Hacker News Reader

Comment by abeppu | Hacker News Reader

Most of this thread is giving entirely good reasons why this is worse than markdown; longer to type, kinda worse to read, misunderstands how markdown aligns with _patterns people were already using_.

I agree with those, and I wouldn't reach for dumbdown either.

But maybe more interesting is the tree notation thing that this is meant to be a use case of. So far as I can tell, the tree-notation thing is a declarative DSL for grammars and textual "compilers", like maybe a very simplified ANTLR? Dumbdown is a ~100 line example, so even if it's not useful as a "real" format, it does seem illustrative of the tooling. And that tooling makes me think ... maybe we would have been better off if markdown (and extensions like GH-flavored markdown) had been implemented and shared as declarations in a language-description DSL.

View on HN

You beat me to it. Any time treenotation.org is involved, the most fascinating part is how you can build compilers for interesting use cases in a very simple declarative way.

Usually, creating a new DSL is a complex programming task; yet with Tree Notation, a domain expert may tweak an existing DSL or even build a simple one from scratch. As you say, maybe the markdown ecosystem would benefit from this approach.

In today's world of a thousand package managers, why isn't there a package manager for datatype schemas? Then, when I'm looking for an "addressbook" schema, I can just download the most popular one. It could include tooling/plugins to convert schemas to different data or programming languages, so create a Haskell datatype for "addressbook" or create the JSON representation of "addressbook".

I think the Semantic Web was supposed to be that. Yet it was waaay too abstract for the majority of people, even programmers.

Schema as a Service here we goo

Schema.org is probably the closest thing to that.

I call the concept "Type the World" (https://breckyunits.com/type-the-world.html)

Here's an early hacky prototype called "World Wide Types" (WWT): https://jtree.treenotation.org/designer/#standard%20wwt

Personally I find Markdown extremely inconsistent and hard to use. I find it much easier to just use a subset of HTML. It's also much easier that way to just slap a stylesheet on it and make it look nice.

Are images [img.jpg(alt text)] or (img.jpg[alt text])? I forgot. <img src="img.jpg" alt="text"> is way easier to remember. Is it ''bold'' '''italic''' or "bold" 'italic' or 'bold' ''italic''? <b>bold</b> <i>italic</i> is MUCH more intuitive.

Intuitive but involves a lot of typing. I find it much easier to remember “square for text, round for attachment “ than write/read through so many html tags. Many text editors, including MacDown for instance, have a split view so you can hardly get it wrong more than twice.

So much so that a Language Designer is part of the standard toolchain https://jtree.treenotation.org/designer/

Reply on news.ycombinator.com