Python, Julia, Jupyter .... clouds

[maajdl]

Hello,

Since a few months I enjoy learning and using Python, Julia, R, Jupyter.

I abandoned Fortran long ago.
Today, I decided to break with the more recent past.
I will stop computing in Excel and C#, (mainly for process simulation, since more than 10 years) ...

I would also like to stop sharing Excel files with my colleagues and avoid asking them to install my C# stuff with admin rights.
Here is where "cloud computing" would be very exciting.

Would some of you have experiences to share about "cloud computing" for technical and scientific applications?
(in an academic setup or an industrial setup)

I am especially interested to learn about how we can share applications in this way, with maximum simplicity.
Could it be as simple as sharing holiday pictures on OneDrive?

Thanks,

Michel

 

[jedishrfu]

Im puzzled by your rejection of tools that work quite well. Programmers and analysts select the tools that are right for the job and right for the times. Fortran was a great innovation when it came out and in some circumstances is still in use today in academic circles using legacy fortran code.

In general, when I want to share information with others, I like to prepare a report document where the data is visualized. I don't want someone to look at the data, alter it in some way and then come back with questions about the altered data. However, there are times when this makes sense and so various data formats are available to do this from CSV files for spreadsheet data to NETCDF files for scientific datasets.

Sometimes we share workspaces s that other can run my java code on their machine or we use a common repository for code sharing.

I like the Jupyter notebook concept. Its like an online interactive tutorial imitating how a prof might rpresent various concepts to students. It can be used to document what you've investigated and is easy to share. Its like having a book where you can interactively explore the displayed equations to get a better understanding of things. Some universtiy profs use it today in their teaching.

 

[maajdl]

My rejection is not so much about the tools.
It is more because of my professional environment where standards are pushing people to do crazy things.
We have hundreds of engineers all using Excel as computing tool.
I have tried for many year to overcome these limitations, by pushing some java code, some vsto C# addin into Excel, ...
I used Excel-DNA to push even further.
But that doesn't solve the problem at the root.
For example, I do a lot of mathematical optimization (without the Excel solver) and had recourse to many solution, including the (failed) Microsoft Solver Foundation which was very promising.
That was all re-inventing the wheel as there are many excellent solution, free, open-source, and very efficient if one agree to break from the business routine.

You can't imagine my pleasure when I first decided to try JuMP on Julia and more recently Pyomo/Ipopt in Python.
I am very happy with this change of mind but a bit angry that I did not try earlier.

Therefore, now I am eager to find ways to share with my colleagues without fighting with the IT dept.
Sharing means that I want to do my job without troubling our IT.
But I need to know if cloud computing could help and how.
(I am also ready to leave 10% of my pay for this life-enhancing aim, paying some CPU time for colleagues)

Thanks for your suggestions

Michel

 

[jedishrfu]

Cloud computing has a fundamental issue of how it to protect your data from breaches of security. You don't have this issue as much with in-house servers.

This article gets into a lot of cloud features, architectures and issues:
'
https://en.wikipedia.org/wiki/Cloud_computing