4 ms·

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 wit

by abeppu 6y ago

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.

TuringTest 6y ago
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.
- lambda_obrien 6y ago
  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".
  - TuringTest 6y ago
    I think the Semantic Web was supposed to be that. Yet it was waaay too abstract for the majority of people, even programmers.
  - joshxyz 6y ago
    Schema as a Service here we goo
  - breck 6y ago
    Schema.org is probably the closest thing to that. I call the concept "Type the World" (https://breckyunits.com/type-the-world.html 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 https://jtree.treenotation.org/designer/#standard%20wwt
timClicks 6y ago
So much so that a Language Designer is part of the standard toolchain https://jtree.treenotation.org/designer/ https://jtree.treenotation.org/designer/
dheera 6y ago
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''? bold italic is MUCH more intuitive.
- njsubedi 6y ago
 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.

by abeppu 6y ago

TuringTest 6y ago

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.

lambda_obrien 6y ago
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".
- TuringTest 6y ago
  I think the Semantic Web was supposed to be that. Yet it was waaay too abstract for the majority of people, even programmers.
- joshxyz 6y ago
  Schema as a Service here we goo
- breck 6y ago
  Schema.org is probably the closest thing to that. I call the concept "Type the World" (https://breckyunits.com/type-the-world.html 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 https://jtree.treenotation.org/designer/#standard%20wwt

timClicks 6y ago

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

dheera 6y ago

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''? bold italic is MUCH more intuitive.

njsubedi 6y ago
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.