It is surprising to see a blog post like this published in 2014. Admittedly is refers to a 9 year old book, but the point it makes has been rendered moot by developments since.
A publisher recently contacted me to explore the possibility of me writing a book for them. The terms, which I believe are fairly standard, are you make about $1 for every book sold, or about 5% of the total revenue. A technical author can expect to make about $10'000 if their book sells well. The publishing house makes about $200'000 on this, with the rest going to distribution and book stores.
Now compare to people like Brennan Dunn who have made $100'000 or more from their books, by publishing online and capturing 90%+ of revenue. You don't need a publisher for distribution. You don't need a publisher for typesetting. If you need an editor or proof reader you can hire one for maybe $10K.
In short, publishing houses serve no necessary function in modern technical publishing.
In short, publishing houses serve no necessary function in modern technical publishing.
...aside from the economies of scale that they can apply to activities like editing, typesetting, graphical presentation, technical review, release scheduling, marketing, publicity, and cross-format distribution. There's also the risk mitigation that happens when you get an advance that you can't deal with yourself.
There's a world of difference between "publishing on the Internet" and "publishing in the age of the Internet".
Not vouching for him, but Aaron Shepard, a children's author, maintains you can self-publish efficiently and profitably these days.
"In fact, by aiming at Amazon, I was able to sell close to 30% the number of copies sold by one of those nationally-distributed competitors including its bookstore sales. And because the profit from my publishing system is so high -- generally half or more of the cover price -- I was earning about twice as much in total as that author would make with a normal royalty."[0]
He has a site on the subject[1], which looks like it was designed by me (it wasn't - I don't know him).
The problem is that he is biased. He is selling books explaining how to self publish on Amazon. So it is better for him to say that you will earn more money doing so.
Fair point. One more thing while I remember. He recommends a firm called Lightning Source[0]. I got the impression (when I read his book a few months ago) that they weren't too keen on dealing with small scale stuff, but just looking there now it seems they have another brand or a partner called Ingram Spark for dealing with independents. Might be useful for someone here, maybe.
This is pretty well known now in the publishing world. There are a large number of authors in genre fiction making a living self-publishing, and many more that do some mix of traditional and self-publishing.
You don't really need this guy's book though, there's plenty of information out there on how to do it (KBoards' Writers Cafe is a good place to start).
> ...aside from the economies of scale that they can apply to activities like editing, typesetting, graphical presentation, technical review, release scheduling, marketing, publicity, and cross-format distribution.
Most of which you don't care about, unless you have to edit, typeset, review, market, format 100s of books. They are one time costs and with the internet they are becoming very cheap already.
The important number is copies_sold * my_income_per_copy. Publishers might raise one of the number 2-5x but lower the other to 1/10. Cost of editing is a very small factor.
Most of which you don't care about, unless you have to edit, typeset, review, market, format 100s of books. They are one time costs and with the internet they are becoming very cheap already.
Maybe you didn't mean it the way it sounds, but I think any author cares about what their reader gets, and if I buy something I certainly care about its quality. I know for a fact that no matter how much effort I put into writing a book, there are always things that can be improved because I missed something I didn't know, or didn't know about a particular design or grammatical principle, or simply made a mistake. More eyeballs make a better product, and eyeballs cost money.
Breaking it down: Let's say I write a 1200 page technical book. Roughly speaking, services charge on a pages-per-day rate. Sake of argument, let's guess 250 pages per day for proofreading, 100 per day for graphical layout and enhancement, 150 per day for technical review. That's roughly 20 days, for which we could elance out separately for $100 per day. That's $2000. Up front.
...which brings me to my next point.
The important number is copies_sold X my_income_per_copy. Publishers might raise one of the number 2-5x but lower the other to 1/10.
I'm guessing you're estimating with these values (which is OK, because I'm estimating too). However, a quick look at Amazon and other online bookstores shows that ebooks tend to sell for a fraction of the price of paper books. $1.21 seems to be a common price point, and even for more expensive books (for which there is already a large demand) the price point seems to be about half that of a paper book.
Continuing our assumptions, let's say a published book sells 10,000 copies. Assuming the self-published book is 50% as popular as an equivalent published book (your most generous estimate) and therefore sells 5,000 copies, it still looks like 90% of a smaller price (let's say $1.21, so a total of $5,445 from 90% of 5000 * $1.21) works out at less than 10% of the larger value (assuming the parent post's assumption of $1 per copy, so 10,000 * $1). All handwavy values based on various people's assumptions of popularity, but useful for a starting point.
Now, back to the one time costs: When compared to my $5,445, that $2,000 sounds really bad now, doesn't it?
As someone writing a book, I can tell you I care a lot about the quality of the finished product but it takes far too much time to not care about the financial return as well. (I'm literally breaking off from proof reading to write this comment.)
The up front costs are negligible for most programmers in the Western world. $2K is a few days work for a contractor. If you've gone through University, as most of us have, you'll have been drilled in the rules of writing and probably know people who excel at the written word. In my case my colleagues are assisting with the proof reading. Not everyone has the cash, connections, or training for sure, but most of us do.
Finally the sales model should be very different for a self published technical book. Best practice is to offer bundles, like the Rails Tutorial does, which fall under the "don't give a shit" threshold for individuals and companies. See: https://www.railstutorial.org/ Your average income per sale should be more like $100 than $1.20.
> However, a quick look at Amazon and other online bookstores shows that ebooks tend to sell for a fraction of the price of paper books. $1.21 seems to be a common price point, and even for more expensive books (for which there is already a large demand) the price point seems to be about half that of a paper book.
eBook pricing is very different between fiction and technical non-fiction. The former (due to, I think, a large supply) has had prices driven down to a couple of bucks for all but big-name authors. The latter, since the supply is lower and readers can see tangible financial value from reading it, is much higher and not that different from print pricing.
My programming book is $25 in eBook form, and I priced it competitively with other game programming eBooks. If it was a novel, I'd have to do something much lower.
From a quick search, Higher Order Perl actually costs more on kindle than on paperback. But you are right, I am sure we could tweak those numbers either way.
> ...aside from the economies of scale that they can apply to activities like editing, typesetting, graphical presentation, technical review, release scheduling, marketing, publicity, and cross-format distribution.
I don't think there is much economy of scale in many of those. Publishers that have "house styles" can reuse a lot of layout to speed that up, sure. But editing, type-setting, technical review, etc. is just page-by-page grunt work. I'm not sure having a full-time staff for those is much more efficient than freelancers.
As someone who once tech-reviewed a Microsoft Press book, I don't see how there's any economy of scale there. They just hired me as a freelancer, after seeing a post I made on a forum.
I think the point you make about an editor is incredibly important. I've gotten quite a few books from self-published authors - including some pretty big names in the programming world - who are adamant that they can do it all themselves, and the prose is a train wreck. It's not just about catching typos - it's the actual developmental editing that needs to happen. Ensuring things are in the right order. Ensuring there's a common voice in the book. Things like that.
As an author, I've never actually asked for a refund for these terrible books, because I know how hard it is to write one. But I don't buy self-published books anymore unless I have incredibly strong recommendations from others.
The problem is that you assume that everyone wants to make a business from book writing. Most technical authors don't want the trouble of book production and sales because they already have a good paying job. They are just interested in having their names in the cover of a book. For this purpose, publishers do a reasonably good job -- certainly better than having to do all by yourself and end up with a low-quality, low-selling "book" in the Kindle store.
"A publisher recently contacted me to explore the possibility of me writing a book for them. The terms, which I believe are fairly standard, are you make about $1 for every book sold, or about 5% of the total revenue. A technical author can expect to make about $10'000 if their book sells well. The publishing house makes about $200'000 on this, with the rest going to distribution and book stores."
I'm not directly in the publishing industry, but I've been a technical reviewer on several books and have a fair number of friends who are authors/editors - mostly in the non-fiction market.
From what I've experienced those publisher profit numbers don't make sense.
Let's assume 5% author royalties (which, in my experience, is the low end 5-15% is the normal range, upper end for track record very successful folk — more for some publishers in some contexts), so that $10k = $200,000 total sales for the book.
Typically the publisher will take around 45-55% of that $200k (including what they pass onto the author). The rest will go to distributers & retailers. Let's take the upper end of that.
So 55% of $200l = $110000 to the publisher.
- the author royalties = $100k
- 10% printing costs = $90k
- 15%ish for pre-production = $75k
So absolute best case it's more like $10k vs $75k. Then there's marketing, the fact that most authors don't earn out their advance, the shenanigans over book discount, etc. So the profit to publishers coming out of that is often significantly lower.
(The slice taken by marketing in the first example above seems high from what others have told me. Its also a somewhat simplistic breakdown since it ignores some of the long-term costs from publishers that aren't related to "books" directly. e.g. the advances to authors who never get published, etc. - so the potential publisher profits aren't quite as large as they appear here - there's other overhead outside the printing/distributing/selling books bit)
"Now compare to people like Brennan Dunn who have made $100'000 or more from their books, by publishing online and capturing 90%+ of revenue. You don't need a publisher for distribution. You don't need a publisher for typesetting. If you need an editor or proof reader you can hire one for maybe $10K."
That's the best-case success case. Which is nice. But not common. Almost everybody I know who has gone the self-publishing route has ended up with the high-hundreds and single-digit-thousands.
They'd have got more from a publisher advance. They'd have also learned a lot more about how to write a good book from the publishers — since they're in the business of helping their authors be profitable.
Is a traditional publisher the right route for everybody. No. But neither are they the rapacious idiots who are exploiting authors that some folk portray them as. Almost everybody I know who went the traditional publisher route felt they got value out of it. Many of them are more than technically literate enough to go the self publishing route if they had wanted to (and some of them have for other kinds of product/market).
> Almost everybody I know who has gone the self-publishing route has ended up with the high-hundreds and single-digit-thousands.
For what it's worth, I self-published my programming book[1] about a month ago and I'm beyond that threshold already. This is (depending on how you look at it) despite the the fact that the book is also freely available on the web, or in large part because of it. (I believe the latter.)
Either way, I'm very happy I self-published now. Not only did I get the satisfaction of feeling in control of every aspect of production, but it turns out that getting a larger slice of the royalty pie is a really nice bonus. I figured that pie would be pretty small so it didn't matter much, but I've been pleasantly surprised at the number of sales.
I don't have anything against traditional publishers and can see myself using one in the future. But one thing that feels strange to me is that most of what they provide: editing, layout, type-setting, proof-reading, is a fixed quantity of work, but they are compensated for that with a fairly high percentage of ongoing royalties.
If you think of a publisher as providing those services to the author, then it means over the long lifetime of the book, they can end up costing you quite a lot.
Self-publishing basically means you accept a bit of risk by paying for those up-front (either by hiring freelancers or doing it yourself). But, in return for that, you get a larger share of the long tail of the revenue. That's a worthwhile trade-off for me.
The first link is (a) talking about John Grisham, not technical books or even most fiction ("$2.00 - Marketing - Book tour, NYT Book Review ad, printing and shipping galleys to journalists"), and (b) double counting the wholesaler/distributor ("$2.80 - Wholesaler - The take of the middlemen who handle distrobution for publishers" as well as "Most books are sold to retailers at X% discount of the cover price by the wholesalers. ... [S]o we can assume a midlist novel might be 20-30%").
The five thoughts at the end of the second are...interesting:
"There are a lot more stages and costs than anyone would imagine.
"This convoluted process can be optimized in innumerable ways – Mr. Bezos would be hard pressed to find a more ‘improvable/kaizenable’ business model.
"Everyone who works for publishing companies claims that there is little money to be made and everyone does it for the love of it. Not sure if this is a function of profits not being shared with employees or there simply are no profits.
"There is huge consolidation of power...
"The number of people employed is staggering. A bail-out for Publishing doesn’t seem as crazy an idea.
"[There are immediately obvious optimization opportunities.]"
Thanks for adding the detailed breakdown. I was concentrating on my side of the deal so I probably misremembered their side.
> Almost everybody I know who has gone the self-publishing route has ended up with the high-hundreds and single-digit-thousands.
A technical book should be selling for about $50, more if you bundle correctly like the Rails Tutorial (https://www.railstutorial.org/). 1'000 copies sold is income of about $40K (accounting for processing fees) if you self publish and don't bundle. About $100K if you bundle and have an average sale price of ~$100. I was told a typical advance is ~$5K. To earn that you'd need sales of just over 100 copies, which seems very achievable.
Just to add another data point here, I just crossed $40k gross revenue on my self-published book in a year and change. I sell entirely from my own site, bypassing Amazon entirely.
Assuming writing a technical book will at least take 3 months of your time (probably 4-6 months more realistically for a good book), making ~$10K is outrageous. And this is in case the book sells well. Who'd accept any other job if someone comes to you and tells you 'look I'll give you a single shiny dollar for every twenty dollars you make for me'.
This is slavery. Publishing houses need to be more realistic.
long live the internet... authors everywhere - self publish!
I've written for PragProg numerous times and will continue to do so. Full disclosure: I also edit books for them.
The royalty rate is 50% of the profit. The other 50% goes to pay the editor for his or her time, and other production fees. It's incredibly fair, the help authors get is great, and I earn a lot more than $1 per book sold.
I think self publishing is great but it's not for me. I am happy to give up 50% if I can focus on writing and delivering quality content while someone else spends the time making it look good, indexing it, helping me develop the content (development editor) and helping me fix mistakes (copy editing.)
The Prags' deal and commitment to actually, you know, helping their authors is unfortunately a beautiful unique snowflake amongst publishers. I've been involved with projects with all the major tech publishers in some form or another, and they're all horrible terms (usually 12-15% of proceeds [i.e., post-costs] or less).
Even worse, they do almost no work for you these days. The editors they find are 99% completely non-technical and on contract, so they're just trying to squeeze hours out of reviewing something they know zilch about. The publisher will stick your book on a website and put it in bookstores, but outside of that, don't count on any marketing or sales help. PROTIP: They don't want big reputation authors because they're better writers. They want them so they can do minimal work and ride off the authors' social media and conference speaking coattails to sell copies with next to no investment. Then they'll sell your work to places like Safari for a pittance (on the order of the cost of about 100 print copies if I recall correctly) and give you as small of a piece of the pie as humanly possible. I understand they're a business trying to survive these days, but seriously, the whole thing needs to be re-evaluated.
As GP or someone said, self-publishing is way more profitable if you can find a niche that's not being served. I've made about 10x more money off my self-published works as I have my "published" works, and that's a very sad statistic given how big the book industry really is. Not saying all publishers will end up with figures like that, but the industry as a whole is wrapped up in an author-hostile business model and a crappy, slow process.
"Unusually good book" may be too humble to be correct. It is a truly excellent book.
Many higher order concepts eluded me for years (SICP was a good introduction, but this goes much further, as an intermediate level book), but this was very enlightening. I recommend it for anyone working in just about any language. I was able to apply the concepts in Perl and JavaScript quite readily. And, there are translation s of the examples to other languages on the web these days.
I've been out of touch for many years, now, but the Perl community was one of the most generous I encountered. And they were relatively early to this, setting many an example that has been carried forward by other communities. (There were earlier communities, of course, but the Perl community seemed to be ascendent around the time that "programming" became "mainstream".
+ CPAN
+ YAPC -- a progenitor of the "just cover expenses", do it yourself un-conference
+ Frequent, and sometimes surprising, contact with the luminaries. I remember the ongoing bait and switch BS at a few Microsoft events I attended.
I don't know... I was always on the margins, not in the midst. But Perl, as a community, was always willing to show you how, and to lend a hand and a few resources as appropriate/helpful. Very generous.
This is a very good thought! It's important for an author to break the barrier to entry for the potential readers of his/her work. It's not always easy or practical. Another way of doing this is publishing something in book form what you have already published elsewhere, e.g. your blog. After all, what a reader of free books, user of free and open source software is spending is the often neglected non-renewable source of energy -- her/his time!
Sometimes however, authors consider their book(s) as their 'life's work'. It's hard to imagine giving it away for free. Psychologically speaking, though best things in life are often free, some people think that something that is 'expensive' is 'good' -- they somewhat erroneously believe that 'cheapness and quality don't go together'.
A related point that I have always wondered about is deciding the price of your creation. For stable businesses it is perhaps a straightforward thing to name the price of a creation, but I imagine it would be hard for an author to come up with the price of her book. By making it free on the website and leaving it up to the publisher to do the hard work seems like a reasonable way to get around this problem.
> Sometimes however, authors consider their book(s) as their 'life's work'. It's hard to imagine giving it away for free.
I don't think of my book as necessarily my life's work, but it's certainly one of the most significant things I've done. People can read it online without paying cash, but I don't think of that as being "free".
I do get compensated, even then. Every time someone tells me they liked my book, or enjoy my writing style, or finally understand something they've struggled with, it feels absolutely fantastic. Given how much money we spent purchasing good feelings, in many ways I feel like I just cut out the middle man. :)
> A related point that I have always wondered about is deciding the price of your creation. For stable businesses it is perhaps a straightforward thing to name the price of a creation, but I imagine it would be hard for an author to come up with the price of her book.
I just did the obvious thing: looked at a bunch of similar products and priced it in the same ballpark.
All of this is well said. This bears mentioning, though:
"For stable businesses it is perhaps a straightforward thing to name the price of a creation, but I imagine it would be hard for an author to come up with the price of her book."
Pricing is an extraordinarily difficult problem in any business endeavor, be it publishing, software, semiconductors, or lightbulbs. Obviously it's easier in long-standing industries, or in quasi-monopolistic markets. But it's never straightforward or easy.
Pricing is even harder in content businesses. First, because at least in the big-business side, so many parties have a finger in the pie, and need to be accounted for in the margin. Second, because there is no generally accepted consumer "spectrum" of price/length, price/format, price/category, price/perceived-value, etc. The guessing process is 50% science (SQL, Excel, R, benchmarks, etc.) and 50% art ("WTF will Segment A consider a decent price for Version X? Or Segment B for Version Y?")
Of course, the other big constraint is rigidity. You can't really test, adjust, or offer promotional pricing on most publishing platforms or book markets. (And to whatever extent you can, you risk consumer backlash and confusion.) So whatever you decide right out of the gate, you're usually stuck with for awhile. A promising practice -- as the author of this blog post alludes to -- is to offer different prices (and sometimes, slightly different versions) to different audiences, who might come across the book at different touchpoints. One size does not fit all, and I suspect we'll see a lot more recognition of that fact in the book market in the coming years.
This has worked very well for Eloquent JavaScript as well -- I suspect it'd have remained obscure if it hadn't been available online (as a proper, web-friendly HTML page). As it is now, it's being read by a lot of people online, and though I suspect less people read it on paper than online, it is also being recommended and bought enough to financially compensate the time that went into it.
Cory Doctorow also put together a well-written account of his views in the recently published "Information Doesn't Want to Be Free: Laws for the Internet Age." He covers DRM, copyright, publishing, the music industry, etc. The book's target is aspiring artists, but the general audience will find it valuable as well.
I have yet to go wrong taking advice from Philip Greenspun. The whole idea that YC has been espousing recently about domain-specific knowledge often being more valuable than pure technical ability, which was brought up in the context of the female founder interviews, is something that Philip has been talking about for 10+ years.
Part of this is that it's an unusually good book. But I think the longevity is partly because it is available as a free download. Imagine that person A asks a question on an Internet forum, and person B says that HOP has a section that could help with the question. If B wants to follow up, they now must find a copy of HOP. If the book is out of print, this can be difficult. It may not be in the library; it almost certainly isn't in the bookstore. Used copies may be available, but you have to order them and have them shipped, and if you don't like it once it arrives, you are stuck with it.
Small publisher here. I believe other factors have contributed to its longevity. I just did a quick lookup of the book on Amazon (http://www.amazon.com/Higher-Order-Perl-Transforming-Program...) and it has a couple of things going for it that help contribute to long-tail sales:
* Availability of new copies of the book, which signals the content is probably still relevant/not obsolete and may also indicate it’s a classic/foundation title considering how long it’s been in print.
* A relatively high number of professional reviews and reviews from experts in the field
* A moderate number of great reader reviews, including many “Verified Purchases” from the past year
The current sales rank (579,897) indicates it's probably selling 1-3 copies per month on Amazon U.S. The fact that there are so many used copies available, not to mention the free PDF, no doubt reduces sales of the new book, but there are still enough people out there who want a new copy of the print edition for their bookshelves to keep the trickle of new sales.
Regarding the convenience issue: These days, it’s possible to buy a new book online and return it or resell it later (as 63 other owners are doing right now on the Amazon U.S. site). It’s a pain to list it and handle the packing, but so is driving to a bookstore to bring back a return.
A note about free book downloads: Free titles are indeed very convenient for those who are unable/unwilling to purchase the print edition, but they are less likely to be read. I currently have a title available for on Amazon and it’s been downloaded thousands of times, but through various mechanisms (including reviews, follow-on sales of the 2nd volume, clicks to the website from the ebook edition, etc.) I have determined it’s seldom read -- I'm guessing less than 5% of the people who have downloaded the Kindle edition have actually read any part of it. I think many free ebooks and PDFs end up on people’s devices and don’t end up being read because of a lack of time and all of the other free content that’s available out there.
That's why you don't put the free e-book on amazon but on your website.
People who see your book on amazon or other book sale websites yet go trough the effort of searching for it somewhere else are probably more likely to be stripped for funds but still want to read it.
On the other hand a lot of people on the kindle store download books just because they are free and never get the time to read it.
No, it isn't. Some books that would continue to sell have great trouble due to the publisher not supplying copies for a period. Lovasz's "Combinatorial Problems and Exercises" was one such book which North Holland essentially "held out of print" by refusing to make more copies available at a reasonable price (and I hear refusing to admit it was out of print). This went on for years until a new edition came out under AMS Chelsea.
To add some data around the self-publishing space overall, the best source by far is Authorearnings.com which is run by Hugh Howey (who is most famous for having become a success by giving away his book "Wool"). According to his data, the small press, indie authors and alternative press options are dominating the Amazon top 100 lists compared to prior years where traditional publishers owned 90% of the market (genre fiction doing the best overall).
On the other hand, to self-publish non-technical books, assuming you use quality service providers, you are probably looking at about $10K before factoring in marketing (according to people like Guy Kawasaki). Juxtapose that against the fact that 400k self-pubbed ISBN's were issued last year, and the true indie success story seems like the proverbial unicorn.
* Although clearly in deep technical knowledge, there will be less competition for the quality material (short supply and all that).
Of course to further muddy the issue, what market share will go to the subscription services as they become more popular? According to Authorearnings, early data shows they are grabbing a lot of eyeballs from those top-tier super-users.
Maybe his book is still relevant today because Perl didn't evolve much during these past years ? (no troll, I just read a few articles about Perl, I have no "insider" knowledge about the language situation except they're all waiting for Perl 6).
Imagine a book about web development from 10 years ago, I guess nothing would be usable without a lot of modifications and side researches.
HOP is a brilliantly-written exposition of a set of techniques that are actually pretty ancient, but only commonly known in parts of the functional programming and specifically Lisp community. It's real connection with Perl is slim. The techniques are easily applicable to Ruby, Python, modern Perl, and Javascript.
Web development? Well, HTTP and the underlying technologies are roughly the same as they were 10 years ago; HTML, too. Relational databases haven't changed much. The underpinnings are probably going to be as recognizable 10 years from now. But web development as a whole is incredibly fashion driven. What will be popular in the future? I'm not even going to try to guess.
* The concepts of functional programming that he discusses in the book are over 30 years old but still very relevant today.
* The syntax of languages don't change that dramatically at the core. Libraries, yes, but this book is about concepts using the core language not specific libraries.
* Nobody is "waiting for Perl 6". The Perl 5 community has generally moved on (or left). We are still writing Perl 5 everyday and don't think about Perl 6.
Not really; the techniques are still relevant in Perl 6 (which you can try out today, no need for waiting), even though the syntax changed slightly.
In fact they are pretty much relevant in all dynamic programming languages that have a convenient closure mechanism/syntax. There's also a "Higher Order Javascript" page somewhere, which translates the Perl examples to javascript. (EDIT: link: http://interglacial.com/hoj/hoj.html)
Perl 5 has evolved quite nicely lately, and I don't wait for Perl 6 (the biggest problem with Perl 6 is that "6" is bigger than "5").
While I haven't read this book (only skimmed parts), this book is not affected by the language evolution. Of course, it doesn't follow 'modern Perl' practices, but mostly it wouldn't need to.
It is more concerned about concepts that can be applied regardless of language.
Eloquent JavaScript's 1st edition is 7 years old today, and still selling well. It is showing its age, and I am bringing out a 2nd edition, but if you stick to the fundamentals and don't run with temporary hypes too much, it is possible to write a relatively longevous book even on web development.
"Web development" isn't a programming language. I'd imagine most decent books on reasonably mature languages from 10 years ago would be pretty usable today.
One of the main problems for books, and particularly technical books, is breaking out and getting word of mouth, not avoiding piracy. There are a lot of technical books out there, and a limited amount of time to find the good ones.
"But I think the longevity is partly because it is available as a free download. "
I'm not sure that's the correct conclusion to be drawn. I suspect the longevity of the book is more due to the fact that it's a perl (or rather, perl5) book. And perl5, as we all know, is now officially an eternal language, never to be supplanted with a newer version. perl6, that mythical beast, is now officially a separate language not really meant to supplant perl5.
A publisher recently contacted me to explore the possibility of me writing a book for them. The terms, which I believe are fairly standard, are you make about $1 for every book sold, or about 5% of the total revenue. A technical author can expect to make about $10'000 if their book sells well. The publishing house makes about $200'000 on this, with the rest going to distribution and book stores.
Now compare to people like Brennan Dunn who have made $100'000 or more from their books, by publishing online and capturing 90%+ of revenue. You don't need a publisher for distribution. You don't need a publisher for typesetting. If you need an editor or proof reader you can hire one for maybe $10K.
In short, publishing houses serve no necessary function in modern technical publishing.