Installation of a Personal Python Version

General Information

Python is available in different distributions and versions ( It works on platforms like Linux, Windows, Mac. For example, the obsolete but still working version Python 2 is currently available in the last release 2.7. The newer Python distributions are written in Python 3, the current stable version is currently Python 3.7.

In LINUX, parts of the operating system are written in Python, which is why Python 2 and Python 3 are automatically installed when the LINUX system is set up, in a file system that is only accessible with administrator rights. For example, in Ubuntu 16.04 it is Python 3.5. If you modify these Python libraries by installing a new module using a typical command with administrator rights such as sudo apt-get install ‚Pythonmodul.xy‚, it can happen that background files still required by the operating system are overwritten with other versions and unintentional problems occur afterwards. It happened to the author that after such an „update“ the screen remained black after the PC restart, because the graphics system (X server) could no longer find its libraries. Fortunately, in LINUX there are still consoles that work without the graphics system and with which the errors can be repaired or reversed.

For this reason, I recommend everyone who wants to program freely in Python to set up his own Python version in his user directory. And then later to call this Python for your own work. So the operating system Python and your own never come into conflict.


There is another possibility to install a very extensive Python3 distribution for scientific work in one step: Anaconda . The free version contains over 1400 data science packages and manages all background libraries and scripts automatically with its own administration program conda. The author had initially worked with this system. But finally some packages are missing that can be used in spectroscopy, so the author turned away from it again and installed all necessary packages manually with pip, as described in the following section. So far this works well.