Retro Basic

The final project

Aim:  you build a compiler that reads input (Basic) file and output B-code. 

Due date  Thursday 9 April 2026, at 11pm.

To do list:

1)  Write Grammar for this Retro Basic language.
2)  Choose your method to write compiler.
3)  Build compiler.  Provide me with the source and executable file so that I can test your compiler
4)  Write a 3-4 pages report.  Describing your compiler (what method you use, what tool you use, what is your internal data structure for the parse tree etc.)

How do I grade your project

1)  I will test it with 5 unseen Retro Basic files.  It is pass/fail result for each test file. (score 0..10)
2)  I will read your source (to make sure your executable is from this source).  (score 0..10)
3)  I will read your report:
3.1)   The method is of your compiler is sound and correct.  (score 5..10)
3.2)   Extra score is you have some interesting idea or observation about your compiler. (+5)

total score 30 (+ if extra)

I will not "judge" how good/bad your programming skill is or the efficiency of the compiler.  As long as your method is correct and the compiler produces the correct output. It is good.  Please put effort in writing a report.  I am a human who must read it.  So, my bias is to give good score to a well written report: good format, fluid writing etc. Of course, total AI generated report will have not-good score.  Don't write too long a report, I have not much time to read it.

Here is the language

in the memory of Microsoft Basic (4K Basic)

A program to print 1 to 10

10 A = 1
20 IF 10 < A 60
30 PRINT A
40 A = A + 1
50 GOTO 20
60 STOP

A program to sum 1 to 10

10 A = 1
20 S = 0
30 IF 10 < A 70
40 S = S + A
50 A = A + 1
60 GOTO 30
70 PRINT S
80 STOP

A program in Retro Basic consists of lines.  Each line starts with line_number follows by a statement.  Statements are a) assignment b) if c) print d) goto e) stop.  An assignment is "id = exp" where id is {A..Z}.  An expression is binary op +/- between id and constant.
An if statement is "IF cond line_number", where cond is binary op {<,=} between id and constant.  A print statement is "PRINT id".  An goto statement is "GOTO line_num". A stop statement is "STOP".  A constant is {0..100}. line_num is {1..1000}

Your job is to write a compiler to translate the source of Retro Basic to the "intermediate code".  Here is the intermediate code (B-code).

B-code is stored in an array of cells (32-bit) where the input source is tokenised (to help speed up the run-time interpreter).  Each token consists of two consecutive cells: type, value (a tuple). 

line_num    {#line, num}     num is {1..1000}
id          {#id, ref}       ref is index 1..26 corresponded to A..Z
const       {#const, value}  value is {1..100}
IF          {#if, 0}
GOTO        {#goto, num}     num is {1..1000}
PRINT       {#print, 0}
STOP        {#stop, 0}
+           {#op, 1}
-           {#op, 2}
<           {#op, 3}
=           {#op, 4}

Example of the translation of a line into B-code

10 A = 1
{#line, 10} {#id, A} {#op, 4} {#const, 1}

30 IF 10 < A 70
{#line, 30} {#if, 0} {#const, 10} {#op, 3} {#id, A} {#goto, 70}  

***  line_num became "goto"

Here is the coding of B-code type:

#line   10
#id     11
#const  12
#if     13
#goto   14
#print  15
#stop   16
#op     17

I will provide an B-code lister so that you can "run" your compiler and check the output.  The lister will "reverse" your output and print the Retro Basic.  You use it to check that your compiler outputs the correct B-code.  The input to B-code lister is a sequence of integer representing the B-code cell, with 0 represents the end of file.

For example the above two lines

10 A = 1
30 IF 10 < A 70
...

B-code

10 10 11 1 17 4 12 1
10 30 13 0 12 10 17 3 11 1 14 70
...
0


Depend on your programming skill, this project can take a few days up to a week of programming.  You can check your output B-code with my lister.  You and use any tools to make your project.

Due date  Thursday 9 April 2026, at 11pm.

Lister

Here is a tool (lister.zip ) to print Retro Basic program from B-code (a listing).  Use it to check the correct output from your compiler. The lister package includes an executable file (for windows platform. I am building a MacOS version and will upload it soon).  Here is a sample of the session to use lister.

Assume this is a Retro Basic program to print 1 to 10:


10 A = 1
20 IF 10 < A 60 
30 PRINT A
40 A = A + 1
50 GOTO 20
60 STOP

It is translated into B-code like this (showing tuples of B-code):

10 A = 1
(10,10) (11,1) (17,4) (12,1)
20 IF 10 < A 60 
(10,20) (13,0) (12,10) (17,3) (11,1) (14,60)
30 PRINT A
(10,30) (15,0) (11,1)
40 A = A + 1
(10,40) (11,1) (17,4) (11,1) (17,1) (12,1)
50 GOTO 20
(10,50) (14,20)
60 STOP
(10,60) (16,0)

The valid output B-code is (assume it is in the file "print1-10.txt") :

10 10 11 1 17 4 12 1
10 20 13 0 12 10 17 3 11 1 14 60
10 30 15 0 11 1
10 40 11 1 17 4 11 1 17 1 12 1
10 50 14 20
10 60 16 0
0

Use lister:

c>basic\lister\lister print1-10.txt

10 10 11 1 17 4 12 1 10 20 13 0 12 10 17 3 11 1 14 60 10 30 15 0 11 1
10 40 11 1 17 4 11 1 17 1 12 1 10 50 14 20 10 60 16 0 0


10 A = 1
20 IF 10 < A GOTO 60 
30 PRINT A
40 A = A + 1
50 GOTO 20
60 STOP

c>basic\lister\

Enjoy

last update 25 March  2026