[phantomics]

Hi everyone, developer here. Seed is not abandoned but its original codebase has been; it is now in the process of being rebuilt. You can check on its progress in the revival branch of the repo.

Here is a quick video demo of the old Seed: https://vimeo.com/237947324?fl=pl&fe=sh

You can see my LispNYC presentation of it here: https://www.youtube.com/watch?v=nnec6_7PWkc

In my initial attempt at Seed I didn't yet understand Common Lisp well enough. That's part of the reason I wrote the April APL compiler between now and then, to gain enough understanding of the fundamentals to pursue such a comprehensive project.

The original Seed was based on React.js for the frontend and attempted to bypass textual Lisp programming as much as possible. This was an ambitions goal and React wasn't the right fit for such a project; it's a heavy system that's undergone rapid development over the years and Seed needs something that's lighter and can function more as a simple outgrowth of the underlying Lisp code. The current Seed codebase uses HTMX and Alpine.js as its main frontend tools, and the model also isn't tightly coupled to browser interfaces; it could later be possible to build terminal UIs and desktop interfaces with the same set of UI classes.

Regarding CLOG, Seed and CLOG are both interface-oriented projects but beyond that they're quite different. The idea of CLOG is a toolkit to specify interfaces with CL; it offers you many elements you can build into interface. The idea of Seed is to extend the manifestation of symbolic expressions beyond text, to have list structures that manifest interface elements that a person can interact with to create and modify programs.

For instance, you can have a series of CL function calls that perform transformations on an image, like lightening or blurring it. Using Seed's model, this list of function calls can be represented in a user interface similar to the layer lists seen in graphics software like GIMP and Photoshop. With the right combination of interface elements it could eventually be possible to duplicate the functionality of these image editing tools in a general-purpose programming platform. It would be like building GIMP into Emacs with the ability to instantly open, edit, and save a changed version of any drawing tool and have it show up in the toolbar.

This model guides the development of both the user interface and the underlying Lisp code, so it's about more than building an interface toolkit. It's about the UI as the outgrowth of the underlying software structure.

Feel free to drop more questions here.

[corinroyal]

Wait, you created April, one of the most amazing hosted languages on Lisp just to learn Lisp? Wow! And here I am toiling over a simple CRUD library management app trying to figure out CLOS and it's not even going well.

April is a revelation and I'm putting it in my framework. APL reminds me of Leibniz's alphabet of concepts for his calculus ratiocinator. So gorgeous and efficient. I'm getting the keyboard.

[phantomics]

Thanks, the utility of April was a big motivator as well, I used it in my Bloxl project (https://vimeo.com/935602359). Still, one of the reasons I focused on it as much as I did was because I knew that if I couldn't execute such a project well, I would have no chance to realize Seed. Implementing APL is a big undertaking but nonetheless one that has a clearly defined and testable goal. Seed is something with no reference to base a design off of, an objective toward which many paths appear to lead but only a few have the potential to fulfill in the best way. The challenge of such a project is on a completely different level from an APL compiler.