I am working on a similar idea. You don’t need a b...
# introduce-yourself
g
I am working on a similar idea. You don’t need a better programming language, you just need better, relational, programmable APIs. frest.substack.com/p/what-is-frest and computer-generated UIs. And a secure composition model.
👏 2
c
What does "frest" mean? And I'm not sure SQL is as fundamental as you presume – maybe to certain areas of software, but plenty runs on key-value stores and basic file systems (still)
g
FREST was an acronym, but now it’s just a name. So it should probably be Frest? There are exceptions, but the vast majority of let’s say networked software of any complexity is based on SQL.
SQL is an abomination we’ve been stuck with for 50 years. I’ll soon post a piece on my substack about what our APIs might look like if only we had something more practical (ie Datalog). As a brief taste, imagine that your drawing program had a Datalog API. All the objects and settings and every other thing that constitute the user-visible state of the program is available to you in relations. Say you want to color all the circles within 20 pixels of a square yellow.
Copy code
circle(c)
square(s)
distance(c, s, d)
d < 20
:set color(c, yellow_color)
😭 1
Most of most software can be expressed in first order logic, which is what the relational model implements. FOL is much more malleable, less error prone, easier to understand and reason about, and has most of the other properties folks in this community care about. We just don’t use FOL in our programs because the only practical way to do it has been SQL, which is abominable.
But now we have cozodb.org
I guess my answer to “what’s the future of coding?” is: “less of it”. 🙂
d
I've only skimmed your post (I should be dayjobbing 😅) but it sounds like there's a lot of overlap between your interests and the semantic web / rfds / sparql community?
g
Insofar as both are concerned with enabling reasoning with logic. The semantic web I take to be primarily concerned with providing a means of enabling ontology-and-logic-based search and simple reasoning about the content of web resources. Frest is concerned with enabling the use of logic also, but in the form of the logical reasoning capability already defined by the Relational Model. It is also concerned with making APIs more useful by defining protocols that enable distributed resources to be accessed relationally — I want to turn the world’s APIs into a huge relational database. I also want to automatically generate user interfaces for that distributed database, much as Access and FileMaker do, but more generalised. In the end, much of what folks use Turing Complete programming languages for should become unnecessary.
I also have a novel security and composition model, which I have not yet properly described, based on the obvious merge operation between namespaces.
c
Have you chosen a proof of concept?
g
I’m currently unemployed so I need to either find a job or finish an app I’m working on before I can turn to this.
There’s a solid way to do this open source while still making a good deal of money, so my near-term goal is to make something demoable so I can seek venture funding.
Or bootstrap.
c
My unsolicited advice, don't even think about your own business (bootstrapping or fundraising, either way) until people are already knocking on your door to pay you for something of yours they're already using – "just starting a project" as your only hope is a great way to both grow poorer and incur opportunity cost (shaking my head at my own past self)
g
Thanks.