It's certainly not as difficult as it was a few years ago. However, unless you're comfortable being thrown in the deep end, it might be a good idea to test the waters first by installing free software applications that are available for both Windows and GNU/Linux on your existing system, so that when you do migrate, you're not going somewhere totally foreign.
Here are a few you can download and install for Windows to get you started:
Once you've spent a bit of time using free software on WIndows, you'll be that much happier when you migrate to a free operating system, and find your old friends already there.
As for the best operating system for a newbie user, everybody who knows me knows what I'd recommend, so I'll shut up and let others do the work for me:
Nobody is a bigger fan of standards than me, but I think the Linux Standards Base is a bad answer to a pointless and misguided question.
Firstly, there is not, nor will there ever be, an operating system called "Linux". There is an operating system kernel called Linux. The confusion caused by calling any operating system that uses this kernel "Linux" has been incredibly damaging. From the ridiculous assertion that the "Linux" operating system must be built on plagiarism because one person couldn't possibly create an entire operating system (which continues to be the basis for a lot of legal FUD), to the continuing nonsense about "Linux is not ready for the desktop/enterprise/whatever" made possible by the judicious selection of which "Linux" you are comparing to your favored proprietary product.
Second, there are plenty of standards available for operating system developers to pick and choose from. The creation of another set doesn't help.
The POSIX standard was devised to enable "source code compatibility" between software for Unix machines. In the real world, this is as good as you're going to get. Remember, key components of the "Linux" operating system, such as KDE and GNOME are designed for "Unix" (i.e. POSIX-compliant) systems, not just "Linux" systems. The upstream developers of these systems still have to cater for the BSD Unix variants, Solaris, SCO Unix, etc. The LSB is not going to help them.
Above POSIX there are a host of other standards and de-facto or evolving standards (FHS, the GNOME and KDE Human Interface Guidelines, "Section 508" requirements, etc.). Operating system developers should be free to pick the standards that are relevant for the purposes their operating systems are designed for. It doesn't make sense to insist that the manufacturer of a network router or a control system for an industrial robot should follow guidelines designed to make it easy for me to install "Quake".
Third, the goal of universal "binaries" for software packages which work on any "distribution" is unachievable and arguably undesirable. The Linux kernel works on many kinds of devices, not just desktop and server computers. Debian alone supports eleven different processor architectures, and Linux runs in all sorts of additional non-PC "embedded" environments. It is not techically possible to create a single binary package that will run unmodified on all these machines.
And the benefits of being able to distribute a single binary package for all "Linuxes", even if it were technically possible, are dubious at best. Free software operating systems currently enjoy a degree of modularity that is impossible in a proprietary world. For example, most of the multimedia software on my computer shares the same back-end "libraries" to do the heavy lifting for them. Some of the really cool software which can rip CDs to ogg files, organise and play these files, compose mix playlists, and burn these mixes back to CD are small Python programs, just tens of kilobytes in size. In the proprietary world, comparable software such as Nero is by necessity a bewildering behemoth of a thing, because proprietary software developers don't have access to such a rich array of shared components.
The "convenience" of having a single installable package puts you squarely in the Nero world, whereas a framework where your software is a collection of small interoperable modules that perform their little set of functions well and in concert with other parts of the system is infinitely more powerful. The downside of this is "dependency hell", which (I maintain) is a non-problem in modern distributions with comprehensive package repositories and good package management systems.
I think a lot of the push behind standards for universal binary distribution is from proprietary software vendors, who don't want to either produce multiple binaries for each release of their software, or release source code for others to compile binaries tailored to their system. To me, this isn't a problem, but a tactically expedient part of the solution to the problem of non-free software. I won't be shedding any tears about the difficulty of installing Photoshop or Dreamweaver on free operating systems.
Live CD's mean that you can get a peek at a number of distros without making a commitment. They run the full operating system from the CD drive, so it's a bit slower than an OS running from your hard drive. There are stacks of great advantages to this for techies and for people that just want to see how one distro or another looks running on their own PC. I would also suggest Ubuntu and gnome as a great starting place. Live CD's mean you can check out what others are using and explore with convenience at your leisure.
Migration
It's certainly not as difficult as it was a few years ago. However, unless you're comfortable being thrown in the deep end, it might be a good idea to test the waters first by installing free software applications that are available for both Windows and GNU/Linux on your existing system, so that when you do migrate, you're not going somewhere totally foreign.
Here are a few you can download and install for Windows to get you started:
Once you've spent a bit of time using free software on WIndows, you'll be that much happier when you migrate to a free operating system, and find your old friends already there.
As for the best operating system for a newbie user, everybody who knows me knows what I'd recommend, so I'll shut up and let others do the work for me:
Linuxes plural
LSB
Nobody is a bigger fan of standards than me, but I think the Linux Standards Base is a bad answer to a pointless and misguided question.
Firstly, there is not, nor will there ever be, an operating system called "Linux". There is an operating system kernel called Linux. The confusion caused by calling any operating system that uses this kernel "Linux" has been incredibly damaging. From the ridiculous assertion that the "Linux" operating system must be built on plagiarism because one person couldn't possibly create an entire operating system (which continues to be the basis for a lot of legal FUD), to the continuing nonsense about "Linux is not ready for the desktop/enterprise/whatever" made possible by the judicious selection of which "Linux" you are comparing to your favored proprietary product.
Second, there are plenty of standards available for operating system developers to pick and choose from. The creation of another set doesn't help.
The POSIX standard was devised to enable "source code compatibility" between software for Unix machines. In the real world, this is as good as you're going to get. Remember, key components of the "Linux" operating system, such as KDE and GNOME are designed for "Unix" (i.e. POSIX-compliant) systems, not just "Linux" systems. The upstream developers of these systems still have to cater for the BSD Unix variants, Solaris, SCO Unix, etc. The LSB is not going to help them.
Above POSIX there are a host of other standards and de-facto or evolving standards (FHS, the GNOME and KDE Human Interface Guidelines, "Section 508" requirements, etc.). Operating system developers should be free to pick the standards that are relevant for the purposes their operating systems are designed for. It doesn't make sense to insist that the manufacturer of a network router or a control system for an industrial robot should follow guidelines designed to make it easy for me to install "Quake".
Third, the goal of universal "binaries" for software packages which work on any "distribution" is unachievable and arguably undesirable. The Linux kernel works on many kinds of devices, not just desktop and server computers. Debian alone supports eleven different processor architectures, and Linux runs in all sorts of additional non-PC "embedded" environments. It is not techically possible to create a single binary package that will run unmodified on all these machines.
And the benefits of being able to distribute a single binary package for all "Linuxes", even if it were technically possible, are dubious at best. Free software operating systems currently enjoy a degree of modularity that is impossible in a proprietary world. For example, most of the multimedia software on my computer shares the same back-end "libraries" to do the heavy lifting for them. Some of the really cool software which can rip CDs to ogg files, organise and play these files, compose mix playlists, and burn these mixes back to CD are small Python programs, just tens of kilobytes in size. In the proprietary world, comparable software such as Nero is by necessity a bewildering behemoth of a thing, because proprietary software developers don't have access to such a rich array of shared components.
The "convenience" of having a single installable package puts you squarely in the Nero world, whereas a framework where your software is a collection of small interoperable modules that perform their little set of functions well and in concert with other parts of the system is infinitely more powerful. The downside of this is "dependency hell", which (I maintain) is a non-problem in modern distributions with comprehensive package repositories and good package management systems.
I think a lot of the push behind standards for universal binary distribution is from proprietary software vendors, who don't want to either produce multiple binaries for each release of their software, or release source code for others to compile binaries tailored to their system. To me, this isn't a problem, but a tactically expedient part of the solution to the problem of non-free software. I won't be shedding any tears about the difficulty of installing Photoshop or Dreamweaver on free operating systems.
Live CD's (sang the chorus)