does try really need to be a keyword?

[Gribouillis]

i could, instead, argue, that ":" isn't needed, in general (and can be optional for enhancing readability for humans)

I proved that "." is not needed in python by implementing a program that adds the dots automatically. If you also remove the ":" and the fact that "try" is a keyword, it may stop to work. I'd much prefer a python without "." than a python without ":".

[Skaperen]

FYI i'm working on the design of a data format using indentation. it is a potentially new data format for recording genealogical data.

---

how do you omit the "." in your trial branch of Python? just leave white space? if i can make a trial branch without ":" then maybe i can make one without both.

i'd like to see the program that adds the dots automatically. are you giving it away?

---

i'd like to leave out the dots in COBOL.

[Gribouillis]

i'd like to see the program that adds the dots automatically. are you giving it away?

I don't publish the whole module yet but I can give you the heart of the algorithm that adds dots to a python source code. You need to install module 'untokenize' from pypi first.

With this function, the dots become optional in the source code. The function adds only the missing dots.

I supplemented that with a code that watches directories with the help of the inotify module and transforms file in real time as one edits them. I don't publish this part know.

# module dotless.adddot

from keyword import iskeyword 
from tokenize import NAME ,OP ,STRING ,COMMENT ,NL 
from tokenize import generate_tokens
# uses module untokenize from pypi
# https://pypi.org/project/untokenize/
import untokenize
TokenInfo = type(list(generate_tokens(lambda: ''))[0])

def dot_augmented(itoken):
    # originally inspired by the author's quail.texmacs.pythonFluide module
    """Add missing 'dot' operator tokens in a sequence of python tokens
    """
    insert_dot = False
    end_lin, end_col = 0, 0
    for tok in itoken:
        if tok.type == NAME:
            if iskeyword(tok.string):
                insert_dot = False 
            elif insert_dot:
                lin, col = tok.start
                start = lin, col - 1 if end_lin == lin else col
                yield TokenInfo(OP , ".", start, tok.start, tok.line)
            else:
                insert_dot = True 
        elif tok.type == STRING or (
            tok.type == OP and tok.string in ")]}"):
            insert_dot = True
        elif tok.type not in (COMMENT ,NL ):
            insert_dot = False
        yield tok
        end_lin, end_col = tok.end

def adddot(infile):
    """Add dot operators to python code.
    
    take a file object, return a string.
    """
    itok = generate_tokens(infile.readline)
    itok = dot_augmented(itok)
#    itok = list(itok)
#    for t in itok:
#        print(t)
    return untokenize.untokenize(itok)

[Skaperen]

when it processes code does it care about indents or just pass them along as it gets them?

[Gribouillis]

It doesn't need to care about indents. I mean the 'untokenize' module cares about them but adddot doesn't. Here is an example (taken from pymotw 3)

import io
# this is PYTHON CODE WITHOUT DOTS
code = r'''
def show_tree(tree, total_width=36, fill=' '):
    """Pretty-print a tree."""
    output = StringIO()
    last_row = -1
    for i, n in enumerate(tree):
        if i:
            row = int(math floor(math log(i + 1, 2)))
        else:
            row = 0
        if row != last_row:
            output write('\n')
        columns = 2 ** row
        col_width = int(math floor(total_width / columns))
        output write(str(n) center(col_width, fill))
        last_row = row
    print(output getvalue())
    print('-' * total_width)
    print()
'''

result = adddot(io.StringIO(code))
print(result)

Output:
def show_tree(tree, total_width=36, fill=' '):
    """Pretty-print a tree."""
    output = StringIO()
    last_row = -1
    for i, n in enumerate(tree):
        if i:
            row = int(math.floor(math.log(i + 1, 2)))
        else:
            row = 0
        if row != last_row:
            output.write('\n')
        columns = 2 ** row
        col_width = int(math.floor(total_width / columns))
        output.write(str(n).center(col_width, fill))
        last_row = row
    print(output.getvalue())
    print('-' * total_width)
    print()

As you can see, it added the dots at their proper places without modifying the indentation.

I should modernize the code one day or the other because generate_tokens() is deprecated in favor of tokenize() which takes bytes instead of text... A minor change.