Yes. I consider this to be a bioinformatic question. Around me, there are a lot of people using Windows 10 PCs that are dreaming about having a native bash kernel working with this operative system. It will allow the using of many programs without the hassle of needing to use virtual machines or install an alternative start with Linux. It will allow the sharing of the same file directories used for Windows
I've been able to use kallisto successfully with the test data provided under this system. Istvan Albert, however, has pointed out an issue using fastqc under the command lane. He was able to run it, but the system crashed after finishing the analysis. I installed java SRE on my own computer and FastQC. The fastqc --help works, but the system hangs if I provide a working command with a known fastqc file that I have used many times before
So this forum is to try to share our experience with this new kernel if you are interested in it
While having bash shell available in windows is a convenience it may be a stretch to expect to use it as a full fledged replacement for unix in this 1.0 release. Being able to ssh/sftp without need of an external program is a good start (for me).
I was excited about this Linux kernel on Windows 10, since it is a truly Ubuntu Linux with Canonical support. I must recognize that I was trying to use it as a replacement and that my expectations are very high. This is in part due that it is very much convenient for my students.
to teach my students how to use a program under the command lane. This is my purpose.
I know that it can work under native windows, but there are many programs that need to be used under the command lane. In addition, we have access to supercomputers that need to be used under the command lane. No graphical environment is provided
About fastqc, it's a bit odd it causes problems since it's Java and should run anywhere with the appropriate Java virtual machine. Maybe it's the perl wrapper script that has issues...?
While wsl might have some advantages in some scenarios, I do not fully understand the current hype around it. I guess the main advantages in the area of bioinformatics are:
bash: finally a powerful command interpreter in windows, one can share some simple workflows and scripts with biologists who are using 99% windows and use grep,sed,awk
bioinformatician will seem less windows illiterate ("what was the command for grep -ce '>' again in win-cmd?")
use of ubuntu binaries, but how many headless bioinformatics apps are there in the ubuntu repository?
But there are some serious limitations also, which need to be pointed out by the documentation:
No GUI applications, I guess that refers to any X11 applications, also those linked against X11 but running headless?
I guess one cannot compile own binaries, even if all dev-packages are installed? What about perl modules with compiled code?
Or can one cross-compile applications on an Ubuntu system to run in wsl and how? Or can only canonical do that?
I am aware that a GUI can be implemented into the Windows 10 Linux. A google search does the magic
The Windows 10 Linux has a compiler installed, and you can install and update and upgrade anything since the apt-get utility is provided. In fact, I compile the Kallisto program using the source code. This includes the possibility of adding repositories.
You don't use a windows cmd. Linux within Windows 10 uses its own command lane window
There are binaries that I did compile in a true Ubuntu machine, that I moved to the Windows Linux, and it worked for me
I did not play with perl and/or python extensively, but I can recall that they are installed already when you install the Windows 10 Linux
While having bash shell available in windows is a convenience it may be a stretch to expect to use it as a full fledged replacement for unix in this 1.0 release. Being able to ssh/sftp without need of an external program is a good start (for me).
I was excited about this Linux kernel on Windows 10, since it is a truly Ubuntu Linux with Canonical support. I must recognize that I was trying to use it as a replacement and that my expectations are very high. This is in part due that it is very much convenient for my students.
Sorry to say that I am already disappointed
Btw, FastQC should run in windows java, shouldn't it, why run it in wsl?
to teach my students how to use a program under the command lane. This is my purpose.
I know that it can work under native windows, but there are many programs that need to be used under the command lane. In addition, we have access to supercomputers that need to be used under the command lane. No graphical environment is provided
About fastqc, it's a bit odd it causes problems since it's Java and should run anywhere with the appropriate Java virtual machine. Maybe it's the perl wrapper script that has issues...?
Adding to this Nextflow does not work on new windows bash either.