On Thu, April 10, 2008 10:47 am, Paul Moore wrote:
On 10/04/2008, Stanley A. Klein email@example.com wrote:
In summary, Python is being used on systems that have very different underlying OS use cases. To some extent, the natural use case for Python is closest to that of Linux/Unix. Running Python on Windows/Mac requires adapting for those platforms some of the kinds of tools that simplify operations on Linux/Unix systems. This discussion is essentially about how far that goes, how to accomplish it, and how to remain compatible with the existing tools on Linux/Unix.
Thanks, that's a good summary. I would dispute your comment that "the natural use case for Python is closest to that of Linux/Unix", however. I think Python is perfectly adaptable to both environments, and from *my* point of view, the issue is that Python is currently well adapted to a Windows environment. It seems that the Unix/Linux users find it less well adapted, and need changes as a result - but in doing so, their changes are disrupting the Windows situation.
However, this is from the POV of a Windows developer, who has no sysadmin experience on Windows, and little experience with Unix. So it's certainly biased. But from where I sit, there's no Windows issue to solve, and while I'm happy for the Unix people to address the problems they have, I'd be unhappy if in doing so they *make* problems for Windows.
A windows sysadmin may have a different perspective, though.
The reason I say that the natural use case for Python is closest to Linux/Unix is that Python is FOSS and its natural approaches encourage dependencies that are not hidden from the user. It is natural in Unix/Linux to install dependencies that are not compiled in as part of a monolithic application and are not bundled with the OS.
Although Python is designed to be cross-platform, porting a "FOSS software with dependencies" environment to Windows is outside the natural Windows use case. The natural case in Windows is that all dependencies are either compiled in as part of a monolithic application or are bundled with the OS. One problem is that an excessive focus on tools that solve the "natural use case" issue tends to break the already existing solutions and distribution methods for Unix/Linux.