Comparing Oberon+ with Oberon-2 and Oberon-07

(Updated 2023-12-27)

The Oberon+ language definition started as a union of the Oberon-07 and Oberon-2 language definitions. Since both Oberon-2 and Oberon-07 are an extension of the first Oberon version (1987 and 1990), the Oberon+ language definition therefore also includes the original Oberon language. This means that each program written in Oberon up to Oberon-2 and including Oberon-07 is compatible with the Oberon+ compiler.

Oberon-2 added the following elements to the original Oberon:

• Type–bound procedures
• Read–only export
• Open arrays
• Extended WITH statements (with variants)
• FOR statements

Oberon-07 (revision 3.5.2016) also describes itself as an extension of the original Oberon (thus ignoring most of the Oberon-2 extensions); here is a list of the major changes in Oberon-07:

• Removed the LOOP and EXIT statements
• Added the FOR statement
• Added the optional ELSIF part to the WHILE statement
• Removed the WITH statement
• Extended the CASE statement so that the case variable may also be a record or pointer to record type (“type CASE”, as a replacement of WITH)
• RETURN is no longer a statement but part of the procedure body (which supports RETURN without BEGIN)
• Support direct array and record assignments, making the COPY procedure obsolete
• Variables imported from other modules are now read-only
• Removed the concept of type inclusion, as well as the SHORTINT and LONGINT types
• TRUE and FALSE are keywords, no longer global constants
• Removed access to local variables and parameters of outer procedures from nested procedures (“non-local access”)

Since Oberon+ includes a union of Oberon-2 and Oberon-07, all removals done by Oberon-07 are ignored. The resulting redundancy with WITH and type CASE is tolerated in favor of compatibility. To accomodate the effect of the Oberon-07 RETURN, procedure bodies can either consist of BEGIN plus statement sequence, or a RETURN statement. Oberon-07 programs with read-only access to imported module variables still run on Oberon+ without modification (they simply do without write access); if the author of an Oberon-07 module wants to enforce read-only exports of the module variables, visibility mark must be changed to “-“.

In addition to the above, Oberon+ adds the following:

• Lowercase keywords and base type identifiers
• Simplifications and shortened keywords (^ instead of POINTER TO, [] instead of ARRAY OF, etc.)
• Underscores in identifiers
• All semicolons optional, most commas optional
• UTF-8 source files with Unicode string literals and comments
• WCHAR type
• String concatenation using the + operator
• Line comments
• Import paths, see this post
• Flexible declaration sequences, more than one CONST, TYPE, VAR or IMPORT section in arbitrary order
• DEFINITION modules
• Enumerations
• Procedure local records with type-bound procedures
• Delegate procedure types (procedure types for bound procedures)
• Constant VAR parameters using the IN keyword
• ANYREC, implicit base type of each record
• Generic modules, see this post, with support for literals and procedure references as generic arguments
• FFI language for cross-platform C shared library integration, see this post
• Variable length arrays (VLA)
• Conditional compilation (as recommended by the Oakwood guidelines)
• Exception handling, see this post

Features in evaluation:

• Array and record literals
• COMPLEX type, as defined in the Oakwood guidelines, including standard libraries
• Covariance of overloaded method return types and fields (like Component Pascal)
• Concurrency, see this post

• Code Examples using traditional Oberon syntax
• Motivation for a new Oberon version
• Comparing Oberon+ with Oberon-2 and Oberon-07
• Considering Generics
• Considering Imports
• Towards an Oberon+ Foreign Function Interface (FFI)
• Towards Oberon+ Exception Handling
• Towards Oberon+ Concurrency