User Suggestions

In order to tailor the neuron user experience, we constantly asked for suggestions on current and proposed features. User suggestions followed by the corresponding action taken by the team have been highlighted below:

What features of current IDEs/developer tools attracted you to use them?

"Simple Interface"

"Easy Learning curve"

"A lot of functionality that simplifies computation"

"Ease of use"

"Easy I/O so I don't have to do a lot of plumbing"

"Variables explorer"

"User Interface and very high compatibility with many languages and software"

"Simplicity and customisation"

"Quick scripting and incremental development"

"Popularity and good support"

"The fast repl, the fact that they are free, the fact that you can edit your code and not have to run all of it again"

"Simplicity, a shallow learning curve and being well documented online"

"Optimised linear-algebra and GPU support"

Do you have any specific recommendations in terms of features to help improve this extension?

"Look at Jupyter NBExtensions, pretty much all of those modules have some use to a large number of end users, especially things like execution time, prettify and other quality of life plugins"

"Make the most of the things you have described as options. You'll never please everyone otherwise. Also I find myself regularly changing coding environment settings depending on what I am working on"

"It would be nice to have the option to package your Jupyter notebook and related data files and then have an easy way to open them on a different machine"

"I tend to use free hosted notebooks quite a lot, so integration with them is always nice"

"When it recommends libraries to install, it should give the library name to prevent malicious code execution"

"With a highlight and run feature, the biggest problem would be maintaining context, i.e. not a lot of code can be run in isolation, it would be nice to still have access to previous variables without needing to rerun the required cells"

"A big gripe with Jupyter is that the code completion is poor/slow. Integrating Vs code's intelligence would be nice"

Action taken by the team as a result of the suggestions above:

• Hid some of the options under drop-down menus to simplify the user interaction.
• Added full setup for the Jupyter Notebook session used to allow usage of existing and remote instances.
• Automated kernel switching on file change.
• Added the option to restart existing kernels.
• Decided to use same kernel for every file written in a specific language to maintain contex for the user.
• Added import/export of Jupyter Notebook files (.ipynb).
• Changed module install user interaction model.