Syntax and API fine tuning

1. Binary object syntax changed from {: ... } to [# ... ].
2. Array-like APIs, like map(), filter(), reduce(), etc., have been made to work when the 'length' property is set on the Mark object.

Syntax and API fine tuning

1. Changed .length() to .length, so that Mark object is now array-like, thus functional libraries that work with array-like data can also work with Mark object without change.
2. Added Mark.lengthOf() to get content length of a Mark object. This is useful when the .length property is defined on the Mark object.
3. Introduced double ?? escape to Mark pragma.
4. Mark stringify() now omits comma.

Syntax fine-tuning

Some minor syntax fine-tuning:

1. String value need not be quoted if it is an identifier.
2. Block comment /* ... */ can now be nested.
3. Added an alternative pragma syntax (? ... ?).
4. Content normalization for

Changed pragma delimiters

1. Changed Mark pragma delimeters from back ticks `...` to round brackets (...).
2. Made the symbols used by Mark public.

Simplified pragma syntax

In this beta release, Mark pragma syntax has been simplified from {...} to `...`.
As its syntax is now clearly distinguished from Mark and JSON object, it's parsing and usage is much simpler.

Binary data support and separated mutative API

1. Binary data can now be encoded in base64 or ascii85;
2. Mutative API functions are now separated into sub-module mark.mutate.js;
3. Added new API functions .each() and .text();
4. Added Mark('{source}') shorthand support;
5. Added XML and HTML stringification indentation support;
6. Some minor bug fixes and code clean up;

New release to support optional comma

A minor release to support optional comma in Mark.