by Clinton Parker
Atari users have a surprisingly wide selection of programming languages from which to choose. We've got three dialects of BASIC, four C compilers, eight or nine FORTHS, a pair of Pascals, PILOT, Logo, WSFN, a Lisp interpreter, numerous 6502 assemblers and a couple of hybrids like BASM and Mirth. Not bad for a "game machine," eh?
Leave it to Optimized Systems Software to come up with yet another way to tell your Atari what to do. OSS has been the leading purveyor of alternative operating systems and languages for the Atari since before I can remember. Action! is only the first of a whole new line of OSS products that's been causing quite a stir in the Atari underground. It's been touted as the first programming environment developed specifically for the 6502, and the fastest high-level language available for the Atari. These are pretty strong claims which, after playing with the system for several weeks, appear to be totally justified. As you are about to read.
In syntax and overall structure, Action! bears a strong resemblance to Pascal, C and other members of the Algol family. It's a procedure-oriented language featuring global and local variables, user-definable functions, parameter passing and powerful structures like DO loops, FOR-TO, WHILE, UNTIL and IF-THEN-ELSE. Three basic data types are recognized: 8-bit BYTEs (or CHARacters), 16-bit signed INTegers and 16-bit unsigned CARDinals. The system also supports a variety of extended data types including pointers, subscripted arrays, strings and records.
Listing 1 includes all of the keywords reserved for use by the Action! system. These are used to declare variables, define new procedures and/or functions and to control the operation of the compiler. BASIC veterans will note with alarm the total lack of keywords that do interesting things in and of themselves, like SETCOLOR or DRAWTO. They're missing for a very good reason. Unlike BASIC, Action! does not limit your programming to a limited number of safe little commands. It invites you (indeed, forces you) to invent the commands you need to solve problems yourself. The keywords in Listing 1 are the tools the system gives you to, in effect, write your own language. If this prospect doesn't excite you, maybe BASIC has been holding your hand for too long.
Don't get the impression that Action! leaves you completely on your own, though. The cartridge includes a library of useful I/O, graphics and system-level routines that you can use to start building more elaborate programs. Listings 2 and 3 will give you an idea of what's available. The resemblance of many Action! library words to Atari BASIC commands is intentional; the kindly folks at OSS want to make your transition from BASIC to Action! as painless as possible. This concern for familiarity unfortunately extends to the Action! graphics library, which offers exactly the same (limited) access to the hardware as Atari BASIC. Other weak points of the cartridge library include inadequate control over memory allocation and a mysterious lack of support for the Atari's built-in floating point math package.
Most of the elements in an Action! program are delimited by space characters - as many as you like! You don't have to keep track of line numbers, semicolons, brackets or any other nuisances that can make you feet more like a bookkeeper than a programmer. Just follow a few simple rules regarding commas and parentheses, and you're all set. Action!'s modern design encourages a wide-open style of program composition, with plenty of freedom regarding the use of blank lines, upper and lower-case characters, indentation, comments and other flourishes that improve readability and make coding more fun.
Internally, the Action! system consists of four distinct modules. There's an editor for creating and modifying program source text, a compiler which translates source text into executable machine code, a run-time library that supports the compiled code (described above), and a monitor which acts as a switchboard between the other three modules and (if you're using a disk drive) DOS.
A very important distinction between Action! and every other compiled language for the Atari is that these modules do not have to be loaded in separately from disk. All four are tucked away inside the SuperCartridge, safe from accidental erasure and ready whenever you need them. Further, the system is arranged so that your source text and compiled code can reside in memory at the same time. This self-contained design combines the performance of a compiled language with a degree of interactiveness usually associated with an interpreter. A stroll through the modules will show you what I mean.
Somebody at OSS once told me that the text editor in the Action! cartridge was originally going to be marketed by itself as a word processor. It isn't hard to believe. There are so many features and options in the Action! editor that I can only touch on the most interesting here.
Action!'s editor uses your TV as a virtual window into a text area that can extend well beyond the edges of the screen. Unlike the standard Atari screen editor, you can type up to 240 characters on a single line with no cursor wraparound. How? When your cursor reaches the right edge of the screen, the line you're working on (and only that line) starts to coarse-scroll to the left. You can keep right on typing until a buzzer informs you that you've reached the rightmost position in that line - the right "edge" of the text window. Move your cursor back towards the left, and the line scrolls to the right until you hit the left edge of the window. This design neatly eliminates the usual confusion between "logical" and "physical" lines of text.
Hitting CTRL/SHIFT/">" or "<" instantly moves you to the rightmost or leftmost character in the current line, respectively. You can also change the maximum width of the text window to any convenient value, such as how many characters will fit on your printer.
The Action! editor allows you to create a second text window, co-resident in memory but otherwise completely independent from the main window. The 2-window editing mode is represented visually by a split screen, with the bottom half of the image devoted to the auxiliary window. You can jump back and forth between the two windows and transfer blocks of text if desired; the editor remembers where you were working in each window and automatically returns you to that point when you return. Additionally, you can save, load or delete text in one window without disturbing the contents of the other. That means, for example, that you could load a library of routines into the auxiliary window, review them and copy the ones you need into your main program, which has been in full view the whole time! Sure beats LISTing and ENTERing lines of BASIC, doesn't it?
Other noteworthy capabilities of the Action! editor include global search and replace, instant access to the beginning or end of a file and the ability to delete, move and copy, selected blocks of text. The block move and copy functions are implemented so nicely that I have to tell you about them. When you hit the SHIFT/DELETE keys, the line you're working on disappears, just as with the Atari screen editor. But the line isn't gone forever. It's being held in a buffer, waiting to be moved or copied to anywhere else in your text window(s). Simply move the cursor to a likely spot and hit CTRL/SHIFT/"P" (for paste) to dump the contents of the buffer. Several adjacent lines of text can be sent to the buffer by repeatedly "deleting" them with SHIFT/DELETE. Action!'s method of picking up and dropping blocks of text feels very natural if you're used to the Atari screen editor, and it also eliminates the annoyance of losing a line of work by accidentally hitting SHIFT/DELETE. Incidentally, you can automatically undo any changes you have made to a line of text by hitting CTRL/SHIFT/"U".before leaving the fine. Luxurious.
Before you toss out your AtariWriter cartridge, let me point out a couple of small but irritating problems in the Action! editor. There's a feature called tagging which allows you to mark any location in your text by assigning it a unique one-character identifier. You can later return to that point in the text at any time by calling its ID code. It's a good idea that, unfortunately, isn't pulled off particularly well. If you set a tag in a line and change even a single character in that line, the tag disappears. This restriction (which is documented) considerably reduces the usefulness of the tagging option, to say the least.
My other gripe is with the way the cursor appears to flash and jump around the screen when it is being moved up or down, as if it isn't sure where to go next. The solid command line on the bottom of the screen also seems to jerk occasionally as you cursor around. Minor cosmetic points, perhaps, but an unstable cursor seems out of place in this otherwise superb little text editor.
After you've put the finishing touches on an Action! program and saved it out to disk, what next? Press the CTRL/SHIFT/"M" keys simultaneously and you'll find yourself staring at a barren white bar across the top of your screen. This is Action!'s monitor, the central interface between the editor, compiler, machine and user.
Monitor functions are invoked by typing a one-character code letter. You can select various compilation options, save and load compiled programs, examine the values of variables and memory locations and trace the execution of your programs. You can even use the X (execute) directive to interactively test almost any procedure or function. This capability is very unusual (and useful) in a compiled programming language.
Unlike Atari BASIC, which compiles each line of program text as it is typed, Action! requires that your program be explicitly translated into machine code before it can be executed. This isn't nearly as formidable as it sounds. All you have to do is type the letter C from within the Action! monitor.
The compiler accepts source text from either the editor (default), or from a text file saved onto cassette or disk. If you've been using both text windows, Action! will compile only the text in the window you last edited. Compilation is almost unbelievably rapid, especially when the source is the editor. I've never seen Action! take more than a few seconds to compile even a fairly large program that was in the editor. Small programs are compiled before you take your finger off the RETURN key. You can optionally instruct the compiler to list each line of source text to the screen or a printer as it is being compiled. This slows the compilation considerably, however.
A compile error causes the system to display the line where the error occurred, along with an error message number. Surprisingly for an OSS product, there are no English error messages. If you re-enter the editor after a compile error, you'll find the cursor obligingly positioned over the questionable spot in your text.
Successfully compiled code is executed by typing the letter R (run) from within the Action! monitor. If you're accustomed to the leisurely pace of Atari BASIC, get ready for a shock. OSS isn't kidding when they say Action! is fast.
Execution speed is very important to Atari programmers. Why? Because much of the software written for the Atari relies heavily on graphics, where a few extra machine cycles in the wrong place can make the difference. between a spectacular special effect and an interesting but unmarketable demo. High speed isn't likely to hurt a non-gaphics program, either. This is in accordance with Moriarty's Maxim: It is much easier to slow down a computer program than it is to speed it up.
A number of attempts have been made to devise a universal method for comparing the speed performance of computer languages and hardware. In September of 1981, Byte magazine published an iterative number-crunching algorithm called the Sieve of Eratosthenes, which calculates all of the 1,899 prime numbers between 3 and 16,384.* The Sieve has since become the informal industry standard for clocking the speed of microcomputer languages.
*Jim Gilbreath, "A High-Level Language Benchmark." Byte, VI, 9 (September 1981), pp. 180-198.
Listing 4 is an implementation of the Sieve in Atari BASIC. It requires 19,490 jiffies or approximately 5 1/2 minutes to execute on an unmodified 48K Atari 800 system. I recognize that Listing 4 is not the most efficient way to write the Sieve in Atari BASIC, but it is the clearest and most portable way, and that's what counts in this application. You might like to try rewriting the Sieve for better speed performance. I've achieved improvements of better than 30% with tricky recoding.
ACTION4.LST is available in ATASCII format.
10 DATA 941,347,921,5,772,308,90,393,3 14,892,919,849,588,596,860,8795 25 DATA 623,689,395,581,796,17,419,352 0
Although I love standards, I don't like the Sieve. It's not easy for beginners to understand, it takes too long (in BASIC, anyway), and it doesn't test the Atari under real-world conditions, with lots of 6502 processor time being "stolen" by Antic for video DMA. I wanted a benchmark that anybody could appreciate, operating under the kind of DMA conditions an Atari program is likely to find itself up against.
Back in Issue 11, I devised a little program that fills a GRAPHICS 24 screen with color, one byte (eight pixels) at a time. It was used to compare a couple of BASIC compilers at the time, but it's equally valid in any run-time environment. My definitive BASIC implementation of this test appears in Listing 5. Screen Fill, as the program shall henceforth be known, executes in 4025 jiffies or about 67 seconds on a 48K 800. (Again, improvements are possible, but for the sake of clarity let's stick to Listing 5.) I'll be using Screen Fill in conjunction with the Sieve to judge the performance of every new language I review from now on. So let it be written; so let it be done.
ACTION5.LST is available in ATASCII format.
10 REM * SCREEN-FILL BENCHMARK 11 GRAPHICS 24 12 POKE 19,0:POKE 20,0 13 SCREEN=PEEK(88)+256*PEEK(89) 14 FOR I=0 TO 31 15 FOR J=0 TO 239 16 POKE SCREEN+J,255 17 NEXT J 18 SCREEN=SCREEN+240 19 NEXT I 20 TIME=PEEK(20)+256*PEEK(19) 21 GRAPHICS 0 22 PRINT TIME;" JIFFIES"
10 DATA 206,5,2,185,233,103,695,394,78 6,399,557,157,527,4249
OSS includes a implementation of the Sieve benchmark in their Action! documentation. I rewrote the code slightly to make it match my BASIC implementation more closely; the modified program is shown in Listing 6. It executes in 89 jiffies or just under a second and a half. I'll save you a calculation by pointing out that the Sieve runs about 219 times faster in Action! than it does in Atari BASIC.
ACTION6.ACT is available in ATASCII format.
BYTE RTCLOK=20, ; addr of sys timer SDMCTL=559 ; DMA control BYTE ARRAY FLAGS(8190) CARD COUNT,I,K,PRIME,TIME PROC SIEVE() SDMCTL=0 ; shut off Antic RTCLOK=0 ; only one timer needed COUNT=0 ; init count FOR I=0 TO 8190 ; and flags DO FLAGS(I)='T OD FOR I=0 TO 8190 DO IF FLAGS(I)='T THEN PRIME=I+I+3 K=I+PRIME WHILE K<=8190 DO FLAGS(K)='F K==+PRIME OD COUNT==+1 FI OD TIME=RTCLOK ; get timer reading SDMCTL=34 ; restore screen PRINTF("%E %U PRIMES IN",COUNT) PRINTF("%E %U JIFFIES",TIME) RETURN
Unconvinced? Listing 7 is an Action! implementation of Screen-Fill. This demanding little gem executes in 32 jiffies (slightly more than half a second), or 126 times faster than its BASIC counterpart under maximum DMA handicap. And if you cheat by replacing ing the nested FOR-TO loops with an Action!, SETBLOCK procedure in the form:
you'll obtain an execution time of just five jiffies. This is essentially the same amount of time it takes the equivalent machine-language code to do the same job. No other high-level Atari language that I am aware of can match this kind of speed performance.
ACTION7.ACT is available in ATASCII format.
BYTE RTCLOK=20, ; addr of sys timer SAVMSCL=88, ; lsb of screen addr SAVMSCH=89, ; msb I,J,TIME ; declare variables CARD SCREEN PROC BENCH() GRAPHICS(24) RTCLOCK=0 SCREEN=SAVMSCL+256*SAVMSCH FOR I=0 TO 31 DO FOR J=0 TO 239 DO POKE(SCREEN+J,255) OD SCREEN==+240 OD TIME=RTCLOK GRAPHICS(0) PRINTF("%E %U JIFFIES",TIME) RETURN
Once I got a taste of Action!'s dizzying speed, I had to find out what was going on inside that demonic little cartridge. So I used the W (write object code) option of the Action! monitor to send a copy of the compiled Screen-Fill benchmark to a disk file. Then I read it back into Ralph Jones' Ultra Disassembler (published by Adventure International), massaged the labels and commented the code to make it correspond to the Action! source text, line by line. The result appears in Listing 8.
Assembly programmers will appreciate the extraordinary efficiency of the Action! compiler. The code in Listing 8 is totally non-recursive. It uses no special stacks or indirect pointers to control the flow of execution, just pure in-line machine code with an occasional JSR into a cartridge library routine. This is "native mode" compilation at its best: simple, clean, and very, very swift. The output of a typical C or Pascal compiler looks like spaghetti by comparison.
Because compiled Action! programs refer to subroutines that reside inside the Action! cartridge, you can't run a program without the cartridge in place. This may come as a disappointment to users who want to give copies of their latest Action! game to friends who don't have Action! OSS plans to remedy this situation by offering a Personal Run-Time Package to licensed Action! users for around $30. It's a utility that will let you turn any Action! program into a self-standing entity that will run with no help at all from the Action! cartridge, thank you. A commercial run-time package will also be offered for a one-time licensing fee of approximately $300. Both may be available by the time you read this; contact OSS directly for more information.
Another $30 will get you OSS's programmer's Aid Disk (PAD), a collection of demonstration programs and library routines that wouldn't fit into the already crowded Action! cartridge. The libraries include badly- needed support for player/missile graphics, memory management and floating point math, precisely the weaknesses I noted above. The demo programs are very instructive and help to clarify some of the obscure features of the language. You even get a full-blown game program, written in Action! by our very own Joel Gluck.
The PAD squarely addresses many of the shortcomings of the Action! cartridge and documentation, and is an absolute must for all serious owners of the Action! system. In fact, this material ought to be included with every new system sold, even if it means bumping up the price a bit.
The 16K Action! "SuperCartridge" is a technically interesting device in and of itself. It employs a hardware technique called bank-selecting to make itself "look" like an 8K cartridge. This gives you access to the 8K of RAM between $8000-$9FFF that is de-selected and thus rendered useless by a conventional 16K cartridge, such as AtariWriter.
The bottom half of the SuperCartridge ($AOOO- $AFFF) is divided into three independently addressable 4K banks of ROM, which are automatically switched in and out depending on what part of the system is in use. If your Atari has 48K or more memory, it's even possible to address the 4K bank of RAM that resides "under" this half of the cartridge. OSS's new DOS XL operating system takes advantage of this capacity in a most ingenious manner. Look for a report in a future issue.
The bank-select cartridge is a nearly ideal home for Atari software. It gives the cartridge designer a full 16K to work with, enough room for plenty of bells and whistles. It gives the user an instant-loading, highly reliable environment with up to 40K of workspace. And because three of the memory banks occupy the same 4K address range, a bank-select cartridge is very difficult to pirate. Let's hope that more manufacturers start taking advantage of bank-selecting to enhance the value and security of their products.
I'm sorry to report that the Action! Reference Manual doesn't do the language justice. In a commendable attempt to satisfy beginners and experts alike, the Manual suffers from lack of confidence, uncertain organization and a shortage of good, hard technical data. Thank goodness for the numerous sample programs, which communicate a lot more about the system than the text surrounding them.
Having once written the manual for a new (and mercifully obscure) programming language, I can appreciate the difficulties involved in deciding how much needs to be said, to whom, and in what order. Nevertheless, a new language can only be as good as its documentation. Until somebody sits down, rolls up his or her sleeves and writes an authoritative book about Action!, it will have a hard time attaining the wide acceptance it so obviously deserves. I conclude this diatribe by acknowledging that the latest edition of the Reference Manual (in the small yellow notebook) shows a marked improvement over the first release.
The Action! cartridge itself has gone through a couple of changes since its first appearance in August 1983. You can tell which version you have by using the "?" (display memory) command in the monitor to examine cartridge address $BOOO. If this byte equals $31 hex, you have the original Version 3.1. A value of $33 indicates Version 3.3, in which a number of minor 3.1 bugs have been corrected. The final version is 3.6 ($36 at $BOOO), which should be ready soon after you read this. OSS has always been very good about maintaining their products, so you shouldn't have any trouble getting an upgrade if you need one. Consult OSS for prices and availability.
I hope my kvetching about the documentation doesn't scare you away. if sensible, structured code and edge-of-the-art speed are what you crave in a high-level language, Action! is exactly what you need. OSS's hideous orange cartridge joins the ranks of valFORTH, Omnimon!, ABC and MAC/65 as one of the most valuable development tools ever published for the Atari. Congratulations and thanks to Clint Parker and OSS for bringing us such an advanced product. You can expect to see plenty of support for this exciting new language in future issues of ANALOG.
ACTION8.ASM is available in ATASCII format.
0100 ; DISASSEMBLY OF COMPILED 0110 ; ACTION! SCREEN-FILL 0120 ; BENCHMARK (LISTING 7) 0130 ; ----------------------- 0140 ; 0150 ; DEFINE ADDRESS CONSTANTS 0160 ; ------------------------ 0170 RTCLOK = 20 0180 SAVMSCL = 88 0190 SAVMSCH = 89 0200 ; 0210 ; GLOBAL VARIABLE STORAGE 0220 ; ----------------------- 0230 *= ORIGIN 0240 I *= *+1 ; reserve 1 byte for 0250 J *= *+1 ; each BYTE variable, 0260 TIME *= *+1 0270 SCREEN *= *+2 ; 2 bytes for CARDs 0280 ; 0290 ; PROC BENCH() 0300 ; ------------ 0310 JMP START 0320 ; 0330 ; If our procedure used local variables, 0340 ; they would have been stored here. 0350 ; That's why the above JMP is included. 0360 ; 0370 ; GRAPHICS(24) 0380 ; ------------ 0390 START 0400 LDA #24 0410 JSR GRAPHICS 0420 ; 0430 ; RTCLOK=0 0440 ; -------- 0450 LDY #0 0460 STY RTCLOK 0470 ; 0480 ; SCREEN=SAVMSCL+256*SAVMSCH 0490 ; -------------------------- 0500 LDA #0 0510 STA TEMP1+1 0520 LDA SAVMSCH ; move SAVMSCH into 0530 STA TEMP1 ; TEMP1 0540 LDA # >256 ; msb of multiplier 0550 TAX 0560 LDA # <256 ; lsb 0570 JSR MULTIPLY 0580 STA TEMP4 0590 TXA ; save (256*SAVMSCH) 0600 STA TEMP4+1 ; into TEMP4 0610 ; 0620 CLC 0630 LDA SAVMSCL ; add SAVMSCL to 0640 ADC TEMP4 ; (256*SAVMSCH) and 0650 STA SCREEN ; store in SCREEN 0660 LDA #0 0670 ADC TEMP4+1 0680 STA SCREEN+1 0690 ; 0700 ; FOR I=0 TO 31 DO 0710 ; ---------------- 0720 LDY #0 0730 STY I ; init I-loop 0740 ILOOP 0750 LDA #31 0760 CMP I ; reached limit yet? 0770 BCS JINIT ; no - do another J-loop 0780 JMP GETIME ; else get timing 0790 ; 0800 ; FOR J=0 TO 239 DO 0810 ; ----------------- 0820 JINIT 0830 LDY #0 0840 STY J ; init J-loop 0850 JLOOP 0860 LDA #239 0870 CMP J ; reached limit yet? 0880 BCS DOPOKE ; no - poke another byte 0890 JMP ADD240 ; else update screen 0900 ; 0910 ; POKE(SCREEN+J,255) 0920 ; ------------------ 0930 DOPOKE 0940 CLC 0950 LDA SCREEN ; add SCREEN to 0960 ADC J ; J, and 0970 STA TEMP2 ; save in TEMP2 0980 LDA SCREEN+1 0990 ADC #0 1000 STA TEMP2+1 1010 ; 1020 LDY #255 1030 LDX TEMP2+1 1040 LDA TEMP2 ; poke (SCREEN+J) with 1050 JSR POKE ; a 255 1060 ; 1070 ; OD (for J loop) 1080 ; --------------- 1090 INC J 1100 JMP JLOOP 1110 ; 1120 ; SCREEN==+240 1130 ; ------------ 1140 ADD240 1150 CLC 1160 LDA SCREEN ; add SCREEN and 1170 ADC #240 ; 240; store result 1180 STA SCREEN ; in SCREEN 1190 LDA SCREEN+1 1200 ADC #0 1210 STA SCREEN+1 1220 ; 1230 ; OD (for I loop) 1240 ; --------------- 1250 INC I 1260 JMP ILOOP 1270 ; 1280 ; TIME=RTCLOK 1290 ; ----------- 1300 GETIME 1310 LDA RTCLOK 1320 STA TIME 1330 ; 1340 ; GRAPHICS(0) 1350 ; ----------- 1360 LDA #0 1370 JSR GRAPHICS 1380 ; 1390 ; PRINTF("%E %U JIFFIES",TIME) 1400 ; ---------------------------- 1410 JMP OVER ; skip over in-line string 1420 STRING 1430 .BYTE 13 ; length of string 1440 .BYTE "%E %U JIFFIES" 1450 OVER 1460 LDA #0 ; msb of TIME 1470 STA TEMP3 ; into TEMP3 1480 LDY TIME ; lsb into Y 1490 LDX # >STRING ; msb of string addr 1500 LDA # <STRING ; lsb 1510 JSR PRINTF 1520 ; 1530 ; RETURN 1540 ; ------ 1550 RTS ; from procedure 1560 ; 1570 RTS ; back to Action! monitor