Choosing Linux is not an all-or-nothing option. Thanks to some handy tools, IT managers can let some users cling to some Microsoft applications, even if their desktops and servers run on Linux, said Robin Miller, author of a new book, Point & Click Linux: Your Guide to Trouble-Free Computing, from Prentice Hall PTR. In part one of this interview, Miller provided tips on getting started with Linux. In part two, he focuses on choosing tools and distributions and supporting Linux.
In your book, you discuss CrossOver Office, software that runs many Windows applications and plug-ins in Linux. Is CodeWeavers' CrossOver Office a good fit for enterprises?
Robin Miller: Absolutely. They now have a version that runs on servers. You can run Microsoft Office in CrossOver Office running on a Linux server, and then desktop users can access that server. So you only need one copy. Of course, you do have to have the desktop licenses for it, because it is Microsoft.
How does Win4Lin from NeTraverse Inc., also discussed in your book, make Windows and Linux play well together?
Miller: With Win4Lin, you can take an old version of Windows and run that inside Linux. You actually have to have a licensed copy of Windows to run it, of course. Here is a funny thing: With Win4Lin running Windows and Windows software under Linux, the software runs faster than if you are running Windows natively. And, if you crash -- which is what usually happens with Windows -- you can recover in about three seconds.
So, Win4Lin might be an option for small to medium-sized businesses that are still working with Windows 98.
Miller: Yes. No question about it. It is a popular option.
Actually, though, the advantage of CrossOver Office is that it does not run the Windows operating system at all, so you don't need to have Windows [operating system] licensing and the problems that come with it.
By the way, a big cost issue that people forget is that on top of the licensing cost, keeping the licenses updated and audited is ultra-expensive. You can get into a mess if you miscounted the number of your users on proprietary software. If you go to a Linux and open source on the desktop systems, that worry goes away.
Some IT shops choose Red Hat or SuSE distributions because both have vendor backing and a corporate user-focused support infrastructure behind them. What is right and wrong about that approach?
Miller: It depends on what you are dealing with. If you are running a lot of proprietary software on Linux -- for instance, Oracle application packages -- Red Hat or SuSE may be your best bets because they tend to publish the binaries for those two flavors of Linux first. But, if you have smart system administrators in-house, and you are not locked into proprietary software, Debian is probably a better choice. The pure support for Debian is unmatched.
My company's servers and critical apps all run on Debian Linux. While Debian is not the latest, it is absolutely stable. Debian is rock solid, 120% reliable for us at all times, as are the updates.
Can the peer support offered by Debian and other open source projects be enough to support IT managers in a corporate environment?
Miller: Many IT managers and system administrators tell me over and over again that they get superior support through the peer system compared to what they get from proprietary software publishers. When you have a problem with an open source package, you can usually talk to developers directly, and, if not immediately, you can be in touch with them much faster than you can within a proprietary software environment. With proprietary software, you have to go through the steps from tier one up to tier 500 technical support to get [to the developers].
With open source, you can suggest changes to software that go into the product very quickly. Here's an example that happened to me personally. I am using an open source HTML and text editing program called Bluefish. When it was first developed, it did not have a word-count function, and I felt that it should have one. I contacted the Bluefish developers via their e-mail list on a Tuesday and made my suggestion. On Thursday, we were discussing the three best ways to add a word-count function to Bluefish. And we had all three working that day. So, in 48 hours, a new feature went from a proposal by an individual user to being in the product.
Now, if I suggested a feature for Microsoft Word or anyone of the many proprietary word processor or text editor programs out there, how long do you think that it would be before my request was considered? It surely wouldn't be incorporated in the product within days.