Following posts by a few bloggers last week I started writing about book publishing. In that blog post I provided a few links and wrote my impressions. Here is more of my reasoning.
Developers read fewer books, but how many books are worth their price?
It is well known that developers read fewer books than in the past. The amount of free information on the Internet and newsgroups, the extensive help files coming with products, the extra online documentation services provided by software companies, all play a considerable role. But that's not all. The quality of books has been getting worse over time. You often see books published with information that is not accurate, without good flow and organization, without good editing. Many publishers simply want the first book out on a subject, regardless of the quality. They are willing to add more authors, to increase production speed, making the overall flow and organization much worse (at times really badly confusing readers, with the same topics covered from different perspectives). This is true also because books are often bigger than they should be (mostly to justify their price), and have nonsensical rules such as an evenly sized chapters (all chapters should have the same page count) and the like.
Another element I find disturbing is that amongst the last few books I bought, I saw many that seem to be "adapted from a blog". I don't mind if a book covering programming from a wide perspective uses this approach, but if it is a instructional book the result is often badly organized material. A collection of examples without a higher level perspective. Specific best practices, but without any indication when you want to apply them. Or maybe these books have just been rushed, written by authors with no experience whatsoever (as authors, I mean), as writing good quality technical articles is rare these days, often replaced by blogging (which is a very relevant but totally different experience).
Quality and reviews
I agree that book quality has suffered. Also, many books still use a "reference-based" approach, at times without even a good index, making it much easier to use the web to find similar reference material. Not only quality suffers, but finding good reviews before buying a book has become tricky. Amazon for a few years was a good place to find reviews, now the number of reviews has decreased significantly, making those left much less significant of the entire readership of a book. Don't know if LibraryThink can fill that void left by Amazon (or better, its buyers) but I doubt. You often have to trust fellow bloggers to provide book reviews, but a single review (even from a source you trust) is not always better than a dozen with multiple perspectives.
Ultimately, it is getting harder to find good books, not because there aren't good books being published, but because good books seem to fall outside of classic series from classic publishers and there is no good reader ranking system these days, which is somewhat ironic, given the pervasiveness of the Web.
Some of the blog posts I referenced, complained about the uselessness of the traditional publishing industry. I partially disagree. In my career as a writer I did find some exceptional editors, who really made my books much better. Still, I agree that the overall process, with its very clear cut steps and strict deadlines that make little sense when working on a not-yet-released product. For example, I remember having to delay writing the chapter covering the IDE (one of the initial chapters) up to the very last moment to avoid having to recapture all screenshots on a more stable release. Or the nonsense of not adding than 10% of content in the second-pass editing even if a relevant feature was added to (or dropped from) the product.
In the traditional publishing model there is also an excessive separation of editorial roles (5 or 6 editors working on a single book, each from a very specific perspective, often don't help the coherence of a book, unless the author can handle them and smooth their requests), and the time taken to complete the books since you submit the latest changes (final editing pass, indexing, blue-prints, actual printing, shipping from warehouse) is incredibly long. These are all reasons I find a self-publishing model interesting.
Self publishing and do-it-all-yourself are not the same
There are many caveats, though. From bad looking covers, to unedited books, to odd formatting choices and fonts, to the lack of a good alphabetical index, to some lack of organization and flow... self-publishing should not be mistaken as "do-it-all-yourself" books. Now I'm against the "do-it-all-yourself" model.
For the books I've self-published I found a few tech editors (generally for free), I paid an editor to help with flow and English language, I paid a cover designer, I found some help on the typographical side. I did a lot of formatting myself, had to pick the proper tools, make an index of one of my books for the first time, investing some extra time for these extra efforts and for coordinating everyone. I think it was well worth it, but my impression is that "do-it-all-yourself" is way more common among Lulu authors...
Lulu and more
Coming to Lulu, I think I like it for the freedom is gives to publishers. However, as I've already discussed in the past, there are two totally different ways to use Lulu. If you buy a distribution service (get the ISBN; have your book on Amazon, at least in theory) you are back to situation very close to the traditional publishing model or the Amazon supposedly "self-publishing" model. The largest amount of money goes to the distribution channel. Publishing times change a lot, you need to set a date in advance, you have a lot of extra burden. On the other hand, if you use Lulu as a pure print-on-demand service, you have full control, but have to find your own marketing. This is good for publishing a work for a community in which you are known, but very hard if not, despite Google AdWords and other focused advertising services you can use these days.
Still there are things I don't like about Lulu (I wrote many times about the positives, I don't want to repeat them all here). They are quite slow in printing (one week average is quite a lot in the Internet age, since you have to add shipping to it). Compared to Amazon fulfillment service, there is a huge difference. Their customer service is OK, very responsive and helpful, once you get in touch. Other facts, like not listing all the countries they ship to, is turning away readers in a rather silly way. Their bulk discount is also quite a nonsense. You need to buy 50 books to have 10% off... or something similar.
The other thing I dislike is that by shielding buyers from publishers (understandable from the point of view of their business and for privacy reasons), they made it impossible to provide special deals for updated books, or a PDF plus printed book bundle, which is makes it harder to build a loyal following. Still, there is a lack of alternatives. At least, I haven't found one so far.
Electronic vs. paper
Certainly moving to a PDF format has many benefits. One could do direct sales, offer immediate delivery, update the content more frequently. But while an electronic format is great for a reference, I prefer paper if I want to delve into a topic and read hundreds of pages of a book. I can read a printed book in many more places, with different positions of my body and back. I already stare to a computer screen too many hours to be willing to do all of my reading there as well... Seems I'm not alone: even while sales of the recently released Essential Pascal are still limited (about 50 copies), the sales of the printed version double the sales of the PDF edition.
I hope to be able to improve the delivery time (offering alternatives to Lulu) for future books, as a way to mitigate the problem... If you have suggestions, let me know. I think an electronic version is great as a reference, but at this point a continuously updated web site could be even better (provided one can find a business model behind it). I'll soon publish a totally new book (not on Delphi) "as a web site", beside Lulu, and I'll see how it works.