4 ms·
There is a plugin system indeed but its not really accessible for non-technical users and cumbersome for developers. What would be a "sweet" way of integrating
by college_physics 4y ago
There is a plugin system indeed but its not really accessible for non-technical users and cumbersome for developers.
What would be a "sweet" way of integrating python more deeply?
Probably many ways and different visions are possible, but here are some ideas:
* be able to discover and install libraries from the interface (think a bit like RStudio / CRAN) into some local dedicated virtualenv
* be able to work with spreadsheets / documents interactively in notebook style, e.g. in a split window that sits side-by-side with the content.
* be able to share spreadsheet and user generated scripts in a simple package
- benj111 4y agoIs python the best choice for spreadsheets? The whitespace thing would seem to be a liability. I'm surprised there haven't been more experiments with specialised languages. I'm further surprised Unix culture hasn't produced any languages for the purpose. We've got awk and then nothing really until SQL. It's bizarre to me that we've got this entire programming paradigm and it's basically, well, Basic.
- college_physics 4y agoI mentioned python mostly due to limited imagination and because its data science ecosystem is a natural fit for working with spreadsheets. But the comment is indeed more about to think how scripting could be integrated into libreoffice more effectively both for more technical users that would write and share specific libraries and non-technical users that can use these libraries easily in "visual" point-and-click mode. People routinely load and work with spreadsheets in IDE's, or the command line etc. For example Spyder/RStudio/PyCharm etc have some rudimentary means of displaying dataframes. What would be more empowering for many more users is for the spreadsheet application to offer a IDE/REPL type environment. Once the rails for this "extension" have been laid out all sorts of alternative semantics for working with data would be possible.