Using a different storage format while keeping compatible with text

30 views
Skip to first unread message

eec...@gmail.com

unread,
Oct 1, 2018, 4:29:06 PM10/1/18
to Augmented Programming
Hi there,

I am slowly working my way up the list (reading messages) but I wanted to share something that might be interesting. A lot of the discussion I am seeing in the earlier threads is about text v.s. graphs (or any other structured format).

All (or at least most) of our tooling can deal very well with text. So in order to use another storage mechanism, it has to be compatible with text. This can be generalized and a solution would be virtual file systems. This is not my idea, credit has to go to Paul Phillips. You can see an introduction of Suffuse at a keynote from LambdaConf in 2015. There was a followup session called "The Virtual Filesystem: This Changes Everything".

The basic idea for such a thing is this:
  • Have the editor edit the contents of a specialized file (or database)
  • Create a projector for your favorite language (for example JavaScript)
  • Now, when you ask the filesystem for `myFile.js` (which does not physically exist), the projection is triggered to provide the javascript file
This is only the basics, you could theoretically go both ways and even go above and beyond.

We have done some experiments back in the day (see https://github.com/suffuse) and the reasons we stopped working on it were not related to its viability.


In any case I hope it inspires someone here and can possibly remove one unsolved problem from the list.


Cheers,

Erik

msfcl...@gmail.com

unread,
Aug 23, 2021, 10:33:48 AM8/23/21
to Augmented Programming
Hello

>> "All (or at least most) of our tooling can deal very well with text. So in order to use another storage mechanism, it has to be compatible with text."

Thanks for sharing :D

Two-way visual tools that can convert (Text to Visual) and (Visual to Text) is more practical

In PWCT 2.0 we did that, The ability to convert textual programs to visual programs (Program Visualization) so we can do (Visual Programming) and once we click (Save) the (Code Generation) works and the textual code is generated. So at any point of time for each source file we have two files (The textual source code file and the visual source file) : https://github.com/ring-lang/ring/blob/master/marketing/pwct/README.md

Research Papers:


Greetings,
Mahmoud
Reply all
Reply to author
Forward
0 new messages