It is now Spring in the Northern Hemisphere, and our Anglo-American tradition calls for spring cleaning. So, in the spirit of the season:
-
Fundry is closing down. So you will be unable to make donations or fund new features for the Data Journalism Developer Studio there. I don’t have any plans for an alternative revenue stream at this time.
-
Speaking of revenue, I’m not making enough revenue from Powell’s, Amazon, Clicky, Viralheat or HootSuite affiliate links to justify the effort involved. Over the next few weeks I’ll be removing nearly all those links from this blog. I’m staying in the affiliate programs for the moment, but I don’t have the time to be an active full-time affiliate marketer, and that’s what’s required to make any money out of these programs.
-
I am shutting down the separate blog for the Data Journalism Developer Studio on Github. Octopress is way too geeky and maintaining two blogs is diluting my focus. It was a noble experiment, but it has failed and it’s time to move on. I’ll move the important posts there over to this blog before I shut it down.
-
I will most likely be moving this blog from a self-hosted WordPress site to a free blog hosted at WordPress.com. Again, the amount of time I’m spending in plug-in and theme maintenance isn’t justified by a payoff. If I were a professional WordPress developer, selling my services, that would be different. But I’m not and don’t plan to be.
And what of the future? I’m glad you asked!
I’m currently re-evaluating the strategy for the Data Journalism Developer Studio, given some data I’m collecting on my visitors, looking in depth at the tasks data journalists actually do and the questions they ask, and my own desire to get more involved in advanced open source technologies. I don’t have all the details nailed down just yet, but here’s a rough road map.
-
The current appliance is a Linux desktop. When I look at my visitor data, 56.2% of my visitors use Windows, 23.9% use MacOS X, 10.1% Linux, and 9.7% mobile. I suspect that my Linux numbers are higher than those for most sites, since I’m a long-time Linux professional. But given these numbers, and the fact that all of the desktop tools packaged in the current appliance will run on a Windows or Macintosh desktop as well as a Linux desktop, it makes no sense for me to maintain and re-distribute a Linux desktop and desktop tools when 80 percent of my users already have a perfectly good non-Linux desktop where they can run those tools!
So I will be freezing the current appliance as soon as I have all the documentation done on how to get the desktop tools if you’re on Windows or MacOS X. If there’s enough demand for it, I’ll port the tool install scripts to Fedora and Debian/Ubuntu/Mint.
-
I’ve had a number of requests for the Twitter research tools. You’d think people had better ways of capturing Twitter search data, contact lists and tweets from a user, but I keep getting these requests. So I’m considering packaging them as Windows executables using the ActiveState Perl Development Kit. I can also make Macintosh executables with that kit, but since I don’t own a Macintosh for testing, that’s way down on my priority list.
-
I am creating a new appliance, tentatively called the Data Journalism Advanced Server. I’m prototyping it at present on the SUSE Studio, but I’m not necessarily going to use that as the final delivery mechanism. As the name implies, it will be a server. I’m still evaluating the underlying technologies, but I’m pretty sure of some things:
-
It will run a 64-bit Linux. Some of the underlying mathematical software is optimized for 64-bit processors, and if I include a virtualization hosting technology, that requires 64-bit processors and some extra flags.
-
As always, it will be 100% open source.
-
Much of the software will be built on-appliance from upstream source. My tentative strategy is to build a core openSUSE appliance on SUSE Studio with kvm as the virtualizer, and building the actual server as a Gentoo guest.
-
It will have R and RStudio Server, plus all of the LaTeX tools required to build R documentation in RStudio Server.
-
It will have Node.js and Django for sure. It may or may not have Sinatra and Ruby on Rails; I think there are better ways to deploy those technologies than what I’m building.
-
It will have all of the databases described in Seven Databases In Seven Weeks.
-