ThinApp 4.5 has just been released in the last couple of days, and many of my fellow bloggers have been picking over the new goodies. Unfortunately most articles I have read just highlighted the “What’s New” section of the Release Notes.
However if you wander over to the What’s new? article on the VMware ThinApp blog site, there are some additional highlights. The one that grabbed me in particular is the section entitled “Quality improvements & Wine test”:
Additionally the ThinApp engineering team has been working diligently with the Linux Wine team to collaborate on suites of automated test. A significant number of test fixes made by ThinApp engineering were contributed back to the Wine project, especially targeted at reducing the number of test failures on Windows 7. The ThinApp engineering team has also set up ‘WineTestBot’, a service which allows Wine developers to run tests on VMware Virtual Machines which run a large selection of Windows versions. The result of the collaboration is both Wine and ThinApp improve their quality.
For the uninitiated it is worth explaining what Wine is used for. You could describe Wine as this: software which provides a layer of abstraction between applications and the underling Operating System, resolving required dependencies and enabling greater portability. Now, if you read that last sentence again, you could easily transpose the term ThinApp for Wine. Now, don’t get the impression that ThinApp and Wine are much of muchness as they’re not. Wine is set of APIs whose aim is to provide a Win32 compatible environment, allowing Windows applications to run on POSIX systems like Linux and BSD.
Now its no secret that people have been using ThinApp (and Thinstall before it) to improve compatibility with Wine on Linux. Jonathan Clark of VMware blogged about it over 2 years ago: http://communities.vmware.com/blogs/thinapp/2008/02/26/thinstall-and-wine, and even Brian Madden had a few derisory words about its real-world feasibility at the time: http://www.brianmadden.com/blogs/brianmadden/archive/2008/05/15/thinstall-wine-windows-apps-without-windows.aspx.
So why should this be any more relevant now; why should enterprises care more now than they did when it was being dismissed as nothing more than a parlour trick.
Several things have changed to make this quite an exciting area. Firstly, Wine has indeed improved. When Google bought Picasa, the photo editing application for Windows, they decided they also wanted to support Linux computers. When they looked at how complex (and expensive) a Linux port would be they decide to instead utilise Wine technologies to make it happen. If you install Picasa on Linux today, it comes with its own set of Wine binaries.
Today, devices in the workplace are more diverse than ever before. The lines have blurred, and IT departments are asked to support more and more appliances. It is not unusual to see Apple Macbooks in corporate offices, whereas only a few years before they were something you might only expect to see in a library or studio. Smartphones are smarter, with large touch screens, and plenty of horsepower to run a full stack including Wine and your indispensable Win32 app (and some do).
Within business today, a big change is occurring that will undoubtedly make this particularly pertinent. The enterprise desktop has largely stagnated over the last five years. Most companies settled on XP, built their “SOE” and figured out how to use Group Policy to manage it. Very little has changed since then, and there hasn’t been much pressure to upgrade until now. Companies large and small are looking at ways to upgrade. Now I’m relatively vocal about my support for Linux as a general purpose OS, and its suitability as a desktop replacement. However I’m also pragmatic enough to know that migrating an entire workforce to a new platform is something only the most rabidly enthusiastic administrators would attempt. What will happen over the next eighteen months is that corporations will deploy Windows 7 en-mass, and many of them will hardware refresh and probably switch to 64bit. Their biggest concern is not whether the OS will work, but if their legacy applications will still run, and how to deploy said applications. This is where application virtualisation comes in to play, and why its going to be front and centre in most company’s biggest IT project for 2010/2011.
Now VMware has a great product offering in this space, and a foot in the door with the Architects who have been pushing IT plans out recently. However Microsoft is the obvious elephant in the room. Their App-V product seems to be able to compete on features, and when you couple this with sweet licensing deals and a seamless tie-in to SCCM, then VMware has to look for another angle.
Well, I think this could be it. And frankly it is something that VMware knows Microsoft won’t try to compete on. So, would you pick an application virtualisation product to package, test and deploy everything which only works on Windows machines. Or would you pick one which will happily do this, but might also let you deploy to Linux based desktops, thin clients, kiosks, call-center stations, VM desktop pools, virtual appliances, oh and maybe even the boss’s Apple Mac. I’m not saying that everyone is going to suddenly ditch their Windows workstations. I just think that most people can see this profusion and will see Wine compatibility as a genuinely marketable advantage.
VMware’s open relationship with Wine certainly points to their realisation of this fact. This is significant. Personally, when I think of this, I dream of a stripped down VM with a minimal Linux install, GNOME desktop, vMA integrated Service Console application and a ThinApp version of a vSphere client. However that is me just thinking too small and too selfishly.
What about the bigger picture. It’s not inconceivable to imagine more general purpose Linux based vApps, with the ability to run all the common (and most uncommon) windows applications out there. Isn’t Novell up for sale at the moment?