Page 2 of 2

Posted: Thu Sep 20, 2007 10:43 pm
by Nodtveidt
FB programs will run faster on modern computers than QB programs on modern computers, and in most cases will run faster on older computers as well. One of the most obvious reasons is that QB targets the 8086 processor, whereas FB was designed from the ground up to be run on modern CPUs. FB is ALWAYS going to be faster than QBasic in terms of execution speed on modern computers for the simple fact that QBasic is an interpreter (and a rather slow one at that) and FB compiles to a native executable. From what I understand, QB doesn't even compile to a proper executable but more of a bytecode stub that runs in BRUN.

Obviously, FB programs aren't going to run on a 286 or lesser CPU, whereas QB programs generally will. But who's using such obsolete systems these days? If you're looking to develop performant applications on modern systems, you're not going to look at QBasic or even QB. You can only push the limits of QBasic so much; you're not writing a realtime hi-resolution 3D engine in QBasic anytime soon...not until they invent 9GHz processors that run in 8086 mode. It simply doesn't have the calculation speed, nor can it take full advantage of modern video hardware. In FB, this isn't difficult to do, and has already been done dozens of times. Windows GUI applications, linear framebuffer SVGA modes, applications that run natively in Linux...all things QBasic could never do that are completely brainless in FB. About the only straw someone could grasp for is that QB 7.1 can develop applications that run in OS/2...but that's a major straw.

Lazy, stupid programmers learn to code in C# and buy 3GHz computers to make programs that are less functional and efficient (yet more flashy!) than similar programs written in C twenty years prior on 386-based machines running Windows 3.1. But thinking that "just code it better!" is the solution to the slowness problems of QBasic is silly when it's been shown countless times over the years that it is slow, inefficient, and above all else, limited in what it can pull off by itself. QB removes some of those limitations but even QB has limits that cannot be overcome. It's a harsh reality that a select few people are having a very difficult time coming to terms with.

The beauty of QBasic is in its simplicity. The beauty of FB is in its raw power. The beauty of this post is in its truth.

Posted: Thu Sep 20, 2007 11:00 pm
by Nodtveidt
Oh by the way...that Monopoly program doesn't compile in QB 4.5:

Code: Select all

BC C:\QB45\MONOPOLY.BAS/O/T/C:512;
Microsoft (R) QuickBASIC Compiler Version 4.50
(C) Copyright Microsoft Corporation 1982-1988.
All rights reserved.
Simultaneously published in the U.S. and Canada.
 008D   000C    CONST GO = -2
                      ^ Syntax error
 4025   0DBA    CASE GO
                     ^ Syntax error
 6584   0DBA    END IF
                    ^ Subprogram error
 69DD   0DBA      END IF
                      ^ Subprogram error
 A095   0DBA    Setup 0, GO, "?GO"
                         ^ Syntax error

42029 Bytes Available
11787 Bytes Free

    0 Warning Error(s)
    5 Severe  Error(s)


Press any key to continue

Posted: Fri Sep 21, 2007 6:30 am
by Mentat
That's what I wanted to know. :)

But I can't call QBASIC slow becuase I came from TI-B.

Code: Select all

0->x
Lbl 1
Disp x
x+1->x
Goto 1
Cycles once per a tenth of a second.
QB counts to somewhere in the tens of thousands by the time TI gets to 10.

But:

Code: Select all

For (X,1,1000)
Output(1,1,X)
End
Runs 100 cycles per second (in TI 83 SE). So optimization can make a big difference. Maybe not enough, but still a difference.

Posted: Fri Sep 21, 2007 5:53 pm
by Patz QuickBASIC Creations
Mentat wrote:That's what I wanted to know. :)

But I can't call QBASIC slow becuase I came from TI-B.

Code: Select all

0->x
Lbl 1
Disp x
x+1->x
Goto 1
Cycles once per a tenth of a second.
QB counts to somewhere in the tens of thousands by the time TI gets to 10.

But:

Code: Select all

For (X,1,1000)
Output(1,1,X)
End
Runs 100 cycles per second (in TI 83 SE). So optimization can make a big difference. Maybe not enough, but still a difference.
Leave off your parentheses.
*vomit*

Seriously, on a limited platform such as the TI-83(84)(+)(SE), you need every bit of optimization you can get.

Like when people do...

Code: Select all

900->dim(L6
For(A,1,900
A->L6(A
End
makes me absolutely sick. This can be optimized to:

Code: Select all

seq(A,A,1,900->L6
Or, even better,

Code: Select all

cumSum(binomcdf(899,1->L6
Optimization is everything in TI-BASIC

Posted: Fri Sep 21, 2007 6:21 pm
by Mentat
I usually do leave off parenthesis.

But you forgot something pretty vital to memory
Patz QuickBASIC Creations wrote:

Code: Select all

 900->dim(L6 
For(A,1,900 
A->L6(A 
End 
is better of as

Code: Select all

 GarbageCollect      'its under Catalouge. Saves Memory and RAM.
Unarchive LMYLIST    'just in case it's archived 
900->LMYLIST                   'the calc makes the list if it can't find the name
For(A,1,999                       '999 is upper limit of list size
PrgmFunction               'Some Sub that returns the function of A as B
B->LMYLIST(A                      'B can be a set of changing data from A     
End
Archive LMYLIST        'Saves it to ARC, thus RAM acts like it isn't there
A list of 900 takes up a lot of space.

And many of my programs are modelled more or less this way. Maybe not as fast, but I don't run out of memory and it's more flexible.
TI has too little RAM. Thus Archiving saves valuable memory, and protects data.

But in QB (QBASIC or QuickBASIC), the way to 'archive' is to save to something else and then write over the variables and/or strings.

Posted: Sat Sep 22, 2007 10:31 am
by Patz QuickBASIC Creations
Wow, obviously you didn't read my alternatives. The For( loop area could be removed and replaced, it's slow.

Anyway, if the list is in the archive, how do you plan to access it? I'd like to hear that one.

Posted: Sun Sep 23, 2007 9:42 pm
by moneo
Stoves wrote:Anyone happen to know if the source code for QBasic or QuickBasic (interpreter or compiler) has ever been leaked? and if so, where it might be found?
Stoves,

Even if you had the original source code for QBasic, you would also need the corresponding original version of the C compiler, plus any libraries used back then, and also the assembler. Having all these exact elements would be miraculous. Another variable might be the actual operating environment, the hardware and the original operating system, used at the time to produce the executable code of QBasic.

The bottom line is that even given limitless time and money, I don't think you could ever put together a comparable working version of QBasic, let alone make enhancements to it.

Regards..... Moneo

Posted: Tue Sep 25, 2007 9:28 pm
by Nodtveidt
The complete QBasic sourcecode is a total mess. It would take quite an effort to build. But if it could be done, it could likely be improved significantly if someone has the skill level to do it.