3 ms·

> there is very little explanation of what the heck Tree Notation The docs are weak, but resources are limited. Ideally people will start their own sites/blogs

• ofibrvev 7y ago
  > Just like there are no errors in binary, there are no errors in tree notation. Tools can operate on binary without understanding the semantics. The scope of tools is pretty limited though: compression, encryption, ecc/encoding. > Tools can operate on tree notation without understanding the semantics. This is very powerful. GIGO still applies. So you move your errors downstream, as mentioned. The space where this is a win also seems pretty limited.
• zamadatix 7y ago
  > The docs are weak The docs are more than weak, for something that's supposed to take over the world in 2 years I can't even click half the links in the FAQ? > Just like there are no errors in binary, there are no errors in tree notation. You can make any format "no errors" with little effort, the problem is not creating garbage data in the process. > Tools can operate on binary without understanding the semantics. Tools can operate on tree notation without understanding the semantics. This is very powerful. And there are infinitely many representations which can take this form. The problem is all such representations are no different than each other. I.e. tree notation can't represent anything binary can't. To actually do anything new you need to understand the syntax so if the main selling point is "are syntax accepts anything" well so does a million others before and after. The question isn't "how can generic encoding work" it's "what does Tree Notations syntax do that others don't already". > All of the other notations listed have both syntax errors and semantic errors. Tree notation eliminates the entire category of syntax errors. But they aren't eliminated at all they are just hidden to produce garbage later on instead. In S-expressions if I randomly drop a parenthesis I have mismatched syntax. In Tree Notation if I randomly drop a newline I have some completely different piece of data that validly represents garbage. The issue isn't eliminated just because it parses and this isn't the first format that can hold arbitrary data! . In the end though what is this beyond yet another form of S-expression? Is there something it does that a simple replacement regex would generate an invalid S-expression? If not why blow smoke up everyone's ass instead of just saying "Tree Notation is a S-expression format that uses whitespace instead of parenthesis"?
  • breck 7y ago
    > The docs are more than weak, for something that's supposed to take over the world in 2 years I can't even click half the links in the FAQ? On my goodness. Thank you. Fixed. I have a script that checks for broken links "npx linkinator -r https://treenotation.org" https://treenotation.org" but I must have screwed something up because those were not getting detected. (Nit: 10 years! not 2 ;) ) > You can make any format "no errors" with little effort, the problem is not creating garbage data in the process. This is a very valid point. But Tree Notation does give you a lot of meaning -- lines/words/scopes -- without knowing the semantics of the Tree Language. But you do have a good argument here that if people don't come to a consensus on even these things, then Tree Notation could just be garbage. There will need to develop a common language for the universal tree notation semantics. > And there are infinitely many representations which can take this form. Good point. > The question isn't "how can generic encoding work" it's "what does Tree Notations syntax do that others don't already". Tree Notation is the simplest method we can come up with to do all the things are existing syntaxes and languages can do. By simplest, we mean "tree notation will always have fewer or equal pieces in terms of absolute complexity than a different notation for any given task". And our bet is that simplicity of atomic ingredients triumphs in the long run because less costs less. But the proof is in the pudding. Tree Notation won't really start taking off (if ever) until there is better tooling and some "killer app" new tree languages. > But they aren't eliminated at all they are just hidden to produce garbage later on instead. Clinicians are now using Tree Notation to write their medical records on pen and paper (peer reviewed paper coming soon). There's no such thing as syntax errors. Their "code" looks just like S-expressions, with prettified formatting, except without paranetheses. They can have semantic errors using either one (for instance, spelling "dose as does"). But they cannot have syntax errors with Tree Notation. Imagine instead we were instructing them to use S-Expressions--suddenly their notebooks would be filled with syntax errors for missing paranetheses. They would have 0 benefit and all downside by adding parens. > If not why blow smoke up everyone's ass instead of just saying "Tree Notation is a S-expression format that uses whitespace instead of parenthesis"? I'm putting a small fortune and an immense amount of effort into this (and giving everything away), because I believe it will eliminate drastic amounts of unnecessary complexity in so many domains all throughout the world, and will indeed, amongst other things, help us cure cancer, which is certainly near and dear to my heart and I'm sure to so many other people on here. I want to make the biggest and boldest statements I can here, so I can get harsh and sharp critical feedback like yours, that is very helpful and can save us time by pointing out flaws (both in theory and practice like the broken links). If this were something for Lispers, I wouldn't care so much (I love lisp, but it would be an incremental improvement). This is for all the people of the world. We are going to simplify and make better code in all domains. It's a mammoth effort that's going to take billions and thousands of smart people. This is just the beginning.