I wanted to create a prototype to see what’s worth to implement, but I ended up with a prototype for 4 independent tools. Please help me to understand which tool would benefit you more. It’s just a video and a 2 question survey. It would be about 7 minutes of your time
More information about the project in https://jponline.github.io/looset-landing/
This is very interesting.
It reminds me of something that I wanted to do a few years ago when I started crawling my file system, extracting certain pieces of information about projects, git repositories etc. and imported them all into Prolog, so that I could use Prolog to make inferences and ask questions about what was on my disk and where it was.
Eg. I could ask which projects are git repositories (ie. have a .git subdirectory) but don’t have a README file. Or which projects are in Clojure. Or in C. Etc.
The ideal would be to get a sufficiently rich model that you could start risking allowing the system to refactor (eg. move a project from one directory to another) etc.
And then expressing some higher level schemas in Prolog and checking if the implementations of particular data structures / classes etc. matched. Though I never got anywhere near that .
So I very much like your emphasis on “tags” which could become like facts / relations in a Prolog-like or core.logic like database.
And I would suggest that this is the way to go. Not so much emphasis on ad hoc diagramming, but on building a really solid internal model / meta-data about the projects. That could be quickly exported and dumped in different formats : Prolog, or core.logic, or as a SemanticWeb ontology etc. So that people could build their own tools to explore / visualize them, in whatever they liked.
And if that can be related to code that analyses / creates / tests schemas : say GraphQL schemas, or Clojure Spec etc. That would be even cooler.
Thank you for the feedback. I do think that the Labels are the glue that makes everything interesting, but people are visual, this is why I focused more on the diagrams and there’s no Label in this MVP.
Something similar to your suggestion have crossed my mind. I would like to be able to ask these types of questions “What Code Blocks should I refactor?”, “What Files should I split?”, then the tool would analyse CBs and files that are big AND are changed frequently, as devs don`t need to change something that is rarely touched. “What dev has worked more in a label?”, “Which Code Blocks are changed more?”, “What are the newest and oldest Code Blocks?”
I created the Landing Page and the video to collect funds for this Kickstarter campain to be able to work on it. I haven’t succeeded, so I’m still thinking about the next steps. I need to find a set of features that generate enough interest so I can invest my time on it.