Previous  |  Next

Thu

04.20.06

Tim O'Reilly

Tim O'Reilly

State of the Computer Book Market, Part 2

In this installment, I'll look at specific technology trends as demonstrated by book sales. First off, here's another view of our computer book treemap, this time tweaked a bit to show the hierarchical organization of the category database. As before, click on the image to pop up a larger version.

As you can see, the treemap no longer looks like Iowa from the air, with each category like a farm field of a different size, and its relative health shown by its color. Instead, you can see that the treemap is actually a structure of boxes within boxes, showing the hierarchical organization of the data mart. We've organized the data into six "super-categories" -- Systems and Programming, Web Design and Development, Business Applications, Digital Media Applications, Consumer Operating Systems and Devices, and Other. Within each super-category are sub-categories. For example, Systems and Programming includes programming languages, databases, software engineering, general programming, security, and so on. In total, there are six levels of hierarchy.


Remember that apart from the hierarchical structure, the layout of each block in the treemap is not meaningful -- only the area and the color. The subcategory blocks are laid out by an algorithm that fits them arbitrarily into the required space of the enclosing category block.

Recapping the big picture from yesterday, you can see that the market's growth has been driven by web design and development -- a sign that "Web 2.0" is in fact driving a lot of new learning activity. The second-strongest super-category is digital media applications. No surprise there, with the strength of the iPod and digital photography. Traditional consumer operating systems books, long the mass market mainstay, are off 5%. Presumably this will reverse next year when Vista finally ships. With the exception of spreadsheets and a couple of up and coming Web 2.0 areas (search, up 103% on the strength of John Battelle's bestselling book, The Search, and geocoding and mapping applications, up 284%), business applications are also trailing the market as a whole. [Aside: next generation mapping and location technologies are the subject of our upcoming Where 2.0 Conference.]

I'm assuming that most of my readers are more interested in technical topics, so I'm going to limit my deep dive to the Systems and Programming, Web Design and Development, and Digital Media Applications areas.

Systems and Programming

(Note that because life never neatly fits into hierarchies, some things that you might expect to be part of this category are not. For example, why is PHP not included in Programming Languages? Because it made more sense to include it with Web Design and Development along with related technologies such as Cold Fusion, ASP.Net, and JSP, as it is more of a specialized language, whereas the others are not limited to the web application domain. In order to redress this kind of difficulty, we have other rollups, including one that is specific to programming languages. I'll show that in a minute.)

You can study the visualization at leisure (and feel free to ask specific questions in the comments, if the labels aren't clear in the image), but here are a few of my major takeaways:

  • Microsoft languages (VB and C#) are both showing strong growth, while Java is weakening, continuing a years-long trend. Among the open source languages, Ruby is kicking butt, with Python book sales also growing nicely. Because of the way the data mart is structured, I discuss a better visualization of computer languages in more detail later in this piece.

  • Database literacy in general continues to rise in importance, consistent with my Web 2.0 assertion that "data is the new Intel Inside", and Hal Varian's comment "SQL is the new HTML." Again, more on databases below.

  • The strengthening of Software Engineering suggests that we're moving out of the run and gun stage of new technology development into a period of consolidation and stability, where building right and to scale are once again becoming important to developers. The category Software Project Management is up 114%, driven in part by the success of Scott Berkun's bestseller, The Art of Project Management. "Agile" (up 43%) has replaced "Extreme" (down 41%) as the lightweight development meme du jour.

  • Linux books have lost a lot of their steam, with the category as a whole off 10%, with books on Red Hat hit particularly hard, off 52%. Other distributions, notably Ubuntu and Knoppix, have seen an increase of 37% over the preceding year. As I've previously observed, the decline in Red Hat book sales began when they split off Fedora from their main line development, and they have never recovered.

  • Books on Networking are up significantly, as you might expect, but books on Security are not. That's perhaps a bad sign, given the increased need for secure systems. While still a small category, VoIP is hot, up 96%, led by books from O'Reilly like Asterisk: The Future of Telephony, Switching to VoIP, and VoIP Hacks. (Our commitment to this emerging area also shows up in our Emerging Telephony Conference.)

Web Design and Development

  • As you might expect, Javascript book sales are up 121%, driven by the new interest in AJAX. (We don't yet track Ajax as a separate category, choosing instead to include it with Javascript.) Manning's Ajax in Action is the top title, followed closely by our Head Rush Ajax and Ajax Hacks. Flash, which introduced the "Rich Internet Application" category, has benefited to some extent, but is up only 12%.

  • As previously noted, ASP is up 60%. With the latest version, Microsoft has clearly found their stride in the web application development space. PHP is up only 4%, Cold Fusion up 9%, and JSP off 16%. Ruby on Rails shows in the treemap as flat, up 0%. However, this is a visualization artifact. There were no books about Ruby on Rails in the first quarter of last year, so there was a choice of representing the growth as either infinite, or zero. Given that one book about Ruby on Rails is delivering nearly the same unit sales as the 20 books that make up the JSP category, nearly one-fifth of the sales of the 61 books in the ASP category, and one-sixth of the 52 books in the PHP category, one would have to conclude that RoR is hot. And of course, there are many more books on the way.

  • In the web design tools category, Dreamweaver is consolidating its lead, and is the only application in the category whose books show growth, up 11%.

  • The 30% growth in the Web Services category is driven almost entirely by professional books on topics like Service Oriented Architectures. One could argue that books like Google Hacks, Yahoo! Hacks, Amazon Hacks, Flickr Hacks, and Google Maps Hacks, which (in part) cover these "real world web services", should also be in this category, but we have chosen to include them elsewhere.

Digital Media Applications

The digital media category is up 14% as a whole, led by the phenomenal growth of books on the iPod (up 228%), Digital Photography (general) up 59%, and some specialized areas, like books on the Camera Raw format, up 115%. We also see a slight shift towards books on the integrated CS suite, with sales in that category up 128% vs. 19% for books on Photoshop alone.

Other Dimensions

As noted earlier, one artifact of categorization is that, unlike in a folksonomy, books can belong to one and only one category, so that it's possible to sum the categories without counting books more than once. However, we can get around this by creating different dimensions in our data mart. In addition to the category-by-category rollup, we characterize every book by its operating system, and any languages or databases used in the book. So, for example, a book entitled Game Programming in Java would be game programming by category, but Java by language. Or, more to the point, PHP and MySQL books are counted as PHP books for purposes of the category rollup (because we determined that PHP was the main sales driver) but show up as MySQL books for purposes of the database rollup. Similarly, Ruby on Rails books are in a separate category from Ruby language books, but in the language rollup, they are aggregated together. In short, this is the visualization to look at if you're interested in languages.

With that preamble, here is the treemap for computer languages:

  • ".Net languages" refers to books that cover both C# and VB in the same book. If you give C# credit for all the books in this group, C# now just edges out Java as the most popular computer programming language. And if current growth trends continue, with Java off 6% and C# up 68%, C# will significantly extend its lead next quarter.

  • Ruby continues its meteoric ascent. In book sales, it is now slightly larger than Python, 80% the size of Perl, and 1/3 the size of PHP. As more publishers jump on the Ruby and Rails bandwagons, we expect these numbers to grow even more significantly next quarter. What's more, when you consider that the Pragmatic Programmers, publishers of the two most popular Ruby books, Programming Ruby and Agile Web Development with Rails, have an aggressive direct sales program, including PDF-only downloads, and report that they sold as many copies direct as they sold through retail channels, you could argue that the Ruby book market is now larger than the Perl market, and 2/3 the size of PHP. Of course, those other languages could counter-argue that they have other strong sources of online documentation!

  • The Python book market is now at 80% of the size of Perl, up from about 2/3 the size of Perl at this time last year. Perl remains in the doldrums as its adherents still wait for Perl 6.

  • Lisp, Scheme, and Lua, while small, are clearly making something of a comeback. Sales of Lisp books are up 90%, Scheme up 462%, and Lua up 100%. The folks who are mad at O'Reilly for not publishing a book on Lisp should be happy about that! (Honestly, we've got nothing against Lisp, and would love to see it take off.)

  • Actionscript, which is really just Javascript for Flash, should not be overlooked. While Ajax clearly has the mindshare, there are still signs of strong uptake for Flash. And we know from other sources that Flash on mobile may be ready to explode.

Here's the database rollup:

  • The first thing to note is just how underrated "personal" databases are. The market for Access books dwarfs the demand for Oracle books. And even Filemaker remains a strong database category -- and with its new release, it's seen a 76% increase over this time last year.

  • The SQL server book market is now more than twice the size of the Oracle book market, 50% larger than the MySQL book market, and growing faster than either of them. In fact, both the Oracle and MySQL book markets shrank versus the same period a year ago, with Oracle feeling more of the pain, off 9% to MySQL's negative 2%. DB2 is even worse off, with book sales down 14%.

  • A surprise to many may be the strong growth of PostgreSQL, up 84% over a year ago. We've also been hearing some signs of growth in the Postgres market from our "alpha geek" radar, with reasons given including better support for geo data, and better handling of very large data sets. New companies like Greenplum and EnterpriseDB have also brought a little focus to this market. We're updating our PostgreSQL book, and watching this market closely.

Disclaimer: By the terms of our contract with Nielsen, we are only allowed to share this data in the course of promoting our books. Hence the many references to O'Reilly books throughout the text. And hey, that's not a bad deal, as I don't mind promoting our books!

Tomorrow: how publishers fared.


tags:   | comments: 10   | Sphere It
submit:

 

0 TrackBacks

TrackBack URL for this entry: http://orm3.managed.sonic.net/mt/mt-tb.cgi/1188

Comments: 10

Adam S.   [04.20.06 07:08 AM]

Very interesting data (and cool visualizations). I'd love to know to what extent learning activity reflects secular trends in the usage of various technologies, and to what extent it simply reflects the stability of the underlying technology. For instance, I write a lot more JavaScript than Ruby, but I can also get by just fine with my 2002 copy of Dynamic HTML. I don't suppose you have any way of regressing the data against some sort of aggregate recency measure for the titles in each category? Doing this wouldn't definitively indicate underlying usage, but it might be an interesting way of segmenting technologies into "elder statesman" and "up and comers."

Other measures of usage are harder to come by, but would also be interesting. For instance, it's no suprise that Access books outsell Oracle books, given the relative proportion of Oracle DBAs to, um, Access DBAs. But it would be interesting to know which type of database handles more queries on any given day.

scottgu   [04.20.06 08:11 AM]

Hi Tim,

This is great data -- thanks for sharing!

Just out of curiousity -- what are the dates for the quarter of data?

Thanks,

Scott

Alex Diablon   [04.20.06 11:10 AM]

Wow! This is a gold mine of information. Thanks for sharing, Tim.

Wai Yip Tung   [04.20.06 12:22 PM]

Thanks for sharing this great data.

I would really like to see comparison of international sales.

Noel   [04.20.06 12:37 PM]

Tim,

I know you focused on Systems and Programming, but can you give some info on geocoding and mapping applications?

Also, do you have information on Data Warehousing / Data Mining topics?

Baishampayan Ghose   [04.20.06 02:42 PM]

``Programming Language'' is spelled ``Programming Langauge'' in the application. I wish I could send in a patch for that ;)

Kevin Farnham   [04.20.06 04:04 PM]

I have a question. I assume that Cascading Style Sheets is located somewhere in the "Web Design and Development" chart, but which specific box includes it?

I'm asking because I'm thinking about writing a book about customization of social interaction web pages, including MySpace.com, Yahoo! 360, etc., and customizable blog applications like WordPress, and others. MySpace doesn't allow Javascript, but I see CSS used very effectively there, and I also see CSS used to customize WordPress themes.

So, I'm wondering if whatever category Cascading Style Sheets fits into is on the move in a positive way right now.

Thanks for all the great data.

Tim O'Reilly   [04.20.06 04:20 PM]

Scott -- the data weeks end on Sundays, so the actual dates of the quarter are from January 2 to April 2.

Kevin -- CSS is included in the "web page" category, along with HTML. It's doing pretty well -- up 13% overall. There have been some very strong new books on CSS, notably our Head First HTML and CSS. The book you're thinking of would likely act more like books on the underlying web apps -- and so far, none of those sites have demonstrated their ability to drive technical book sales. So positioning could be a bit tough. You could hit the ball out of the park, or could get lost in the noise. Would love to see a proposal, though...

Baishampayan -- I've nagged at the dev team to fix that spelling bug (and a few others) for a while, but they are knee deep in other projects. Maybe this public humiliation will get it to the top of the list :-) A publishing company that doesn't proofread the text in its apps!

Noel -- Data warehousing (up 64%) and data analysis (up 226%) are both showing lots of growth. But even when an area is growing strongly, it might still not support new books. For example, the data analysis category, while growing, contains 41 titles, with an average quarterly revenue per title of about $11,000, well below publisher targets. In short, a proposal would be welcome, but not a slam dunk. The trend is going in the right direction, but there's always the need to analyze what's already been published, how well it's doing, and what hole your offering will fill. Sometimes a single new book will revitalize a crowded category on its own (e.g. Head First Design Patterns), but many times, if there are already a lot of books in an area, you need to do something distinctive and valuable or get lost in the noise.

Noel   [04.21.06 03:46 AM]

Tim,

Thanks for the info on the DW/ data analysis. I think my original question was lost there. I should have broken them up.

Can you give some info on geocoding and mapping applications? I'd also like to know if they are all related to Mash-ups, Google API, etc, or if they are more to do with desktop apps i.e. Map Info, ESRI.

Thanks,
Noel

Nick Presta   [04.22.06 09:37 PM]

Very cool. I think I will reread the article and make some other conclusions. Of course, the data presented is somewhat misleading at a view of the treemaps alone (like the Ruby 0%, for example) so you need to take everything with a grain of salt.

Great though.


Post A Comment:

 (please be patient, comments may take awhile to post)




Remember Me?


Subscribe to this Site

Radar RSS feed

BUSINESS INTELLIGENCE

CURRENT CONFERENCES