It seems magrittr is not used in the function bodies. Shall it be moved to Suggests:?

Use portable() to attach package scope to class's parent env chain.

myThing <- portable(type(function(a, b){
                add <- function() a + b
}))
## or 
myOtherThing <-  function(){
             value <- 1
           } %>%
          type() %>% 
          portable()

May I ask why this definition is OK

size <- Q7::type( function(size){ }, s3 = "ggsd.size")
x <- size(1)

But below definition result in error.

type_size <- function(size){ }
size <- Q7::type(type_size, s3 = "ggsd.size")
x <- size(1)

@nanxstats

Birds generally fly. Turkey doesn't.
Mammals generally don't fly. Bat does.

canFly() feature can be added to Bat() and not be added to Turkey()

retain class after using extend() -- also, rename extend()?
offer a feature()/implement() friendly way to manipulate S3 classes.

use utils::getParseData() or {xmlparsedata} to....handle binding modifiers? Warn erroneous/harzardous implements? Explore.

require(Q7)

BaseType <- type(function(a){
  b <- 1
  f <- function(){
    a + b
  }
})

bt <- BaseType(2)
bt$f()

ExtendedType <- type(function(b){
  extend(BaseType)(b)
  g <- function(){
    b
  }
})

et <- ExtendedType(3)
et$b #1
et$f() #2
et$g() #1

Users may have assumed of the opposite.

the name is appended to the "feature" attribute of the main obj.

That calls the print() method of an object.

Locks all bindings, but not the envir itself, so new bindings can be added.

myThing <- myThing %>% immutable()

option to lock only functions?

`$.extend(object_a, object_b)

https://github.com/r-lib/evaluate

keeps comments, formats, etc.

if ...s3 = "Foo" then the S3 class of the type would become Foo_type.
This fully differentiates type and object created from the type, but how does this affect features applied to a type?

Create an environment in the class obj's attr.
Refer to this envir with active bindings in the public/private envirs of the obj.

Despite the potential hazard, pay no regard to the consequence of envir serialization now.

put into .private$.s3