusive right for obvious reasons.)

Maybe it could be a good idea to somehow recognize intellectual property as property but end that exclusive part. E.g. for books make publishing agreements public offers instead of private contracts. Just a wild idea, of course :)

> On the other hand my using some technology does not prevent anyone else from doing the same.

Yes, however, for the same reason people can own summer homes they don't use for half a year and retain that exclusivity - even though it wouldn't disturb the people - is because they worked for that. They built it, or paid someone who did. We allow people to enforce their exclusive rights on property even when it wouldn't disturb them to let others use it.

Similarly, if I spend a billion dollars building a piece of technology, I should be able to own that. And decide who gets to use it, even if giving it away to everyone wouldn't disturb me. If someone else built their own version of my software (like building their own summer home) they could give theirs away for free (open source) or sell it. But if they broke into my summer home, rented it out, and then left everything exactly the way it was, that's still illegal because they profited from my property, my work.

If you believe in private property (rather than say personal property, which would allow you to own a house exclusive to other people, but not to own multiple properties and only use one at a time) it's hypocritical to not believe in some form of copyright (unless you are advocating for feudalism, where work does not count as value, and the owners of the land or tools worked get the products of the labor by default).

> e.g. if I built a house on my land that looks exactly like my neighbors, I still own my property and the house; but if I took their house and put it on my property I wouldn't

Uhm, no, that's exactly what copyright prevents (or tries to, anyhow).

Again, you are thinking patents or trademarks.

My house may look the same, but it's been built by different people, different electricians, on different land. It probably has a different internal structure, different materials, and slightly different external marks, because it was rebuilt from the ground up.

Copyright would apply if I used the exact same blueprints, but if I hired an architect to make blueprints with the same external design as another house that would be similar to clean room design and not a violation of copyright. Like how Google can make a copy of the Java API (they only got hit for having the same implementation written by the guy who wrote the version oracle had copyrighten in the first place).

In the analogy it would be theft if I took your house, put it on my land, and claimed it was mine. And that's what it is to commit criminal copyright violation (e.g. you distribute copies for profit, claiming you own it). Yes in the digital world copies don't deprive the other of anything tangible, but like stealing a house, it does deprive them of the work that was invested to build that expression in the first place.

Again that is the difference between parents and copyrights. With a patent you can't make a copy. With copyrights you can't claim you made something. For example if I see a dress in a magazine I can make one, what I cannot claim is that it is Dolce and Gabbana...

> if I see a dress in a magazine I can make one, what I cannot claim is that it is Dolce and Gabbana

That's actually trademark, which is distinct from copyright or patent.

And notably, clothing designs are not copyrightable, because they are "useful articles", not just creative expressions. https://www.copyright.gov/register/va-useful.html

Patent: Practical invention, prevents sale even of independent re-invention. 17-20 years.

Copyright: Tangible expression of original authorship, prevents duplication and derivation except for fair use, common stereotypes, etc. Does not prohibit independent re-invention, does not prevent re-use of ideas, just the expression. Has grown to progressively longer term (currently life + 70)

Trademark: Prevents false claims of source of product, even of "confusingly" similar designs. Includes "trade dress" (packaging/presentation). Valid as long as the trademark is used.

Design patent: hybrid - prevents duplication of ornamental design of a functional product, so it is like copyright, but has a duration similar to patent - 14-15 years.

It just needs to be spun as a free market thing (the professors are entrepeneurs and competition brings down prices) :) And then cross your fingers nobody gets sued into oblivion.

In the math and engineering classes I took, the prof normally just distributed his notes via PDF.

I read this as the "prof had normally distributed notes".

At my US university the school had a (free) text book rental program. I never had to buy any books.

edit: Obviously the initial cost of the books is paid for by tuition, but it's much less than the cost of each student buying their own book.

lucky you. California universities living largely on subsides just moved to Pearson's online homework. Requiring every single student (even the ones taking a extra language class for a single semester) to buy $300+ dollars text books just for the online access code for the lame online homework website.

you simply cannot pass any class without the $300+ ticket from that publisher. Oh yes and the actual books are old and full of misogynistic cultural commentary. The riches generated in this illegal scheme is definitely not going to improve or update the books.

Few student organizations from UCLA tried to fight it, the result was that some graduate students that did not sign the petition won all the TAships next year and that was it.

The autograder homework hack by Pearson is epic. I wonder how many people from ticketmaster and comcast started working at Pearson?

The only time we had to use a professor written paperback textbook the quality was abysmal and the layout unusable.

I don't have any data to objectively support the textbook quality, but overall I generally preferred them to standard textbooks. Perhaps how common it is impacts quality.

This has happened to me in a few courses at Queen's University in Canada

A lot of people tend to harp on the "textbooks are too expensive" issue, and I think this correctly identifies one of the problems: textbook price is not an issue to many professors. Unfortunately, there's no actual solution to that presented.

> If you can’t find one, write one. It’s not that hard.

I've used three or four textbooks written by my professor, and I can't say the quality was all that great. Considering that the set of professors who currently choose to write their own textbooks probably skews toward professors who are good at writing textbooks, I'm not super high on this plan.

> Students: You should go on strike. If your textbook costs more than $50, don’t buy it. If it has more than 500 pages, don’t read it. There’s just no excuse for bad books.

Many students already do this. It's not uncommon for students to not buy a single textbook in a semester. In fact, the professors that do care about textbook price generally make textbooks optional. It turns out that's a lot easier than writing your own textbook and somehow selling it for cheap.

I understand that "just write a book" might come off as glib, especially for profs who teach 3+ classes per semester. So I don't mean that everyone has to, or should, but

1. Writing 150 pages that are targeted for your class and your students is easier than writing "the bible" for your field.

2. The quality of the first draft might not be great, but if you are getting feedback and constantly improving, it's not long before you are better off than using one of the expensive tomes.

3. If you start with other free material, you can get off to a fast start (several people have now written books that started with my material, and then evolved beyond recognition).

This is too disaligned with the criteria that are important to having a successful career as a professor. The problem isn't doing the work, it's that it takes a lot of time and is generally not a priority for the university and department vs research.

I've had some professors create fully self-contained slide decks for their course with references. Other professors teaching the same course often shared slides. Occasionally a textbook emerges from this material, but not most of the time. I think this is an approximation of your idea and probably the closest we'll get in practice.

Edit: Okay, I see in another child comment that you are a CS professor — how do you make the time to do both?

That's true. I have the good fortune to work at a college (Olin College) that recognizes that my work developing textbooks is aligned with the mission of the institution, and can have as much external impact as research. But you are right -- most places give professors zero credit for writing textbooks.

Hey man,

Just wanted to thank you for all the stuff you have given us. I have found your books to be both delightful and enlightening, and the fact you give them away is pretty astounding. I myself buy the printed copies just to support this work, however I know many many people not of means that I point your resources to with great success.

I don't know if you hear it a lot or not enough, but thank you, sincerely. Solid material you got here.

Seconded, I have Green Tea Press bookmarked and I've been meaning to read through The Little Book of Semaphores for a couple months now

We could use a few more schools that think like that!

> there's no actual solution to that presented

I think you missed this part

> All of our books are available under free licenses that allow readers to copy and distribute the text; they are also free to modify it, which allows them to adapt the book to different needs, and to help develop new material.

Along with the source to all the books, https://github.com/AllenDowney

We need to find a way to make 100 more Allen Downeys

Fortunately, from first-hand experience, Allen is churning them out himself in the form of students. He's a truly amazing professor, so while not exact, the clones are effective.

Many of his students are even involved in education in some way (incl. me, I make software for university bookstores).

> We need to find a way to make 100 more Allen Downeys

Well, as this reminds me of a discussion on here that illustrated why 'hopeless idealism' in the face of actual problems of market economics tends never to amount to much, the "answer" readily jumps to mind:

"Create a model for a market which rewards actors who behave similar to [Allen Downeys] and punishes those who do not, and migrate communities to instead use such a system."

Of course, it's a nontrivial (monumental, tbh) solution, but frankly, so is the problem it addresses - we'd have them solved if they weren't. Nonetheless, that is what is needed to 'solve' the problem (rather than merely sidestep it - though that could also be done as a non-lasting 'patch' solution).

Fortunately in this time of flux, few are confident in what "will" actually happen, and thus many are willing to try out new systems. Perhaps someone like you (or even you) will construct such a model in Ethereum or some other market platform which ends up being the dominant one in 20 years.

Why can't students just borrow textbooks from the library? I see the 'textbooks are expensive' complaint from Americans a lot and I'm confused. I never had to buy a textbook at my UK university, there were usually plenty of copies in the library. What's different about the US system?

While this doesn't apply to every case, this certainly affected me as a university student:

1. Mandatory online assignments for credit that require a code from a new textbook. (Usually because then the professor doesn't have to grade the assignments.)

2. New editions of textbooks every year, where the order of chapters or questions is shuffled, requiring you to have the latest edition of the textbook in order to do an assigned problem set ("Read Chapter 3, do question 2, 3, 5, 7 on page 148" only applies in the latest edition).

3. Libraries often only have two or three copies of the textbook, and all the students want to access them at the same time (cramming the week before exams).

4. The university library has hours and closes at night.

5. Fewer people buying textbooks leads to textbook publishers increasing per-book costs to cover the fixed costs of a print run, leading to fewer people buying textbooks.

6. Different professors teaching the same course having different preferred textbooks. Both get listed on the course description, but you don't know which one you should buy until the first day of class.

Oh right, I think a big difference is that in American courses it seems that there is one course textbook, and the assignments are taken from that. Our courses weren't designed around a textbook.

For us, assignments (both problem sheets and essay questions) were written by our tutors. They came with a reading list that would include a selection of relevant articles, books, and textbooks, but there was no single one that you had to use.

UK redbrick graduate here. My courses mostly followed textbooks as well. The library usually had enough stock for 10% of the student body to have a copy at any one time.

Librarian here...many university libraries here actually have policies against purchasing textbooks.

The reason being, there is just not enough in the materials budgets to purchase textbooks for every class (especially considering that after a year, many will be "outdated" and need to be replaced) and still have enough money to purchase all the other necessary materials (mostly database subscriptions which cost an arm and a leg, but all the other books they need to purchase as well). However, practically speaking, we can often get copies of individual textbooks on an as-needed basis through interlibrary loan. Usually someone somewhere has a copy, but it's not a solution that can be applied on a university-wide level.

Often a professor will put a copy (that they own) on reserve so it may be used within the library, but those can't be checked out and taken home for use for the entire semester.

Personally, I think it's shameful we don't have a better solution, but that's the state of things here. Many schools are turning toward open educational resources (OER) but it's slow to adopt, because many professors have a favorite textbook and OER texts aren't always of the same quality as mainstream texts. It's a really frustrating situation.

> Many students already do this. It's not uncommon for students to not buy a single textbook in a semester.

I never bought a textbook unless I absolutely completely needed it. Otherwise I did anything possible to get it if I needed it before outright buying the book. In some cases just using Google was enough for the class. Especially for most of my programming courses. Textbooks are so expensive, and half of the time you don't even go through the whole of the book, not even getting your money's worth because the semester is not long enough for said book. Math textbooks were my only exception, though the school did allow me to borrow them for a few hours.

I had a different experience in school, I kept the text book and re-read it. Eventually I had read every text book that I kept (although I did not keep them all) at least once and some of them several times. In my "first" senior year I discovered it was really useful to re-read a text book about a year after I had finished the class, apparently a lot of stuff had settled out by then and since it wasn't coming at me all new at that point I found it really helped me deepen my understanding of the topic generally.

  >> re-read a text book about a year after I had finished the class...
  >> it really helped me deepen my understanding of the topic generally.
As an endorsement of this comment, 'ChuckMcM has the deepest/broadest knowledge of an enormous range of hardware/engineering issues of anyone I am familiar with who posts frequently on HN.

> I think this correctly identifies one of the problems: textbook price is not an issue to many professors. Unfortunately, there's no actual solution to that presented.

Several of my professors solved it by supporting the last couple editions of the book vs just the latest. Works for everyone.

I'd say about 2/3s of my professors in the 90s used huge copy packets you'd buy in addition to textbooks.

My wife got the same degrees 15-20 years later (finance/economics), her courses (at what's now ranked as a "higher" rated university) were the equivalent to what I had in high school.

This wouldn't really have been an issue but when she went to graduate school, it's on par (difficulty wise) with the 90s. So the gap she had to make up was extraordinary.

It's funny, because many of my professors already do write what amounts to a small textbook, in the form of notes/slides. These are then freely available online. By saving, editing, and expanding on these notes over the course of a few years, I'm sure most could have a fairly authoritative textbook for their course.

I agree with the main point that students should read and understand textbooks. But disagree with some assumptions and points.

1. Many textbooks are written to be understood, but they vary a lot by field and class level. Generally, I think lower level textbooks best meet Downey's standards.

As you get into what is junior/senior (300-400 level) classes, there is not always a neat textbook available.

2. I disagree with 10 pages per week per course. I think the expectations of what students can read per week are too low. I attended a couple different schools, and one has a reputation of having high expectations of students, and most students tend to rise up to the challenge. I think most professors don't expect enough, and what a degree represents is watered down.

I do feel strongly that busywork and pointless readings should be avoided. Pages per week should not be some sort of metric for learning, but 10 well-written information-rich pages a week per course is not usually going to be a challenge.

Nationally, most students don't even read much of what is assigned, so telling students to not read a book if it has 500 pages won't change the status quo.

3. The idea that writing a textbook is easy is crazy. Even if you ignore the other requirements put upon professors, it is time-consuming to do it right. Even short niche books, think O'Reilly type stuff, take time to produce.

10 pages a week might be fine and well in a mathematics course and related fields, as it's a field famously characterized by it's use of a concise notation. You can state lots of things in a 10 pages. You can spend several hours working with a proof that you can fit on one page. The softcover edit of Rudin's Real and complex analysis is a small book of 400 pages, and and is supposed to a full-year course.

What about a less-mathematical course, say, history or philosophy? Well, lots of depends how dense and difficult the text is, and how many courses the student is assumed to be taking the same time.

But on the other hand, looking the bare page count, that's not much higher than the reading assignments we had when I was in high school (Finland, 00s), and there I was absolutely bored with the slow pace and had enough free time to read approx. 1 - 1.5 full-length / short-ish novels in a week. (edit. in retrospect, this feels like an overestimate, but I also read the Potters in less than 48 hours / one weekend when they were published, so maybe not.) Because the article mentions "professors", this is supposed to be about university level students: the selected few who actually are academic inclined and moreover, have chosen their field of study out of their free will. I have habit of reading books on my daily commute train trip, most recently Kahneman's Thinking Fast and Slow: I average around 4-5 pages per 20-30 minute trip. I'm mediocre student in internationally mediocre university, so I assume my mental faculties are not exceptionally high.

If there's is a problem of students simply not being able to find enough time to study (because they are working 2 jobs), this is an external problem which is not solved by changing curriculum but introducing financial support. If the problem is that students are not willing to find time to study preferring other activities, this is a problem solved by having different students.

I've had a teacher in high school and a couple in college that were in the process of writing their own textbooks. They would print out and hand out a half of a chapter or so every other week (probably ended up printing out and handing out about a third of the final book over the entire course).

They were amazingly good. I think that teachers who are passionate about what and how they teach, probably have ideas for what they want to see in a textbook. They're probably a bit too busy to have already finished the book, but they're probably working on it and checking its effectiveness on real students.

But I do generally agree with you that "most students" is not a good metric for what is reasonable today. "most students" put in an amount of effort that rounds to zero. This is not really their fault, they're just going through the system that's set up. The problem is how much pressure and reward there is for just going through the system. Classes where everyone really wants to be there just for interest in the material have a very different quality to them.

Its funny. This was my exact shower thought this morning.

I find books overly verbose, and too formal. I do not think there is a need to dumb-down technical content. I also disagree with having a page limit as it would likely lead to omission of topics that might be otherwise useful.

I think the publishing industry has to change or be weeded out by self-publishers and video makers. I have learned many topics of my courses through Slide decks and Youtube videos to avoid reading.

My plan was to start re-learning everything from Uni and write informal tutorial (snippets) of blog posts about the topics and perhaps compiling to a open source book. I'll keep you guys posted so you folks can blindly upvote my fancy submission titles =D

Addendum: I'd really like to pug Brian Douglas' Youtube playlists on Control Theory. AMAZING. Got an A thanks to him.

I just implemented the reading quiz idea this term, and I thought that it went really well. Give the students a manageable chunk of reading material (in my case, the material came from various sources, not a single textbook), and give them an offline quiz to test their reading comprehension. The quizzes were graded by a TA, but the weight was quite small; small enough not to matter if they cheated (and cheating won't help them in the final exam. Aside: is it cheating if they didn't read the material but just hunted for the answers by skimming?) but enough to provide extra incentive to read.

In class we go over the answers to the quiz. I don't post the answers (the TA will have provided them the feedback they need when grading); rather we make it an interactive session. I answer questions the students have, we go over examples, I supplement the reading with slides if need be. Effectively, a flipped classroom.

This was done more out of necessity: first time teaching the course, no proper textbook (and in a quickly changing tech landscape for the topic at hand), lack of confidence in my own understanding of the material (I also tried gathering student questions beforehand so I could investigate them offline and come to class prepared to answer them), but now I think I'm going to stick to this way of teaching this course in the next installment next year.

Here's a trick we used when I was a student: we had an FTP server shared by students of all years of our program, and we put there a copy of every required and recommended textbook, as well as slides from the lectures and every other piece of material that was relevant to our classes.

Honestly, I think this kind of setup is something universities should provide for their students. We live in 21th century, it's not that much work to provide PDFs (with restricted access, if needed, because copyright blah blah).

Here's the list of free books from the website: http://greenteapress.com/wp/

Free as in gratis, not free as in libre. It looks like most of them are under a "Non-Commercial" license.

Richard Feynman's adventures in textbooks 50 years as told in "Surely you're joking" ago are still instructive. It's pretty embarrassing that it is still so bad. More power to Downey, support him. Perfect is the enemy of good and the ally of the status quo, which is horrible.

You refer to this story, right? http://www.textbookleague.org/103feyn.htm

Textbooks are flat out unnecessary. The notes on the board should be enough to understand the material, and the teacher can either write their own problem sets or copy them from somewhere and put them online. There's just no excuse to require a textbook for a class - it means the teacher is unable to communicate the material effectively and needs the students to read it on their own, can't be bothered to write or copy homework sets, or is forcing students to buy the professor's own book out of greed, none of which should be seen as acceptable. If the department makes you have one, just don't use it (happened in a few of my classes). For classes that need some out of class readings like history or English, there's no excuse to make students buy books when the body of freely available, uncopyrighted work out there on the internet is so easy to access. Good example: a history class I had a few semesters ago where the primary documents were a simple downloadable .doc.

I've had lots of classes that worked like this, particularly my Calculus I/II classes where there was a textbook but homework from it was just suggested, not collected, and the lectures were entirely sufficient to understand the material and do well on the exams.

Beyond being a pointless scam, I'd go as far as to say textbooks make professors worse than they would be otherwise by letting professors use them as a crutch.

I disagree fairly strongly. Maybe this is reasonable for lower level courses, but in students' third or fourth year of study it seems short-sighted to expect the bulk of their learning to happen in live face-to-face lectures. Requiring them to learn independently, be it from prerecorded lectures or textbooks, allows classes to cover a lot more ground and (importantly) save lecturer face-time for the task it's actually best suited for: reacting to students' individual needs.

On covering more ground: this isn't just about expecting students to spend "their own time" learning things not covered in class, it's also about more efficiently using their time. With textbooks (and, to a lesser extent, prerecorded lectures) students can learn at their own pace. They can spend more time on the parts that they find difficult without holding up the rest of the class, and they can move more quickly through the material that they understand easily.

>in students' third or fourth year of study it seems short-sighted to expect the bulk of students' learning to happen in live face-to-face lectures.

I'm about to graduate a 4-year Computer Science program. It hasn't generally been necessary in any classes so far to get additional information outside the lectures from the textbooks for me, besides in one online class where there were no lectures, or an instance where the specifics of the virtual machine's language you had to write a compiler for were specified in the textbook - written by the teacher, but free digitally from the campus library, so no big objection in that instance aside from that not being mentioned in the syllabus, and that he could've just put all that in the text file where all the lab-section assignments were defined. But back on track, the general learning of the concept of compilation/parsing/etc occurred in class, and you applied it there. Your notes and the task description were enough.

>Expecting students to learn independently, be it from prerecorded lectures or textbooks, allows classes to cover a lot more ground and (importantly) save lecturer face-time for the task it's actually best suited for: reacting to students' individual needs.

Posting the notes online is great and often necessary, as it was in my algorithms class last semester where long algorithm pseudocode was in the notes, but the way the class worked was that everything was explained in-class, with the premade notes plus running trough things on the chalk board, and only what was in the notes was what you were responsible for. It was an effective format.

>save lecturer face-time for the task it's actually best suited for: reacting to students' individual needs.

People asked questions if they had any, and more explanation would follow, and it wasn't a problem. Taking notes on paper while listening to the lecture in real time has been shown to be very effective compared to alternatives, so using lecture time on that is probably better than leaving any concepts to be learned primarily by textbook reading.

>students can learn at their own pace.

Lecture notes being posted beforehand, as it was in that algorithms class, works for people who feel they're struggling and need to read over the notes before class, and the notes are always there later if you need to review something. Though, the student will have to resist not writing notes just because the notes are already there - aside from the audio-to-paper process's effectiveness in learning, additional explanation/metaphores/comments will be given, especially in response to questions asked. I've almost never seen it be the case that one person holds up a class with questions everyone else knows the answer to - often many others were wondering the same thing, or the answer adds more even if you thought you already knew everything about that specific point. I also think enforcing a certain pace is good - it's happened that I've read on subjects on my own and forgot parts due to going "too fast," but that's never happened in a college class, so it seems like a tradeoff of classes feeling annoyingly slow sometimes.

> I've had lots of classes that worked like this, particularly my Calculus I/II classes where there was a textbook but homework from it was just suggested, not collected, and the lectures were entirely sufficient to understand the material and do well on the exams.

I'm guessing the exams were easy but you didn't learn much calculus.

I definitely learned calculus from the Calculus classes. I took calculus 3 recently for credits with no difficulty remembering anything relevant from 1/2. Some things I could give right now off the top of my head are the definition of an integral/derivative, squeeze theorem, mean value theorem, integration by parts, epsilon delta definition of a limit, second/third derivative tests, and a lot more. Test me with something if you want.

So what's the basic idea behind Taylor series? (A "proof sketch", or even a sequence of theorems relying on each other building up to the lightbulb step.)

One of the things that was used less and less time was spent on, but they're convergent sum of a series that eventually approximates some function's value with increasing accuracy, from what I remember. You can get a sine value from one. Actually proving them, I can't remember how, but the class treated it pretty much as a short bonus topic.

If OP doesn't reply, can you? I'd like any insight I can :-)

(1) Rolle's theorem says that if f(x) if differentiable in the interval (a,b) and f(a)=f(b), there must be a point c such that f'(c) = 0.

Proving this one uses maths that, although simple, more analysis-like than calculus-like, but it should be intuitive if you try to draw a function with f(a)=f(b). It has to come up and down so it has a point where it has a horizontal tangent.

(2) The Mean Value Theorem is like Rolle's theorem, but "tilting" or "rotating" your function, so that if f(b) = r(f(b)-f(a)) then there's a point c with f'(c) = r. And this is already an order 1 Taylor series! It gives

f(b) = f(a) + (b-a) f'(c) for some c

(3) Now what remains is just induction. For an order-2 series, write the order-1 series for f'(c) as a function of f'(b). You'll start to see the pattern. If you really want to put a bow on it, you can assume the order (n-1) series exists with the known formula and prove the order (n) series works out.

So classes on something like British Literature shouldn't have textbooks?

No, I said in the post that those kinds of classes should just use downloads of the material. If some still-copyrighted books are substantially necessary then the college library should get digital copies. Are there cases where a book is substantially necessary to a course like that but publishers refuse to give college libraries digital copies?

I completely agree with the author's comment on shorter books. My biggest problem with instructional books in general is that they're filled with too much fluff. It isn't that I can't read 50 pages a week. I just don't want to, when the usable content could have been written in a page or less. While anecdotes and metaphors are great for inflating page count and price, they do little to help me understand a concept, and just become busywork, which I cannot abide.

I recommend "The no Bullshit guide to Math and Physics". It is about 300 pages, but goes along with what this article is about. This book is such a different approach (combines math with engineering and physics principals) so as to give validity to the maths you're reading. Also, it is written in a casual tone. Highly recommend it.

"Learning from Data" is a reasonable example of the type of textbook the author is asking for.

There is something to be said about the value of "reference" books, however. Maybe reference books shouldn't be used in classes, but there can be great value in a 1000 page book that has a complete discussion of everything you'd expect.

I personally work with one of the authors of this book (Malik). He has mentioned that the book optimizes simplicity and brevity, and that when the authors were writing the book they constrained themselves to an exact number of pages to force themselves to make decisions on what was important or unimportant.

Another interesting tactic they employed was an indicator on the page that the reader could skip ahead to another page (usually before a challenging proof) without losing the chapter's main concepts.

That is excellent. Do the authors have any articles about their process? Sounds like something that should be replicated in other educational material.

Textbooks have gone off the rails in terms of price (high), information density (low) precisely because the goal is profit and not education.

I don't know of any articles they've written about the process, they do maintain a forum where they post extensions to the book (on things that change often, e.g. neural networks) for free.

I agree, though I do think that professors evaluate each textbook to make sure it has the information they'd like to teach within it so the issue becomes more of a professors preference- should they provide a reference textbook or a book that perfectly matches their course.

In the case of Malik, (I believe) he writes books to fit how he thinks the material should be learned then teaches the courses to match the books. Attending the class would usually give answers to some of the example problems and go through the concepts.

The two biggest universities in my city have some sort of publishing agreement where they can print and bind the relevant textbook material to give to students for free and when possible they use open textbooks https://open.bccampus.ca/ though there is a government grant paying professors to maintain the texts.

The best thing about the open textbook site is other professors and TAs review the books like this Precalculus example https://open.bccampus.ca/find-open-textbooks/?uuid=2fdb8a19-...

While I agree with the content of this in general, I find books to be burdensome and would not like a course centered around them. I think a series of well crafted video lectures are a better medium for some people, myself included.

I also find that I can learn everything I need to in my Software Engineering program via a series of pointed google searches much quicker than reading a text. Most courses have 1 or more $100 books which are "required" but I haven't bought them in years.

What I /would/ like, is sample problems with solutions ;)

I think this depends a lot on the particular courses. For example, I nearly always want a textbook in an upper division math class, because they're so heavy on theory, and the concepts are very dense. It's useful to see it all written out on a few pages, and be able to refer to it easily. In most of my engineering classes, well done lectures work well because the concepts can be explained more clearly, and I'm a lot more likely to actually learn the material while it's being explained (because the material is not as dense).

Similarly, google searches are sometimes useful as supplementary reference for math classes, but are much more helpful in programming courses. In the more theoretical upper-division software engineering courses (theoretical computer science, AI, etc.), I find that google does not perform all that well, at least compared to other programming classes.

In contrast (because people have different learning styles) I get almost nothing out of video lectures, and vastly prefer having books. One of the things holding me back from learning Julia is the current dearth of quality books.

The language docs are pretty well written and accessible.

> I think a series of well crafted video lectures are a better medium for some people, myself included.

In a traditional course, isn't that what the (physical) lecture is for?

On your suggestion about video lectures, remember that text is accessible to some people for whom video is not, particularly blind people and deaf people.

For high-level math courses, the best I've seen is a reader written by the professor, combined with an optional 1000 page tome.

The reader goes with the lectures, and is focused on the actual material of the course. The reader, combined with your notes, basically covers the lectures. Meanwhile, if you need another take on the material, or some wider context, the 1000 page tome is always there. This works especially well if the reader points to equivalent chapters in the tome.

To add a different but orthogonal perspective to this:

My higher-level math courses were pure mathematics courses, and we pretty much always used Springer textbooks, which were only a few hundred pages long and the size of a normal paperback (i.e., not the size of, say, CLRS). When we didn't use Springer textbooks, we used other textbooks similar in size and length (e.g., [0]). I found these textbooks to be completely manageable to read as a student, and they were the best textbook-related learning experiences of my undergraduate years.

[0] https://www.amazon.com/Introduction-Galois-Correspondence-Ma...

I've had a similar experience with linguistics. The main textbook I've used was $40, is a bit under 500 pages (but is smaller than a typical textbook, and paperback), and I've used it for 3 classes. That's the kind of book I'm happy to buy. Meanwhile my experience with CS classes has been a > $150 textbook that is never referenced in class, is outdated, and is seemingly on the syllabus just to have something there. I ended up just not bothering with the books even if the syllabus claimed they were required.

What was the book?

Syntax: A Generative Introduction, 3rd edition by Andrew Carnie. I got it used, so it's probably a bit more expensive if you want a new copy, but it's a great book.

Isn't part of that due to math being quite dense? What I mean by dense is that a small amount of mathematics can express things that will take a long time to understand. Because of the density you don't need a lot of pages to contain enough material to occupy a student for a whole semester.

For subjects that are less dense, you need bigger books.

For instance all of the required and recommended books from the courses I took as part of getting a bachelor's degree in math from Caltech take less space on my bookshelf than the required books from just my first semester of law school.

Toss in the second semester of law school and include the non-required books I bought, and that would bring it to more shelf space than all the recommended and required books I bought for all the science and engineering classes I took in four years at Caltech.

As a student in my forties I have been appalled at the price and quality of many of my textbooks. The $300 dollars worth of textbooks in one class could have done with a few youtube videos. The publishers know this cash cow is coming to an end due to piracy and online textbook rentals. Now, they are charging less for textbook but gouging students on the mandatory online components.

Professors often do not have the choice to pick a book. They have to teach off of one recommended by department.

more politics into the education industry.

But 'the department' is nothing more than all the professors grouped together.

Yeah and one of them decided to venture into book writing, had connections and was friends with other professors...

That's one way to sell that pig lipsticks.

Edit: And incorrect about department = array(Professors)

That still doesn't make sense. If a department had a professor notable enough to write a textbook in a particular area, wouldn't they be teaching that course themselves? Rather than pressuring the other professor who is teaching it to use their book?

MATH courses typically have large student count, thus needing multiple professors and sections.

BTW, What I said is anecdotal.

Allen B. Downey needs to be a MacArthur Fellow.

Most commenters here should re-read the article and internalize the body of work created.

At the school I went to, many of the classes had no formal textbooks.

You bought a 50-100 page of not terribly dense text/examples; these were photocopies on plain old letter paper that were stapled together and pre-punched for three-ring-binders. Each class you'd buy one of those per semester and they were developed in-house. That was it. Naturally, it wasn't always this way, but certainly for the basic classes it was exactly this way.

Note this wasn't any sort of engineering field and what they were teaching didn't have a lot of authors writing standard issue coursework to begin with, but it was great material that was very focused to the classes they were teaching. I still hold on to them, too: very concise and a nice reference if I ever need to brush up.

Much of this articles comes across as basically complaining: "these books are long and these books are hard". Why, it's true that sometimes this is a valid criticism, but what about when the subject matter really is long and hard to understand? What then?

> Students: You should go on strike. If your textbook costs more than $50, don’t buy it. If it has more than 500 pages, don’t read it. There’s just no excuse for bad books.

This is bad advice.

You need the book for reference or at least will do better with the book most of the time. If you want to stick one to the publisher, buy used.

Textbooks need a "microservices revolution". And not with these crappy interactive DRM-ridden e-textbooks with exercise codes... the experience with most of those is markedly worse than print books. We need more open content like webpages and journal articles. O'Reilly does it best. Textbooks authors should follow / adapt their model.

Governments should definitely pay content creators to make open-source curricula for schools and universities.

I agree with Stallman that proprietary learning materials are antithetical to the spirit of learning itself. Part of the process of learning is rearranging information, changing it, collectively annotating it, and consuming it through convenient new media (a digitized, hyperlinked document on a open comment platform, for example). There are only limited possibilities with proprietary information in this regard, and even fewer possibilities with proprietary software. Schools should not use either one.

I'd like to see this philosophy applied to some of the open content out there already like the OpenStax textbooks [0]. For example, the Prealgebra text is 1152 pages in the PDF format.

I think there would be a huge amount of value in distilling these down to chapters which are 10-15 pages each instead of 100-150 pages each. Of course you would loose a lot of detail, but they could serve as a summary of 'this is the important stuff you need to know'. The expanded textbooks would serve as reference material if you want to go into more detail.

[0] https://openstax.org/

I've taken a class from Allen and used his books. In the context of his classes they are very good and the short readings can be useful, but taken as a reference like many other textbooks are they don't do as much.

I believe the earliest textbooks contained a gloss in the margin written by students which clarified the main text. It contained translations, notes, references, etc. This helped to circumvent the curse of knowledge whereby most authors have zero memory of their early misconceptions of the subject (a major reason why most textbooks are rubbish).

I think this ought to be reintroduced by major publishers -- new editions to contain copious annotations garnered from students who field-tested the previous edition, explaining how they conquered the parts that they found hard.

I studied Computer Science in Germany, and I didn't use a single book during my entire bachelor. The way it worked was professors always put the slides from their lectures online, so we could reference them as sources. Sometimes, additional references or texts were available online.

In addition to lectures, we had weekly classes where we applied the concepts from the lecture in practice. Exercises for these classes were also available online.

All of this material was free for students, and created by the professors and instructors specifically for the course.

I like the suggested price point of under 50$. Perhaps I'd go even lower and require < $40. This is enough money to keep self-published authors motivated to maintain their books and write new ones, and also affordable enough for most students.

This is the approach I've been following with my MATH&PHYS and LA books, and I will continue to use with future titles.

I guess the ideal case for students would be OER, but then when everybody owns the book nobody is particularly invested in maintaining it and improving it...

"Before long, the students learn that they shouldn’t even try. The result is a 1000-page doorstop."

My oldest two are through college and learned that they shouldn't just purchase the list of books dictated by their classes. As stated by the article, many times the textbooks were not required to pass the test. If I had to estimate, I'd say they spent half as much as their fellow students on textbooks.

Also all CS books (ok, maybe not all but the vast majority) need to ship with code. To quote Linus, "talk is cheap. Show me the code".

I agree that textbook costs are exorbitant, and I use open source, online, or very low cost books when teaching at the community college level. I've been using the interactive version of the "Think Like a Computer Scientist" book when teaching the introductory programming course. The students still don't read the material, at least not before the lecture.

Start with a quiz?

I think this is needed, most courses have one or two textbooks which compounded together with a full courseload is a lot of pages, however the full textbook is also never used, only several chapters are usually recommended reading, really beating the point of buying the whole book

Back in my student days there were really extensive textbooks that no one of us read and short brochures on the subject published by university that was the real source of knowledge and how to apply it in practice.

Many teachers use chapters from several books and their own material.

It should be possible to buy student textbooks by chapter and print your own book. Most cities with college have few high quality printing services.

I'm thinking of writing a ebook that I can use for teaching a course. I would like to see examples of well written textbooks. Are there good examples?

Anything by Kernighan/Pike/Ritchie such as The C Programming Language, The Unix Programming Environment and gopl.io book are all concise and well written books with just exactly the right amount of material it seems. Paolo Aluffi's Algebra Chapter 0 is a good math text as a math example for something that is more than just dumping propositions/theorems and proofs.

I liked this draft book as an algorithms book example http://www.parallel-algorithms-book.com/

MIT's course notes are good as brief material that compliments the lectures https://ocw.mit.edu/courses/electrical-engineering-and-compu...

Knuth also labels the difficulties of each problem set in his books like Concrete Math, something I wish more authors did. Concrete Math is also a good example of a well written and engaging math book accessible to a motivated highschool student.

I had 3 beginning stats books and they were all terrible. I used "statistics for the utterly confused" and it helped click everything together with about 1000 less pages.

Disagree - many students have different learning patterns. Textbooks are only one way to teach, and there are many different ways to learn

"If you can’t find one, write one."

What would be the best way to write a free book please? Any pointers? Experience?

I wrote one last year. It started out as notes for a series of "lunch and learns" I was going to give at work. The notes just kept growing and growing, until over the course of a weekend I ended up sitting down and pounding out 200 pages of content. That was the easy part! The hard part was then learning how to build an effective index (harder than it sounds), coming up with examples and figures, tuning and tweaking, etc.

So, to recap my experience, I would suggest that you think of a topic about which you know enough to have something to say. Not just "pure knowledge," but real opinions founded in real-world experience. Pretend you're going to give a series of presentations, lectures or articles on the matter. Start typing up the notes, looking for organizational structure as you go. Keep typing. Reorganize as you go. Keep typing. At some point you'll realize you're on your way, then the momentum will start to carry you as you figure out to what "depth" you want to go. Don't edit at this point, just keep generating content.

When you think you're close, start "spiraling" back through the book, editing, cleaning, editing, adding, editing...If the material deserves an index, go read up on that (again, it is NOT just a cross-reference of non-trivial words!) Somewhere around now hand it out to some friends to read. Incorporate their feedback. Edit again.

Look! You've written a book! :)

This is very similar to my method. The first time I teach a class, I draft some notes. After the class, I fix the problems, or at least flag them for next time. The second time through, I add, remove, edit, refine, etc., based on feedback from students and my own observation (and, often, what I have learned since last time :)

Many of my examples start with code, so the first draft of the chapter is mostly explaining the worked example.

It's not very different from the work most profs do when they are developing a class, but at the end you have a book that has co-evolved to fit your students and the learning goals.

The author of this article actually has a pretty good write up on how to do just that http://greenteapress.com/free_books.html

In education they are calling free textbooks OER: Open Educational Resources. As far as I know there isn't a single go to resource for OER books but wikipedia's article has some resources https://en.wikipedia.org/wiki/Open_educational_resources

I was just thinking about asking HN about free books that will get me started in phython. This seems to have a quite a few [0]. Has HN read any of these or recommends anything in particular?

(I went through learning python the hard way a few years back and have been slacking off)


[0] http://greenteapress.com/wp/

I know you're looking for free but Humble actually has a pretty great python ebook bundle right now, https://www.humblebundle.com/books/python-book-bundle

If you've done LPTHW, then it's more about deciding what you want to do with Python and finding something in that direction. Do you want to learn computer sciences basics? You might want a data structures book. Do you just want to continue to learn the language and standard libraries? Doug Hellman's The Python Standard Library by Example is a good reference. Statistics? Web development? It depends.

I actually wanted to get into ML (DL as you might have guessed). But I am not sure jumping into tensorflow or keras directly is good or bad.. But I just scanned through think python 2e and looks pretty good to me. This should keep me busy for now.

Why would you want to learn a programming language from a book when there are so many free guides online, including python's official ones?

Amen. This author's thesis sounds pretty darn good to me.

> Choose books your students can read and understand.

A noble and audacious goal.

> If you can’t find one,

Realistically accessed.

> write one. It’s not that hard.


Ok, I can see how this could be either plausible, or utterly absurd, depending on the domain.

> check whether they understand.

Err, does this mean "I think they didn't do too badly on the midterm"? Or daily quizes and clicker questions? Or a grad student, with a focus on the field's education research, dedicated to running concept inventories and stats?

At least in college introductory science education, "check whether they understand" is hard, an area of active research, and historically, a cesspit of professorial self-deception.

> It’s not that hard.

Let's draw a proton. With marker on whiteboard, as a circle (not hard). With an illustration app, as an arbitrarily-sized hard sphere with physically-bogus lighting (not hard). With code, as a gradient, based on the proton mass density curve (not hard, but did eat some hours).

Let's draw atomic nuclei. As balls of red and blue marbles (not hard). As gradients, post-processing from recently published ab initio density functional plots, when available (hard). Background: light nuclei are lumpy.

Ok, so let's aim lower.

Let's draw the Sun. As an arbitrarily colored circle (not hard). What about as a circle, with a color at least vaguely realistic? Demonstrably hard, as it's so rare. You likely can't ask your first-tier astronomy graduate student to do it.[1] Or almost any of the professorial authors of the many introductory astronomy textbooks.

Ok, so let's aim lower.

Last week I was reading an AP Chemistry curriculum standard. Towards the top, "atoms are conserved". Great. Later on, "atoms are neutral"[when charged, they're instead "ions"]. Okaaaaay. So are there any atoms on the right side of H + light -> H+ + e- ? The old "molecules aren't made of atoms, they're made from atoms" school. Two historical threads of definition. Left for students to reconcile, because that's obviously where the burden should lie. And this wasn't Pearson trash content, this was a curriculum spec (albeit a poor one). So what do you tell your kids to make it safe for them to take standardized exams?

"[N]ot that hard." I know wizzy education-focused MIT and Harvard professors who work really hard to raise some small bit of intro physics and biology content from wretched, to very-slightly-less-wretched.

Perhaps for some domains "not that hard" is true. And it helps if the objective is "no worse then the rest of the crap out there". And if "check whether they understand" means "ask a few clicker questions, and give a random quiz" instead of "systematically run formative misconception checks". But, wow. It so doesn't match the areas I'm most familiar with.

Perhaps the manifesto is missing some scope-of-applicability predicate?

[1] http://www.clarifyscience.info/part/MHjx6 "Scientific expertise is not broadly distributed - an underappreciated obstacle to creating better content"

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact