Haskell Web Development: A Freely Available Book
indiegogo.com
indiegogo.com
If the author can crowdsource $20k, more power to her. On the other hand, maybe the monetary threshold needs a reality check. EDIT ADD: I looked again and noticed that the $20k is a "flexible goal" so presumably, the author will write the book even without reaching that amount; therefore, the author already understands market forces at work that might prevent a $20k payout.
(Of course it's possible that the Haskell book market can charge a premium because of a uniquely enthusiastic audience . E.g., doctors pay $500 for a book that shows surgical procedures and the volume of sales is only in the hundreds.)
In the economics of books, the big money is fiction. (50 Shades, Harry Potter, etc). To a lesser extent, you have some non-fiction titles making a decent amount of money (Chicken Soup for the Soul, Eat Pray Love, etc). Way down on the totem pole are the trade publications like computer books. The economics on that level basically dictate that you slave away writing "Java for Dummies" for less than minimum wage. You then use it as a business calling card and let it open doors to speaking engagements at conferences and attract consulting gigs.
Cost of living is listed as $5k/mo on this page, which contrasts with the $6.6k/mo listed on the IndieGogo campaign.
The author is a self-described anti-capitalist but this campaign looks an awful lot like free-market economics at work. Rather than simply writing the book for free and accepting donations, the author has simply stated a price ($20k) for a product (an open source book).
I also found a blog post here https://blog.cppcabrera.com/posts/53-announcing-queer-types.... where the author goes into more personal detail about the project. Apparently the $20k would cover in excess of 3 months living, whatever that means. There was even mention of releasing additional books if funding reached $60k+. Sounds pretty far-fetched.
Might be worth noting that while it's "currently in progress", it's also 912 pages of "in progress" in the current release. ;)
To put it another way, people who donate are paying a price based on perceived value.
Perhaps someone will offer to write (and make freely available) a similar book for a smaller sum. If not, there is only one offer on the table and the market will decide whether it is reasonable or not.
Haskell is written for the Haskell compiler to understand.
Specifications and designs should be written in a spoken language of choice and maths. That's where all the high-level modelling and problem domain specification should go.
Programs, on the other hand, should be written with the greatest amount of mechanical sympathy for the target platform as is necessary to achieve your engineering goals. That means we should prefer to write directly in the language understood by the platform, usually some form of assember, but where this isn't practical we write in something a little bit more complex but staying as close to that lower bound of complexity as is possible.
Before people start piling on and saying, "It's not the 80's anymore you dummy, we have spare cycles and plenty of memory," think about why I have a laptop with so much power under the hood and yet it still takes almost a minute to load up a spreadsheet... long enough that I get to stare, blankly, at a splash logo while my multi-core super-computer chugs along on the code designed in such a frame of mind.
Engineering is applied maths. Specifications are blueprints not implementations.
This, of course, depends on what your engineering goals are. If maintainability, extensibility, portability, amenability to cheap labor, etc. are considered as engineering goals then the opposite approach makes more sense: write all software as a handful of Prolog clauses and let the search algorithm do all the hard work; gradually move to lower levels if performance is unacceptable.
The fact that you're waiting a minute for your spreadsheet to load shows that the performance is acceptable: you've accepted it. If it were unacceptable, you would be spending resources (money, or time to learn/transition-to/improve) on a faster alternative like sc.
Those don't sound like goals they sound like principles.
A goal is I want to create a web server that implements the HTTP 1.1 specification that can process a stream of N bytes per second over X simultaneous connections. There are metrics, tolerances, and limits to consider.
A well engineered program that considers the data first and foremost can be easier to maintain since there are fewer abstractions to consider and reason through. It can be easier to debug since the program execution and memory layout are explicitly maintained. Performance is much easier to reason about for the same reasons. It is actually easier to refactor this kind of code in my experience.
We have to get out this idea that our job as programmers is to write code. It's really about solving engineering problems with computers -- real processors with caches, memory channels, interrupts, and the whole nine yards.
update: To wit, for those that know me, I actually write a lot of Python code. And Common Lisp. Doesn't change a thing; you can write a Python program that is ignorant of the underlying platform the interpreter was compiled for at your own risk... "write once, run everywhere," is a nice idea but a fiction in reality... a cleverly disguised, hand-wavey fiction.
I want to work full-time on this book for 3 months. $20,000 covers
my cost of living for that long.
In other words, you have to consider this as a 3 month contract as an independent contractor with no chance of extension. Personally, that would be pretty far into my discounted rate ;-)
Still, if successful, it has the upside of giving the author chances at a good reputation and follow on contracts. I'm very interested to see how successful this is. I'm sceptical that it can be completed in 3 months, but wish the author the very best of luck!
And I know this isn't indicative of her skill or anything but surely, if you're going to add 10 or so lines of code to a pitch, you'd check it wasn't buggy. Or is that just me?
Should the person making this possible charge reasonably for their time? Yes, in my opinion.
Putting it another way, for almost any other product on HN we would point out that one should price based on perceived value, so I find it curious that we are so happy with some kind of "labour theory" here.
(Of course, there are other side-benefits to simply being an author of a book, so it might be worth discounting)
But why should it even matter? When you buy a car, do you also ask if the engineers received a fair salary?
Long story short, over 3 months after taxes and health insurance, with no other benefits like retirement nor any other business expenses like a computer and internet, that $20,000 becomes $10,884 or $3,628/month.
Even professional authors working with legitimate publishers rarely make $20,000 from a book. I do hope she can write this book cause I would really like to read it, but I also think she's a bit delusional regarding what authors in tech niches actually make for a book.
Using what actual published authors make on book deals is a good way to estimate what the value of such an initiative should be.
Maybe a lot of people will be interested in the project and fund it, then great. But publishers investigate the market thoroughly before signing a deal with an author, and they offer advances and royalties based on what they think the book will make. When an author received a $7,500 royalties advance on a book it's because the publisher thinks this is a good indicator of what the book will generate in royalties.
It's a pretty good introduction to Haskell web development for people who are already familiar with both web development and Haskell.
$20,000 to write a book is an enormous amount. When you work with actual publishers you don't get that much, the upfront royalties usually stay in the 4 digit range.
Maybe it would make more sense to use a platform like GitBook or LeanPup.
If author wants to make same money as work for corporate, no problem, but should say this directly, not say the dubious "cost of living".