rtsi.com/2014.01.www.rtsi.com.tar/www.rtsi.com/OS9/OSK/DEMOS/UB_68020.LZH/unibasic.faq

Text File | 1996-05-06 | 9KB | 170 lines

Q: What is the history of UniBasic?
A: UniBasic is the C output version of CDL Basic which was a native code
   compiler for the OS9/68000 operating system. Since UniBasic now
   produces C code, it is now quite portable to most any platform.

Q: What language was UniBasic written in?
A: UniBasic was written in UniBasic. The early version was written in
   its predecessor CDL Basic. This is not unusual in itself in that most
   C compilers were written in C. What is unique is having a Basic that
   is powerful enough to be used to compile itself.

Q: What can UniBasic do that other Basics cannot do?
A: In general, the answer is produce systems programs. Attempting to
   write systems code in most Basics usually results in the use of
   PEEKing and POKEing and other "tricks". UniBasic has all the power
   and utility of C but without the cryptic and often terse syntax of C.
   These include pointer variables, dynmamic buffer allocation, internal
   and external functions, function pointers, conditional compilation,
   include files, macros, structures, compiler variables (manifest
   constants), command line arguments, and others.

A: What does UniBasic have that C does not have?
Q: The first thing is that UniBasic is easier to learn and use than C. The
   second is that UniBasic provides more meaningful (and accurate) reporting
   of compile time errors than most C compilers. The third is full STRING
   capability (in UniBasic, STRING is a bonafide data type not an external
   afterthought). The fourth is VECTOR (based) variables. These are similar
   in concept to pointers, but as easy to use as ordinary variables. The
   fifth is portability. While C made portability famous, data structures
   in one C and another (even on the same computer) are often "padded"
   differently (just do a sizeof and see for yourself). In Unibasic, all
   data structures are declared as a char array with ALL internal management
   and referencing handled by UniBasic. The sixth is the ability
   to exchange and transfer all data types (including arrays) in a single
   statement. Exchanges are especially useful for sorting data.
   UniBasic has all the features of Basic which means it is a higher level
   language than C. (C is variously referred to a a mid-level langauge or
   universal assembler). This unique set of features makes UniBasic
   equally at home at systems as well as applications programming.

Q: Since UniBasic produces C code, do I need a C compiler also.
A: Yes. The type of C compiler required depends on the system being used.
   In most cases there is a GCC compiler available from the Free Software
   Foundation which is an fine compiler. UniBasic automatically invokes
   the C compiler if you have the standard one for that system. If not,
   the Basic program may be invoked with the -c option which causes the
   compile process to stop with the generation of the .c file. This .c
   file may then be used as input to your C compiler.

Q: Is there any runtime package required?
A: No. Unlike interpreters and I-Code compilers, the output of UniBasic
   is compiled by your C compiler to pure machine code which is not only
   fast and efficient but requires no runtime package.

Q: Since UniBasic produces C code, do I need to know anything about C?
A: No. In fact, unless you invoke the -v (verbose) option, you will not
   even see the C compiler being called.

Q: How does UniBasic implement Basic subroutines?
A: Subroutines are implemented as void functions.

Q: Does UniBasic support structured programming techniques?
A: Yes. Various control structures are available including FOR/NEXT,
   WHILE DO/ENDWHILE, REPEAT/UNTIL, IF/ELSE/ENDIF, LOOP/EXITIF/ENDEXIT/
   ENDLOOP.

Q: Do I have to use structured programming techniques?
A: No. You can make your program as structured or unstructured as you
   wish (or anywhere in between).

Q: Will programs I write in UniBasic on Linux run on MSDOS?
A: Yes, if you recompile the program on the MSDOS system.

Q: Is the C code produced ANSI compatible or K&R compatible?
A: Both. The K&R must have the "void" extensions (which most do). The
   input/output operations make use of the unbuffered calls such as
   read(), write(), etc. (as opposed to fread() and fwrite() etc.) which
   means that your C compiler must have the System V (or Posix) extensions
   (most all do).

Q: IF UniBasic is as powerful as C, then how is it easer to learn and use?
A: First, as a beginner on UniBasic, you can use the more traditional
   Basic features and work your way into the more advance features as you
   progress. Second, the advanced features are easier to learn and use than
   those in C and the syntax is more English-like than C. Also, while
   UniBasic does have pointers like C, the Vector (based) variable allows
   the same power as the pointer without any pointer syntax at all.

Q: Is UniBasic able to use the vast array of C functions and libraries on
   the market and in the public domain?
A: Yes. Some of the C "string" functions may require careful consideration
   but there is already a full array of STRING functions included with
   UniBasic as an integral part of the language.

Q: I notice there are is no graphic capability built into UniBasic. How
   do I do graphics.
A: Since Unibasic can call C funtions equally as well as C can, you have
   full access to the C graphic libraries for your system.

Q: Of what practical use is the conditional compilation capability?
A: One example of this is the UniBasic compiler itself. While there are
   a variety of versions available (regular and demo) (Linux and MSDOS) etc.,
   there is only one master source file in which any differences in the
   various versions are enclosed in conditional code blocks. When a version
   is compiled, the command line specifies (by setting a compiler variable)
   which version will be compiled. You might want to use this concept to
   produce different versions of your own program. Different versions might
   include special features or different language prompts etc.

Q: What is the advantage of named labels over line numbers?
A: Line numbers convey no intrinsic meaning while named labels can be
   quite desciptive, thus enhancing readability. For example, GOSUB 1435
   gives no clue, while GOSUB PrintChecks is quite clear.

Q: What are compiler variables?
A: Compiler variables are named constants (referred to also as manifest
   constants). Like named labels, these increase a program's readability.
   These are not to be confused with program variables but are, in fact,
   named literal values. They are especially useful for naming flags and
   indicators. For example, after setting a named constant (#SET Red=22),
   it has a clearer meaning to write Color=Red than to say Color=22.

Q: How do I look at the C code that UniBasic generates?
A: Just compile your program with the -c command line option. Then list out
   the resultant .c file. Each UniBasic statement will appear as a C
   comment followed by the resultant C code. If you wish, these comments may
   be omitted by using the -r option with the -c option (-cr).

Q: What is dynamic buffer allocation used for?
A: Sometimes when you are writing a program, you have no way of knowing
   how large to diminsion an array. You won't know this until the program
   is run and you can (for example) check the size of an input file. In
   this case, you can check the size of the file (using the FILSIZ()
   function), request the required amount of memory, and set a vector (based)
   variable to the memory buffer address. You can then operate on the array
   just as if it were diminsioned the standard way. In standard Basics this
   problem is solved by diminsioning an arbitrary array size, and then if the
   arbitrary size is too large, you waste memory space, and if it is too
   small, the data must be operated on a little at a time.

Q: Do I have to worry about the fact that in C, arrays start with 0?
A: No. UniBasic automatically compensates for this so that your arrays can
   start with 1. There is a BASE statement in which you can select either
   0 or 1 as the base fore arrays. The default is 1.

Q: Is there bounds checking on arrays?
A: No. This speeds up execution speed but does require some caution on the
   part of the programmer. Strings are bounds tested so that if you input
   22 characters into a 20 character string, the worst that happens is the
   string is truncated. The next variable in memory is NOT clobbered as in
   some cases with C.

Q: Can UniBasic be used to write system utilities?
A: Absolutely! Since UniBasic progams end up as pure machine language
   programs you can write your own system utilities and use them just
   like the ones that came with your system. You can even "read" options
   from the command line and use these options to control the flow of
   your program.

Q: How can I treat the same variable as more than one type?
A: You can diminsion vectors or pointers of various data types and then
   set them to the same address. For example it is possible to diminsion
   x as a long (32 bit integer), y as a 4 byte array vector, set y to the
   address of x and then operate on the same data as either type.

Q: What is the distinction between "program" and "function" in UniBasic?
A: Program refers to the main body of code which in C is the main() function.
   while function refers to any non-main function in C.