Richard Bucker

Next Development Environment

Posted at — Jun 28, 2020

Being a ChromeOS user and programmer/consultant for many years I struggle with the state of the art software development. While Apple wants us all trapped in out iPads aka sandboxed application containers… there is an endless barage of people complaining about usability and ease of use. And that’s the rub; so there will always be a cadre of people wanting a plain PC, Mac, Linux etc machine.

As I have been saying for years… When my highly customized linux or windows machine dies and needs to be replaced it can take days or weeks. When my ChromeOS machine does I head to Amazon or some big box store and resume my work. And while it’s obvious that I’m shifting some risk to my servers… my code runs anywhere. I try to keep my code running on FreeBSD, OpenBSD, PopOS, ClearLinux.

Recently, a reporting system I designed and ran in docker was relocated from one prod server to another. Where the original prod server was CoreOS; which was acquired by RedHat and reached EOL in May 2020… I moved the project to chroot on FreeBSD without so much as a burp.

How did this work? The system is designed in several layers…

The DSL I’m running right now is very custom. It takes some variation of SQL and that’s ok but it means that all of the tweaking is either in the DSL or in the SQL. There are some reporting use-cases that would benefit from having better looping or conditionals. So the next DSL will be a hybrid between Lua, tcl and Markdown.

You see I like the JupyterLab. It makes perfect sense to create formatable/readable code. However, one reason I do not like it is the language limitations. I used to like python but no longer… The use-case for tcl and Lua is much stronger. And while Jupyter is fantastic for interactive development once created I need a scheduler, queuing and delivery mechanism.