Hi Guys!

Anyone interested in a starting BASIC interpreter? I've been around a while,
and came across this one recently while looking into a C cpackage called
Micro C by Dave Dunfield. Micro C is along the lines of BASM286 and Moonrock
BASIC - ie: it put's out assembly code and then requires that YOUR assembler
and linker do the final assemble and link to get eithr a COM or EXE (Small
model) file. It is more hands off than either BASM or Moonrock however.
Makes some nice tight & small COM/EXE's to boot!

Anyway, for one of it;s examples it has a decent little BASIC interpreter
written in Micro C syntax (90% standard C syntax). I found it very
interesting as an interpreter example. Other exampels include a terminal and
assembler (in Micro C).

If interested - www.dunfield.com for the latest version, or hit SIMTEL's C
directory.

Ep
Ed Parry - Southern California, USA

Unauthorized amphibians will be toad away.

        Well, that's a list (in no particular order) of bonuses I would
like to see in an interpreter, and it is by no means complete. I'm sure
I'll think of some more things, but that list was off the top of my head.
Wow! Now I'm really fired up about this!
        Peter

Also, has anyone thought of using The Artful Parser's EVALUATE.BAS
program for variable storage?

Can we build XMS support in here? I'm sure we can use Rick's routine
at www.qbasic.com...!

Lator gaters,

+--Terence Jordan(x7233)----------------+----------------------------+

|(parenthesis)                          |  Or in doubt,         |    |
|---------------------------------------|  Run in circles,    _/ \   |
|Views expressed are Terence's          |  Scream and shout!"    /   |
+-----and of no other.------------------|----------------------------+

Very good, if I understood you well then youre 'user modifiable
language' closely resembles what I said to make some frame/container
fast and let everyone put there desires into it ;). Using the word
'user modifiable language' has a lot of appeal though! But you are not
meaning the 'end-user' I guess ?

I repeat my desire for Ep to take up some decisions fast: what would
be the language to program in, will he be the committee( or at least
the dir ?), what will be desired to make the frame/container and who
will start with that. If we are really gonna do something like this( I
like to think of it as something like the OSdevil thing or the JOLT
thing under way. Of course more at our levels(g). Ep you might contact
Jan Zumwalt about how to direct such a thing if you are uncertain..).
Also I could make contact to two really great BASIC guru's who are not
into usenet, but might be willing to help us( yes Dan is one of them).

My proposition is asm for the language( all basic routines are already
there in STDLIB, or FREELIB, etc etc), and TASM 2.0 for the compiler
( its .386+ and you can get it with HOTBOT! go for it)

But I repeat myself: lets get started fast! Before we have a no_code
threat like GIU!

Rick

Lets get some minimal thing:
1) Editor: support for multiple windoze for subs/functions can be
postponed( see below).
qs:Linelength(256), linedelimiter(crlf?), filelength( unlimited ?)

There is for instance source for some older version of notepad
available which could be a good starting point. Later we only have
to add multiple windoze..Scrolling etc is already there, unlimited
filelength too.

qs:functions(like QBASIC ?), keyhandling procedure, mouse support(
lets not waste energy there at first..)

2)Intrepreter:  

START: what will be the signs for

COMMENTS : (' ?)
METACOMMANDS: (# ?)
add a metacommand for DATASPACE allocation
CODE ( nothing ?)
CODEPARTS( FUNCTION/ SUB ?)
DECLARATIONS: ((RE)DIM for variables, DECLARE for FUNCTIONS/SUBS ?)

2: DATASPACE:

Q: XMS or FLAT SUPPORT ? ( Flat needs XMS lock to)

PREALLOCATED XMS ON METACOMMAND

LENGTH(64 kB LOW + RAM CONNECTION for more ?)
STORAGE: 2/4 BYTE DESCRIPTOR: CONTENT LIKE
DD: XMSOFFSET&
4:DB* 512: LOWBUFFER FOR SWAP LOW/XMS

ELSE    
DW: VARIABLE LENGTH= 0..&HFFFF
2:DB*LENGTH:  

The intrepreter must decide at runtime how to intrepret the DATA( as
ramtype,byte integer, integer, string etc..etc.)

For array's something like:
DIM
RAMDIM
For variables something like:
%,&,#,$,!
only problem there: RAMSTRING: could be something like $$

The rest for DATA is a parsing routine and a case routine.

3: STACK

LENGTH(?)
SEE POINT 6 TOO..

4. CODE
PRESETTED SEGMENT REGISTERS TO 4gb FLATMODE
At this moment we really need flatmode!(routines available)

-Since it will not be pmode intrepreter I guess we have 16 bit code
,but we will use .386 code in a 32 bit segment.
-So we have to check for *.386 cpu( routine available), and for
dosbox( routine available)

1-The intrepreter puts the intrepreted code into our XMS locked
-segment available from real mode.
2-We do a far jump to the code and hopla: it gets executed..

-We need to unlock and deallocate XMS

5: COMMANDS

Lets start with some easy things( all available everywhere)
This is not the most difficult part!

6. SUB/FUNCTION SUPPORT

But this is! Later we might start with making a real intrepreter which
supports subs/functions.

Crucial for the lucidity of the intrepreter might be the way to pass
arguments. Normally this is done with some internal buffer/stack,
which becomes invisible for the end_user.

Also the way to define variables local., global needs careful
consideration. Thats another arrea where most intrepreters get
invisible.

A. PRIORITIES:
Since commands are basically available and easy, and the same holds
for the XMS/FLATMODE support that routines do only need some minor
adjustments.
I would strongly suggest that first we make a suitable minimal editor
WITH keyhandling support. At the same time we could start discussing
DATA issues, so that when the coding of the editor is fini, we can
start with the much more difficult DATA parts. At that time we can
split up so that a few will make the DATA part, others will do some
COMMAND parts, while all can start discussing SUB/FUNCTION issues.

I am afraid that Gunnar might not agree, but I think we
have a chance if we make/ maintain it for free with a group of
programmers, and do not set time limits. Just some goodie.

I've often wondered why someone hasn't developed an user
modifiable language. A language that would allow the user to add
functionality as an integral part of the language.

The essences of any higher level language is the executions of
functions. The value of any higher level language is it's ease
of use and it's hugeness of functionality. A higher level
language dies out if it can't put out.  

Why not make the language so the user can add functionality? The
package would an easy to use IDE and a user modifiable function
library. If you wrote IDE/LIB without mouse functions, then I
could write and add them to the language in such a way that they
would be an integral part of the language. I would then submit
my mouse functions to the committee and they would do their
committee thing. To not only provide a better mouse trap, but to
also provide the means to improve upon the mouse trap.

All of us have wish lists, mine is singular, a language that
will not die.

With that said, let the committee begin.

Have you ever written a compiler? This is really hard work, if you want to
have an example how someone shouldn't write an interpreter take a look at
the old ALPHA version of GBasic.

---
Gunnar Beushausen

http://www.hof.de/~gbasic

Ever heard of Forth?  It allows you to add keywords at will.  Even QB, PDS and VBDOS allow you to create libraries of functions and
subprograms that you can use like extensions to the language.  I created a series of MAT subs that are called like so:

DIM A(10)
MAT READ A
MAT PRINT A
DATA 4,1,6,7,9,1,0,3,2,5

See?  It looks like a part of the language when you call the SUB named MAT.  There ARE limitations, to be sure (the most annoying
is that you can't have a variable number of arguments unless you pass a string and parse it yourself:

y = MULT("3,5,8")
would set y equal to 120
and
y = MULT("7,4") would set y equal to 28

Tom Lake

Any language implementation that allows seperate compilation can be
extended.
Functions that are awkward to write in the target language can be
implemented in
assembly language and linked.  The QuickBasic IDE has an easy way to load
user-defined libraries, which don't have to be written in QuickBasic.

Forth is a small language in which you define 'words' which you use to
define more complex words, until you define the name of your application.
Forth is usually implemented as a threaded-code interpreter.  This gets
amazing performance from a tiny system; the Forth "inner interpreter" for
CP/M was just
six{*filter*} bytes of 8080 code.  That's not a typo. Bytes, not kilobytes.  Of
course,
the editor/compiler was bigger, but that did not have to be present at
runtime.
All this bang-per-buck has a price: the programmer must do much of the
work
normally done by a compiler's front end, such as managing the runtime
stack and
converting expressions to reverse Polish notation.

There was also a series of articles in Byte by Jhonathan Amsterdam, some
years
ago, on compiler construction.  They included C code for several compiler
parts,
plus a stack-oriented virtual machine similar to the "hypothetical stack
machine"
mentioned in the Pascal report.  Does anyone remember this series?  My
Byte
collection was destroyed by a flood.

First, one of my web pages has pointers to source code for about a
dozen Basic interpreters: <http://www.nicholson.com/rhn/basic/>

---

#include <canonical.disclaimer>   // only my own opinions, etc.

As a professional in the fields of computer programming and administering, I
would like to voice my opinion on this matter.  I don't mind if the program is
interpreted or compiled as long as it can access databases that are already
considered standard.  SQL, DBase, MDB...  For me, MDB is the most important.
These programs must also be multi-user accessible.

I occasionally read these newsgroups for any usable information that I might be
able to apply to real life circumstances.  I've found that we have an enormous
amount of talent in this group and others as well.  Unfortunately, most of this
talent doesn't seem to know how to apply it to a useful means.

There are many problems that need solving in today's business world.
Individually we know how to make something.  Software is getting more complex
every day.  To stay in the paying market we may need to start pooling our
talents to a greater project.  We need project leaders with real solutions to
real problems.

One such problem that I am about to undertake (if things don't straighten out on
their own) is a general purpose program that will allow writers (books and
articles) to sell their works over electronic methods.  Writers have this
problem of getting books published without getting ripped off.  They are in the
business of selling information.  It just so happens that information is more
difficult to peddle, than ICE in alaska!

Novels are created on computers around the globe, then sent to publishers for
manufacturing and distribution.  It sure would be nice if there were a way for
authors to be able to sell their works directly (uploading over the internet)
and be able to ensure payment for their works.

Of course, books still need to be printed for those computer deficient people in
our society.  None the less, it would allow unknown authors to become known
authors.

____________________________________
PGP Key=3E634CB9
FP=6F4F3A5D52519646 3BD400A368158C8F

http://cust2.iamerica.net/kdl1
PGP key available at web page too!

This is a multi-part message in MIME format.

------=_NextPart_000_01BC3C10.E424EDE0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

I'd be interested, depending on how much C we would be using. I am about as
close to useless as you can get using C! Don't know it, don't want to.....

Brad
------=_NextPart_000_01BC3C10.E424EDE0
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<html><head></head><BODY bgcolor=3D"#FFFFFF"><p><font size=3D2 =
color=3D"#000000" face=3D"Arial"><br><br><br><br>Ed Parry &lt;<font =

wrote in article &lt;<font =

color=3D"#000000">&gt;...<br>&gt; Hi Guys!<br>&gt; <br>&gt; Anyone =
interested in a starting BASIC interpreter? I've been around a =
while,<br><br><br>I'd be interested, depending on how much C we would be =
using. I am about as close to useless as you can get using C! Don't know =
it, don't want to.....<br><br>Brad</p>
</font></font></font></font></font></body></html>
------=_NextPart_000_01BC3C10.E424EDE0--

This is a multi-part message in MIME format.

------=_NextPart_000_01BC3C12.1E76A0A0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

Brad
------=_NextPart_000_01BC3C12.1E76A0A0
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<html><head></head><BODY bgcolor=3D"#FFFFFF"><p><font size=3D2 =
color=3D"#000000" face=3D"Arial"><br><br><br><br>Gunnar Beushausen =

color=3D"#000000">&gt; wrote in article &lt;<font =

color=3D"#000000">&gt;...<br>&gt; That sounds really good. But how do =
you want to do that? In wich language<br>&gt; do you want to write =
it?<br><br>Well, not to {*filter*}in ..... &nbsp;&nbsp;<br>I'd like to write =
it using QuickBASIC for the first version and then use the resultant =
compiler to build upon itself. That way, I can start with something I =
know and build upon it. Seems to me the ability to add functions and =
subs is a great idea and one worth continuing, but it would be nice to =
be able to do the same thing and make it a permanent (or at least =
semi-permanent) part of the language.<br><br>&gt; <br>&gt; Have you ever =
written a compiler? This is really hard work, if you want to<br>&gt; =
have an example how someone shouldn't write an interpreter take a look =
at<br>&gt; the old ALPHA version of GBasic.<br>&gt; <br><br>I've never =
written one, but would like to. Have you ever written one? I'd sure like =
to have a look at the Alpha version of GBasic. Got it? <br><br>Brad</p>
</font></font></font></font></font></body></html>
------=_NextPart_000_01BC3C12.1E76A0A0--

The reason there's so many databases is the same reason there's
so many different languages. The end-programmer couldn't improve
upon it or was told not to improve on it because these
improvements would be NON-STANDARD.

-AND- before I get flamed again. I agree, MASM, C, Forth, etc.
allow the programmer to enhance, but...

(snip)

(snip)

P.S.
Will you be supplying the uniforms? Can we do something besides
black? How about sea-foam and salmon? Those are up-beat colors.

P.P.S
I don't look good in purple.

P.P.P.S.
I'm suprised a professonal would be interested in my cause, it
is NON-STANDARD.