Threaded Mode | Linear Mode

compsystems · 09-30-2016, 08:45 PM

(09-30-2016 03:16 AM)compsystems Wrote: 1: Use an operator optional as end line, semicolon (; ) for example.

Persistent comments also preserve the newline at the end. This means all you have to do to keep it formatted is add a comment at the end:

Code:

<< @@ This is a persistent comment, which also can be empty as in the next line

DUP @@

>>

> but if the source code has no comments, the formatting is lost, it is necessary to finalize the "sentences" somehow, the semicolon is best

(09-30-2016 03:16 AM)compsystems Wrote: 2: Do not delete comments to store programs

Already explained by Guenter, just use @@ for single line permanent comment.

>OK

(09-30-2016 03:16 AM)compsystems Wrote: 3: use a different delimiter between variables & symbolic expression, for example a polynomial 'x^1' -> 'x' would be treated as algebraic expression and non-variable

It would break compatibility too much.

>partly yes, but this is a serious in handling algebraic expressions problem, create patches within the code to avoid ambiguity. the idea is to work more naturally, separating variables of algebraic expressions

the hp49/hp50 if include this new delimiter, but it is useless because it changes the delimiters try
`X^1` EVAL returns 'X'

(09-30-2016 03:16 AM)compsystems Wrote: 4: use "ENDIF" "ENDCASE" "ENDWHILE" "ENDDO" ETC as option to visually and detect the body block, also this allows easier code converters

This already exists. You can code using ENDIF, ENDCASE, ENDWHILE, ENDDO, and you can also distinguish THEN, THENERR and THENCASE. The code can decompile to either END or all different forms of end. Right now it's set to END and THEN, so even if you code with ENDIF, it will display END when you decompile it. Back when I did this I didn't have flags implemented yet, and then I forgot all about it. I need to find a flag to configure this option.
The accepted constructs are:

Code:

IF ... THEN ... ELSE ... ENDIF

IFERR ... THENERR ... ELSEERR ... ENDERR

CASE

... THENCASE ... ENDTHEN

...[default]

ENDCASE

DO ... UNTIL ... ENDDO

WHILE ... REPEAT ... ENDWHILE

> OK

(09-30-2016 03:16 AM)compsystems Wrote: 5: A flag to maintain always a complex number in format (x, y) dupla

(x, y) [ENTER] (x, y) and not x+y*i and not (r, angle theta)

This already exists, without any flags. A number stays the way you entered unless you explicitly convert it to other form.

> but when the expression is simplificque is preserved?

(09-30-2016 03:16 AM)compsystems Wrote: 6: a new type of data is needed tripla (x,y,z) for geometric objects

??? Please explain better, I don't understand why you would need a separate type, you can use vectors for that.
Complex numbers are not vectors because i^2=-1, (being 'i' considered as the unit vector for the second component of the dupla). This is inconsistent with a standard vector space, hence it needs different rules than a vector. All other nth-uplas are the same.

> for example geogebra has a specific data type to tuple, as a point pairs (x,y) and triples for 3D points (x,y,z), vectors can be used, but the idea is to be more explicit in the source code. compared with hprime that is very difficult to code the lists, in some cases are lists [] -> {} in other vectors [] and other matrices [[]]

Threads

http://www.hpmuseum.org/forum/thread-6669.html
http://www.hpmuseum.org/forum/thread-6628.html

(09-30-2016 03:16 AM)compsystems Wrote: 7: Local storage in RPN

This also already exists.

> OK

8: PART is very important to analyze the input expression in a function, it must be included in the ROM =)

Thanks