Open Apple Vol 1 No 00 JAN 1985

Releasing the power to everyone. January 1985 Vol. 1 No.0 Uncle DOS survives Softalk bankruptcy Matthew Monitor and Dr. Basic also found alive Uncle DOS, the beloved leading character of Softalk magazine's monthly DOStalk column, is in good condition after a close brush with death. He will appear monthly in Open-Apple, a new newsletter for Apple users, according to ‘Tom Weishaar, former author of DOStalk. Matthew Monitor and Dr. Basic, as well as many other mythical figures, have also signed to appear in the new newsletter, which Weishaar will publish himself. Uncle DOS, Matthew, and the good doctor were found buried in Softalk rubble after an extensive search by Weishaar. Softalk entered Chapter 7 bankruptcy in August. When asked what happened, Uncle DOS reported, "We don’t know—the roof just caved in on us.” Weishaar, developer of the Beagle Bros programs ProntoDOS and Frame-Up, had written Softalk’s DOStalk on a free-lance basis. He had no other connection with the magazine and could not explain the bankruptcy either. ™ Softalkwas the most popular and highly-regarded of the Apple-only magazines’ Weishaar said. “Its rapid growth caused problems that its managers were unable to deal with. The magazine's readers were loyal — many still are. One reader told me that if he had known the magazine was in trouble he would have taken up a collection’’ Weishaar said he decided to create Open-Apple to fill one of the information voids left by Softalk’s demise. “There are a large number of Apple-only publications suitable for beginners; Weishaar said. “There are also magazines available for full-time professionals. What's missing is something for intermediate Apple users— people who want to release the full power of the Apple Il in their own area of expertise. Softalk used to Matthew Monitor, Dr. Basic, and Uncle DOS as seen at the “Applelll Forever” brouhaha held in San Francisco last spring. provide this information in columns such as DOStalk and IInd Grade Chats, but few other publications seem willing—or able—to fulfill the needs of the inter- mediate user.” Open-Apple will be published monthly, Weishaar said, with the first issue scheduled for February. The subscription price will be $24 a year. Potential readers are encouraged to encourage Weishaar by signing up as charter subscribers. If you would like more information about Open-Apple, just keep reading. You are holding asample issue in your hands. Complete ordering information is on the back page, or if you are really lucky, you got an order card with this sample letter. Please fill it out now, even if you don't intend to subscribe. Your feedback on this sample issue will be carefully scrutinized and heeded, My Two Bits Tom Weshaar My banker says I'm crazy to start a news- letter about an eight-year old computer. He says eight years is an eternity in computer- time. He says the Apple II should be dead and forgotten by now like the other machines of its generation —the KIM, the Sol, the TRS-80 Model |, the Compucolor, and the Video Brain. For most of the last eight years even the people at Apple Computer, Inc. have privately called the Apple Il a “dead machine”. They have consistently underestimated—often by wide margins—how many Apple lis they could sell (while overestimating the market for the Apple Ill, Lisa, and Macintosh). The high priests of computer science think the machine isajoke. Meanwhile, buyers have managed to liberate more than two million of them from Apple's factories and several hundred thousand compatible machines from other places. In his book Enhancing Your Apple I, Volume 1, Don Lancaster (who because of his many articles and books about electronics is called “the father of the personal computer” by InfoWorld columnist John Dvorak and others) says: The Apple Il is far and away the single most powerful tool ever put in the hands of many individuals on an uncontrolled and unregulated basis. The new personal freedoms and the potential opportunities that result from this are almost beyond belief. I'snowa whole new ball game, ajump into hyperspace from where we are. The Apple Il is far more significant and it will have a vastly greater impact than such short term frivolities as the automobile and television—and, possibly even more than the printed word itself. Like Lancaster, | believe in the Apple ll. This is the computer of the student, ofthe entrepreneur, of the researcher, of the cottage industrialist, of the people. The combination of Apple's fatalistic attitude about the II and the machine's massive popularity have caused a slow evolution. Most computer companies periodically trash everything their customers know about computers through radical redesigns. Even Apple has tried this several times, but because we keep 2 Open-Apple buying the Apple I! instead of more “advanced” computers, our knowledge base has been protected. We have benefitted from a slow progression of enhancements to Steve Wozniak’s dream machine. Much of the software written for Apples in the late 1970s runs justfine on Apples manufactured today. Today's Apples can do more, yetnone of the hours early users spent learning about their machines has been wasted. Today's Apples can still do what yesterdays Apples did. This is very unusual in a field as dynamic as computers. The limits of the original Apple II are now in sight. The full potential of the Ile, however, isn’t known. And we know even less about how to hamess the lic. On the horizon is a 16-bit Apple II, currently known as the IIx. This machine will have the memory and processor capacity to run large, powerful programs, such as the successors to Lotus 1-2-3. But the real beauty of the machineis that itwill also beable torun standard Apple 65028-bit programs, such as Applesoftand DOS 3.5. On the one hand, it will run any Apple I program; on the other, it will have horsepower today's Applell users have never dreamed of. Best ofall, itwill validate your experience with and knowledge about todays Apple Il. This newsletter is dedicated to the Apple Il. It will follow the advances Apple makes to this family and it will follow the advances users around the world make in pushing these machines to their limits. ! want Open-Apple to be the world-wide gathering place for Apple II people. If machines are to be pushed to their limits, it's critical that users exchange information. No one can discover everything alone. Open-Apple will be a Clearinghouse for elegant tips, tricks, and procedures. The information published in Open-Apple will come from several sources. As a software developer, | am able to obtain some material directly from Apple that is generally unavailable to the public. I subscribe to every Apple publication I knowabout; some of the things you'll see here are refinements of the best ideas from other publications. One of the most fruitful ways of learning is to take apart high-quality programs and see how they work. Much of my time is spent searching for beauty in other people's software. You'll read about my discoveries here. Finally, many good ideas come in over the transom—in conversations and correspondence with other users. You may know several things about the Apple most of us have yet to discover. Part of the fun of Open-Apple is your ability to Participate in what's published here. The way to convince my banker I'm not crazy is to show him a stack of order cards. For that I need your help. Ifyou believe in the Apple IL fill out the coupon on the back page of this newsletter and send it in. We have years of adventures ahead ofus. Digging Into DOS Reading a text file with an Applesoft program has always seemed like a pretty straightforward affair. You open the file, tell DOS youwant to read from it, and input it Ifyou know how many lines are in the file, you issue that many input statements, then stop. Often, however, the number of lines in the file isn’t known. In that situation the typical procedure is to read the file until DOS sends an end of data message and then stop. Onerr goto errors. To accomplish this you issue an onerr goto statement. When an error occurs, program control will go to your error-handling routine. There. you check to see if the error was indeed end of data (error number five). Ifnot, you will want to take some true error-handling action. But when the error is simply your signal that the end of the file has been reached, you want to continue processing. At this point most programmers send control back to their main program. This works for awhile, but if you read enough files in sequence your program will die quite unexpectedly. For example, try this little program. Ittries to read an empty file called emptyover and over. It reports its efforts as it proceeds. After it tries the 18th time it crashes into the Monitor—fairly unusual for an Applesoft program this simple. 100 ne 120 FS="ENP TY" DS=CHRE(4) 200 PRINT “PROCEEDING MITH READ Vol. 1, No.0 [ ce ‘e ready protocol in the while the Macintosh uses xon/xoff protocol (switch 2-5 closed). : handy, isnt it? : cae ee : My Imagewriter now has.all the dips on switch one, exc ‘num ‘open. Allare closed on i Gos weet cnmeestr te Mataah use the llc | have to tum off the printer, open switch 2-5, switch cables, and tum the printer back on. A switch box would solve the two-cable problem, | pe bee es on a ee Anybody want to buy a facintosh? ieee : 500 L=1 : ONERR GOTO SSO 510 PRINT D$;“OPEN";FS 520 PRINT 08;"READ"3FS 53 INPUT TS(L) : L=L+1 : GOTO 530 550 PRINT : PRINT D8; “CLOSE";FS S60 IF PEEK (222) = 5 THEN GOTO 200 570 PRINT “ERROR 8”; PEEK (222); IN LINE “: PEEK (218) + PEEK (219) * 256 580 END After the program crashes, return to Applesoft with a 3D0Gand issue a goto 570 command on your keyboard. Our program will report that we were killed by error number 77, out of memory, in line 550. That may seem pretty incredible to you, given the small size of this program. But people often forget there's more than one kind of memory to run out of. The usual way to run out of memory is to have a program so large and with so many variables that it won't al fit in your computer. For example, just type on your keyboard dim a(1000,1000). An array of that size won't fit and you'll see out of memory error appear on your screen. All stacked up. The other kind of memory you can run out of is called stack memory. Applesoft saves all kinds of stuffon the stack, which is a 256-byte area that the 6502 microprocessor can access quite easily. Whenever you do a gosub, for example, Applesoft pushes the current line number on the stack before passing control to the subroutine. That way, when the subroutine’s retum is executed, Applesoft can find its way home by pulling the line number off the stack and jumping to it Modify line 500 of our program to 500 gosub 200and run itagain. This will cause awhole bunch of gosubs to be executed without any returns. You'll again see the Pout of memory error message; this time during the 26th pass. But unlike when you typed dim a(1000,1000), you've now run out of stack memory rather than program memory. Just as gosub pushes a line number on the stack, so does the onerr goto command. And the resume statement is onerr goto’s return. The reason our first program crashed was that we never used resume. However, the solution isn’t as easy as just beginning to use it Resume sends control back to the line in which the error occurred. If we tried to resume after an end of dataerror, we would just getanother one and our program would lock upina tight cycle of onerr gotos and resumes. ‘The solution to all this is to use a call to a machine language subroutine within Applesoft that will clear the stack. Use this call in place of resume. The call is mentioned in Apple's Applesoft BASIC Programmer's Reference Manual, Volume 2 on page 265. The subroutine lies at $F528 and can be executed with a call -3288. Earlier Apple manuals included a short machine language program that could be poked into memory and called that would also clear the stack. Call -5288 does the same thing, and is obviously much easier to execute. Using this call, you can put the portion of your program that reads files into a subroutine. If you try this without the call, you get a retu without gosub error when you try to return to your main program after the first end of data error. So here's the standard way to read a text file from Applesoft. This program will run forever or until you press Control-C. 100 DS=CHRS(4) 110 Nel 120 FS="ENPTY” January 1985 200 210 220 PRINT “PROCEEDING WITH READ “3 GOSUB See GoTD 200 1: NENAL 500 L=1 : ONERR GOTO 55@ PRINT 08; “OPEN” 3S 520 PRINT 08; "READ" ZFS 530 INPUT TS(L} : L=L+1 : GOTO 530 550 560 PRINT : PRINT DS;“CLOSE"3FS IF PEEK (222) = 5 THEN CALL -3268 : RETURN 570 580 PRINT “ERROR f°; PEEK (222); END IN LINE "; PEEK (218) + PEEK (219) * 256 Seen In Print Steve Wozniak told InfoWorld to look for the 16-Bit Apple Ix in 1986. If you missed it, InfoWorld carried a report ofan interview with Wozniak in its November 19 issue. Wozniak said the machine would have a built-in drive, slots, expandable memory, and the 16-bit65816 microprocessor. This chip can directly address up to 16 megabytes of memory. Of course it can also emulate a 6502 and thus run today's Apple II software. In the October issue of Apple Assembly Line Bob Sander-Cederlof has a report on the newly available 65802 microprocessor. This chip can do intemal calculations with 16 bits but handles data eight bits at a time like the 6502. It is completely compatible with the 6502 — Bob installed his in a Ile and everything worked normally. Right now the chip costs about $100. The phone number for Apple Assembly Line is 214-324-2050. I've always said ProDOS would become the Apple Il operating system of choice when the price of hard disk drives arp to $500. We're still not there yet, but we're getting closer. A company named First Class Peripherals announced anew, $700, 10-megabyte hard disk for the Ile in the November issue of most Apple magazines. The specifications look pretty good, including DOS 3.3 and ProDOS compatibility. We'll be watching this one. Their phone number is 800-538-1307. ‘The Apple PugetSound Program Library Exchange — the big Apple user group in the Seattle area that publishes Call -A.PPL.E.—recently changed the legal form of its organization to a cooperative and is now known as A.P.P.L.E. Co-op. They've always had really good mail-order prices on hardware and software for Apples; as a co-op they may sell to non-members. Members will get patronage dividends at the end of the year. For people who don’t need the hand-holding local dealers supposedly offer, APP.L.E. is a good place to buy stuff. Their phone number is 206-872-2245, Go, Logo, Go After releasing the IIc, Apple came out with anew, 128K, ProDOS-based version of Logo called Apple Logo II. It comes with some good documentation —a tutorial and a reference manual— and a copy-protected master disk with the usual “as is” warranty. The package costs $100 retail. It comes in a lclike red and yellow box; don't let somebody sell you the older 48K Logo if you want this one. You might wonder why anyone would buy Logo when Applesoft BASICis included “free” with all Apples. Let me duck the superior/inferior language arguments for a moment and simply point out that the equivalent amount of documentation for BASIC—the Applesoft Tutorial, Applesoft Programmers Kit, and BASIC with ProDOS costs $110. Thus, if your goal is simply to lea a computer language, you can’t use price as a deciding factor when comparing Logo and Applesoft. Both languages can be learned from Apple's materials. They are “easy” in the sense that neither requires formal classroom training, They are also “easy’ in the sense that both are intrepreted languages. You can type commands in on the keyboard and get immediate execution. You can write simple programs within five minutes of cracking open either tutorial. With other languages (take Pascal, please) you have to know half the language before you can add two numbers together. Logo has a built-in full-screen editor for writing programs. This is nice— especially compared to what Applesoft provides. The editor isn’t quite AppleWriter, butit’svery easy to use. It can also create or edit any kind of ProDOS textfile shorter than 6,145 characters. Many people think Logo is a graphics language. While graphics programs are very easy to write with Logo, the language goes far beyond pictures. It is derived from LISP, a language developed by artifical intelligence researchers. BASIC is a better language for dealing with numbers. Logo shines at handling lists of words, that is, data bases. Before yourun out and decide to write the Great American Data Base Program in Logo, however, you should know that there's one critical problem. Apple Logo II is slow. Not only do programs execute slowly, they are loaded by “execing” them into memory. | haven't experienced such long waits since | disconnected my cassette recorder and fedit to the goats. One ofthe sample programs Apple provides takes well over two and a half minutes to load. Nonetheless, Logo has its advantages. It's a nicely structured language. You program by writing procedures, which are kind of like subroutines. The name you Open-Apple 3 give a procedure becomes a new Logo command. Thus you create your own language as you use it. The main thing that attracts me to Logo, however, isa book called Mindstorms by Seymour Papert. Papertisa child psychologist and educator. He and his associates developed the language. In Mindstorms Papert tells why. In Paperts vision, you don’t learn Logo because you want to know about or program computers. Instead you learn Logo because you want to know, period. Papert presents Logo as a language for learning, Programming in Logo provides intellectual models, Papert says, that are otherwise rarely available. Using the computer, these models can be manipulated by people and readily absorbed. Learning Logo is supposed to be an adventre in learning, not an adventure in leaming about computers. In this column we're going to present tips and tricks for using Logo. They'll be based on the same kind of peeks, pokes, and calls our other stuff revolves around. Precious little work that I can find has been published in this area, so a lot of this column will be breaking new ground. Any input you might add to this procedure would be appreciated. Reviewer's Corner 3 se Fontrix. By Data Transforms, Inc. This program has received several favorable reviews and has some impressive advertising. The Apple version of the program has some severe limitations, however. Fontrix bills itself as “extended screen graphics software.’ This is a good description ofits mostimpressive feature — the ability to create and print graphics that are much larger than a single hi-res screen, These graphics are stored in a special file Fontrixcalls a graffile. Graffile size is defined in sectors and can range in size from a single hi-res screen, which is five sectors wide and six sectors high, up to 480 sectors in any combination of width of height that's larger than a hi-res screen. You can create and printa graphic the size ofa single sheet of paper, for example. The exact size depends on your printer; the Fontrix manual includes a table showing the possibilites. Most Epson printers can handle graffiles 17 sectors wide by 26 sectors high. Most other printers, including Apple's, can handle graffiles 11 by 24. The program has three basic modules, a graphic writer, a graphic printer, and a font editor. The Fontrixdisk comes with 11 fonts. You can create more yourself with the font editor, or you can buy additional Fontpack character set disks from Data Transforms for $20 each. The maximum character cell size is 32 by 32 pixels. Normal Apple text is 5 by 7. Characters can be either all the same width or proportionally spaced. The only input device recognized by the font editor is the keyboard. The graphic printer is a limited program compared to many of the graphic printers available today. Itallows inverse or normal printing on 8.5 or 14 inch paper widths. A magnification option accepts choices from 1 to 255, but choices over 3 result in some of the graphic being cut off on the right hand side. The only other options available relate to where on the paper a graphic will be printed—rightside, center, and so on. The graphicwriter is the module you use to actually create displays. Itallows you to type fonts on the screen. Some primitive graphics abilities are also available, including background colors and pattems, but Fontrix is clearly not a drawing program. Graphic input devices can be used, but only to position the cursor. This can be handy when non-character fonts, such as electrical, mathematical, or engineeririg symbols, are being placed on the screen, but is of little use with straight text. Inadvertisments you will see stunning graphics created with Fontrix These were no doubt created on the IBM version of the program, which has more advanced drawing capabilities. Fontrixis presented in advertising as a typesetting program. The Apple version is virtually useless in this application, however, because you can't send text files you've created on a word processor to the graphic writer. This is possible with the IBM version, but not with the Apple version. In addition, the program has no provision for justifying text on both margins. You should also know that the examples of Fontrix printouts included in advertising and the manuals have been photographically reduced. The manual makes this clear and recommends a reduction of about 40 per cent for best results. The day will come when software will be available that will tum an Apple II and upcoming laser printers into typesetting machines. However, we aren't there yet. Fontrix 1.2, by Data Transforms, Inc. (616 Washington St, Suite 106, Denver, Colorado, 80203; 303-832-1501). Unprotected. “As is" warranty. $75, 4 Open-Apple Picking Up Applesoft Taking a poke at the Garbageman Everybody has anemesis. Hamlet has his step-father, Batman and Robin have the Riddler, Peter Rabbit has Mr. McGregor. Applesoft programmers working under DOS 3.3 have the Garbageman, When the Garbageman strikes, your Applesoft program will come to a sudden and complete halt. Your keyboard will go dead. Your face will look confused and angry. You will probably hit control-reset, unless you've met the Garbageman before. In that case you will sit and wait. After a pause of something between a few seconds and a few minutes, the Garbageman, his can full of electronic detritus, suddenly flees. Your Apple will mysteriously start working perfectly again. What strings. Consider the following famous program, which originally appeared in Softalk’s DOStalk a couple of years ago: 10 INPUT “kHAT? ~;AS: 20 GOTO 10 When you run this program, the word "WHAT?" appears on your screen with a cursor beside it. You type WITCHES” and press return. Applesoft puts the word “WITCHES” in an empty place inside your computer. It also stores the length and location of “WITCHES” in a table used to keep track of such stuff. Its called a variable table, The empty place where strings are stored extends from where the variable tables end to where DOS starts. The first string is placed right up against DOS. The next one is tucked in next to the first. Figure 1 will help you figure out where this string storage place is inside your computer. The left-most bar of Figure 1 represents the entire range of Apple II memory and shows its standard uses. The middle bar shows how Applesoft splits up the “free” memory area from byte $800 (in decimal that's 2048) to byte $9600 (38400). Applesoft uses this area for storing your actual program and your program's variable tables and strings. The right-most bar of Figure 1 gives you a peek inside the string storage area itself. Usually memory maps are presented with the high addresses at the top. Some of you may think the Open-Apple map is upside down. We do it that way because it matches what you see when you examine memory using the Monitor. Ittakes onlya few minutes of trying to match what you see in the Monitor's memory display with whatyou seein a standard memory map to realize thatit’s the standard maps that are upside down. Since the first string is stored right next to DOS and the second next to the first, the order in which they appear in memory is backward from the order in which they are used. Note in the right-most bar of Figure 1 that our first string, “WITCHES”, appears as the last word in the bar. Ifline 20 of our famous little program said print a8, Applesoft would look in the variable table until it found the entry for a$, then would use the length and address stored there to retrieve "WITCHES". Why the Garbageman takes solong. Ah, but the next commandisn't print, it's goto 10. Our input statement is repeated. In response to the second “WHAT?”, you typein “GOBLINS”. Applesoft now stores "GOBLINS" in unused memory, right next to “WITCHES”, and changes the length and address for a$in the varlabie table. Applesoft keeps track of where the edge between empty memory and the string storage area is by means ofa pointer keptat bytes $6F-70 (111-112). This pointer is called fretop in the Applesoft literature. It, too, is shown in Figure 1, along with several other interesting pointers. Whenever you enter a new string, Applesoft refers to fretop to figure out where to store the string. Heer {ceca cats ihe prs the return to the screen ae a ‘have to destroy what's already there with another pré3 _Wheneeryoutum onthe tp abet hcan Ups ileal Program later tums on a printer with a pr#1? Vol. 1, No.0 Since your program has changed the value of a$from “WITCHES” to “GOBLINS”, Applesoft correctly assumes you are no longer interested in “WITCHES”. The word “WITCHES” is still in memory, but because its length and address have been overwritten by the length and address of “GOBLINS”, Applesoft can no longer find it. “WITCHES” is now an inactive string, “WHAT?” You type in “BROOMSTICK” and a couple thousand similar words. Eventually Applesoft will go to store a word you have typed in and find the strings have bumped into the variable tables — there is no more empty memory left. At that point Applesoft calls on the Garbageman to examine the variable table, find all the “active” strings, and consolidate them down at the DOS end of memory. After the consolidation, fretop will be changed to point to the new boundary between the active strings and the rest of memory. All the memory held by “inactive” strings will be freed and made available for reuse. The Garbageman has a lot of work to do. He can’t just go in, find the first active string in the variable table, and move it next to DOS. Ifhe did, he might overwrite an another active string. So what he does is search through the entire variable table and find the active string that is already closest to DOS. Then he moves it into the first position in the string storage area. This happens in a flash with the little program we've been using, Since there is only one string variable being used (a$), the Garbageman quickly finds its active value, moves it, and disappears. You probably wouldn't even notice he had stopped by. But programs that have a lot of string variables— particularly programs with large string arrays—really slow the Garbageman down, For each variable in the table, he must look through the entire variable table once. The more variables in the table, the more times he must look and the longer itwill take to look through it each time. A demonstration. If you'd like to measure the effects of the Garbageman yourself, try this little program: 10 20 1250 DIM 65(1) x 40 50 60 FOR X-0 TO I + GS(X)=STRE(x) +: PRINT 68(X) NeXT 70 cr) 8s 30 PRINT “COLLECTING GARBAGE”; CHRS(7) IF PEEK(978)><19@ THEN X=FRE(@) : REN Applesoft/D0S 3.3 IF PEEK(978)=190 THEN PRINT CHRS(4);“FRE” : REM Pro00S PRINT “GARBAGE COLLECTED"; CHRS(7) Line 10 determines how many strings our program will have. Line 20 dimensions Zero Page __.$0000 7—~ Aopiesoft 7 fee — ‘$2000. | Hi-Res Page 1<$3000 ? var rable \Maxttas ¢$67-68> NS EEE Steeno <teo-ee> eis dereso) fees eat) $2079 free Hi-Res Page 2 nenory Free’ memory string storage ‘storage area [$05R@-PURPK INS ‘$9800 |$95°5-SuITCHES Front tor $0000 To SFFFF Figure 1. Applesoft string storage January 1985 a string array of length I. Lines 30 through 60 fill the array with strings. Lines 70 through 90 allow us to measure how long the Garbageman takes to clean up this many strings. You may be somewhat puzzled by lines 80 and 85. To understand these two lines, you first have to know that ProDOS includes its own Garbageman. He is quite a bit faster than the one built into Applesoft, as we'll soon see. ‘A program can tell ifitis running under ProDOS or DOS 3.3 by taking a peek at byte 978 ($3D2). If this byte is equal to 190, then ProDOS is active. If it is anything else, then you are running under DOS 3.3. Ifyouare running with DOS 3.3, you can call the Applesoft Garbageman and ask him to clean up with the command 2=fre(0). To call the ProDOS Garbageman, however, you have to use a DOS command. Thus, in line 85, we say print chr9(4)."fre”. After typing in the program and running it, increase the value of | in line 10. This will increase the number of strings in the program. Garbage collection will take longer and longer. Here are some sample timings: Garbage collection tine in seconds nunber of 00s 3.3 ProDas strings 250 5 on 500 19 0.2 100@ aid 05 Speeding up the Garbageman. Back in January 1981, Call APPLE. pub- lished a program by Randy Wiggington (one of the very first Apple employees— currently famous as the author of MacWrite) that could make the Garbagemaan fly. The program worked by finding the 16.active strings closest to DOS on each scan of the variable table, rather than just the closest one. Rather than searching through the entire variable table once for each variable, the new technique scanned once for each 16 variables. Apple's programmers used a similar technique in the ProDOS Basic.system software. That's why the ProDOS Garbageman is so much faster than the one built into Applesoft. Thus, the obvious way to escape the lulls caused by Applesoft garbage collection is to use ProDOS rather than DOS 3.5. ‘A second way is to incorporate Wiggington’s 505 byte machine language program into your DOS 3.3 program and call it every once in awhile—just often enough to keep Applesoft’s own Garbageman at bay. Both of these solutions have some problems, however. Lots of people still seem to prefer DOS 3.3. ProDOS does give many new features but they come at the expense of memory. ProDOS requires users to have 64K in their machines but leaves less program space than 48K DOS 3.3 does. You could stick with DOS 3.3 and use Wigginton’s program, but it is long, hard to find unless you know where to get back copies of Call -A.P.PL.E., and is copyrighted by Apple Computer, Inc. Selective String Preservation. There is, however, another way to foil the Garbageman. This technique was developed by Clay Ruth and was described in the August 1982 Call-A.PPL-E..The technique, which is relatively easy to implement, is called selective string preservation. Using this technique, you take out the garbage yourself by making a simple poke. Programs that use this technique avoid calls to the Garbageman completely and thus run even faster than ProDOS-based programs. To see how the technique works, let's imagine an elementary order-entry program for a pumpkin factory. Each time the factory receives an order, the name and address of the purthaser is entered. Then the program asks for a pumpkin identification number (they comein many different sizes, shapes, and shades) and the quantity ordered. The program automatically looks up the identification number to gue out if that kind of pumpkin is available and its price. Once the first item is found, the Trogam asks for another pumpkin identifica- tion number (we hope it's a large order for many different kinds of pumpkins). The process continues until there are no more items, atwhich point the program saves the entire order in a disk file and prints an invoice and shipping papers. Then the program starts over and asks for the name and address on the next order. This program, and most other database-+type programs like it, involve two kinds of string variables. First there are permanent string varables. These are variables that are assigned a value only once. They may be referred to over and over again, but they don’tactually vary. The DOS d$is a good example ofa string variable thatis assigned once, is never again changed, but is used throughout most programs. In our order entry pri , we would also have a large, permanent, string array holding pumpkin identification numbers. These would be strings like "RX-11-Pink” or "JC-O-Light’. Whenever a pumpkin ID was entered, we would scan the array to make sure that the ID was valid. The ID's position in the array could also be used as arecord number for a random-access inventory file. In addition to permanent string variables, our pumpkin program would use a number of temporary string variables. These are variables that are assigned one value on the first order, a completely different value on the second order, and soon. After an order has been saved on disk and printed, the strings associated with these variables will not be used again. These temporary string variables hold information such as the name and address on the order and the ID numbers and prices of the pumpkins. Open-Apple 5 Any software similar to our order-entry program is exactly what the Garbageman_ loves to muck up. After entering an order or two, all available memory would be filled up. Garbage collection would be slow because the string array holding pumpkin identification numbers alone would typically have 500 or more itemsin it. Noteven counting all the other strings in the program, the Garbageman would get to scan the 500 array strings 500 times. Your computer will lock up for what seems like forever. Beating the Garbageman to his own game. In the selective string preserva- tion technique, the programmer takes care to make sure all the strings that will never be changed are assigned values first. This will pack them all together in the area of memory right next to DOS. Then the programmer takes a peek at fretop. The first order is taken. Temporary strings are placed in memory, between the variable tables and the permanent strings. When we are finished processing the first order, we poke the value previously found in fretop back into fretop. This tricks Applesoft into storing the next group of temporary strings right over the top of the first group. The second order is taken. Its temporary strings are placed in memory, again beginning right next to the permanent strings. At the end of the second order, we again poke fretop with its original value. Using this technique, we constantly force new temporary strings to overwrite the old, no-longer needed ones. Consequently, the strings never fill up your Apple's memory and the Garbageman is never called. Here's what our order-entry program, which uses the technique, might look like: 100 GOSUB 1000 : REM initialize progran 11@ DIM INS(50@) : GOSUB 1100 : REM load index array 120 FT = PEEK(1LL) + PEEK(112]#256 : REM save value in fretop 138 GOSUB 2000 : 140 GOSUB 2100 : 150 GOSUB 3000 : REM save order on disk 160 GOSUB 3500 : REM print invoice, etc 170 POKE 112,FT/255 : POKE 111,FT - PEEK(112}*256 : REM fix Freetop 180 GOTD 138 : REM do next order REM get name and address REM get items s+sprogram subroutines start here The important lines are 120 and 170. Line 120 saves the value in fretop after all permanent strings have been assigned. Line 170 pokes this value back into fretop when it's time to clear all the temporary strin Implementation tips. Ifyou decide to use this technique, here are some tips. First, use great care in splitting the permanent strings from the temporary strings. In Jong, complex programs, you may have to use more than just two string groups. For example, you may have completely-permanent strings, semi-permanent strings, and temporary strings, In this case you would save the fretop value twice, in different variables. This way you could delete only the temporary stings by poking fretop with one of the values and you could delete both the temporary and the semi-permanent strings by poking in the other value. Don't try to poke new values into fretop until you have all other bugs out of your program. If you change fretop at the wrong place in your program, very strange things can happen. As Ruth pointed out in his original article, itis helpful to know that the pokes to fretop, and not some other bugs, are to blame, When the selective string preservation technique is to blame for the bug, here's what happens. Your program tries to use a variable that was previously given a string value. But instead of the assigned string, the variable contains gibberish, or worse. This is because the original string assignment was placed in the temporary strings area and was later overwritten. Itis easy to get mixed up and think that since you assigned avalue to a variable before peeking at fretop the first time, you can later safely change the value. It kind of seems like the new value should get stored in the permanent string area. But it won't. Remember that the values in your permanent strings cannot be changed. If they are, the new value will be stored in your temporary string area, and the old value will become wasted memory space. One way you can get around this is by assigning values to string variables right inside your program. For example, ifyou havea line in your program such as 1000 fS="scary stuff” the pointer in the variable table will point at the string where it appears in your program. The string "SCARY STUFF” will not be moved into either the premanent or the temporary string storage areas. This only works for program strings assigned with quotation marks. For example, d$=chr$(4) will put a controF-D in the string storage area; d$="": rem invisible control-d between quotes will not. If you are using the selective string preservation technique, avoid the fre(0) command. This command calls the Garbageman and asks him to attack. What we are tryingto do here is haul the garbage out ourselves, thereby completely avoiding the Garbageman’s time-wasting visits. ‘Once you get theselective string preservation technique down, you'll find itquite helpful for those string- and disk-intensive data base programs you are writing, The Garbageman will no longer be yournemesis. Ask (or tell) Uncle DOS Life and times I subscribe to a Time-Life newsletter about Apples that costs $48 a year. Howwill you and Weishaar stayin business with a price only half that? Wiley Catt Okefenokee Swamp The October issue of Time-Life’s letter listed seven executives and nineteen contributors. The committee that puts together Open-Apple is smaller—just Weishaar and six or seven of us imaginary creatures —so our overhead is lower. Time-Life sells its letter to the U.S, business community, which is used to paying too much for computer-related services. Open- Apple's market, on the other hand, isthe world-wide Apple user community, which demands getting its money's worth. Save the trees Why is Open-Apple printed on such lightweight paper? PTT. Bridgeport Okefenokee Swamp It saves postage (particularly to subscribers living outside North America), space in your file drawer, and gives us a distinctive feel (you'll never get Open- Apple mixed up with Byte, for example). The news- letters will be protected by envelopes in all subscrip- tion mailings. From cotton to apples Open-Apple looks interesting and | am consider- ing a charter subscription. But could you please tell me Weishaar's qualifications asa newsletter publisher? Howland Owl Okefenokee Swamp Weishaar (pronounced wise-ar) is a professional journalist who has been working with computers since he was in college. He received a masters degree in journalism from the University of Kansas in 1975. He then worked at Commodity News Services, an electronic news source for commodity traders. At CNS Weishaar started and managed a group of four weekly newsletters. Later, as the company's Manag- ing Editor for operations, Weishaar was heavily in- volved in a changeover of the company's entire editorial department from typewriters and teletype machines to computers and terminals. He was also involved in the development of new electonic termi- nals for CNS customers. The terminals replaced teletype machines and early dot-matrix printers. Weishaar bought an Apple I! in 1980. A year later he left CNS to devote full time to the potential of the Apple. He is an accomplished Basic and assembly language programmer. He has developed two pro- grams, ProntoDOS and Frame-Up, that have been published by Beagle Bros. Both programs appear frequently on best-seller lists of Apple utility pro- grams. ProntoDOS was chosen as one of the best new programs of 1983 in Softalk’s annual reader poll (Softalk, April 1984, page 75). In addition to programming, Weishaar wrote Softalk’s monthly DOStalk column from April 1983 through the final issue of Softalk in August 1984. During that period he also wrote articles (DOS Be Nimble, DOS Be Quick, March 1983; Breaking the Floppy Barrier: An Introduction to Apple's Pro- DOS, January 1984) and was one of the Softalk Sages who answered reader's questions in the IfThen- Maybe column. Weishaar is a member of the Apple Bits Users Group in Kansas City. He is frequently invited to make presentations to groups of Apple users. 80-column puzzle answer (Question on page 6) Toretum tothe 80-column card without clearing the screen do this: 005 3.3 POKE $4,72 POKE $5,195 : CALL 1002 Prod0s PRINT 085 “PRUASC307~ This trick reconnects the 80-column firmware at its warmstart address ($C307), which doesn't clear the screen. With DOS 3.3, you can safely do this by poking the warmstart address into the page-zero output hooks and calling the page-3 vector table routine at $3EA (1002), which forces DOS to use the current contents of those hooks for future output. With ProDOS, this same trick can be accomplished much more easily because the a(ddress) parameter can be used with the ProDOS pr# and in# commands. By the sea Do I need to watch the SuperBowl this year? | missed Apple's Big Brotheradlastyear and my lifehas been a mess ever since, Churchy LaFemme Okefenokee Swamp Things weren't real swell around here, either, though 1984 wasn’t as bad as some predicted. Last year | was in the bathroom during Apple's ad and I, too, missed it. Do you really think that’s the root cause of our troubles? | predict that Apple will buy another SuperBowl spot this year and that it will once again be more interesting than the game. You read it here first. Where is the beginning? I'm interested in doing some interfacing tricks with my Apple, such as hooking two Apples together and connecting stuff to the game port. However, | am a complete beginner at electronics. Where should | begin? Grundoon Groundchuck. Okefenokee Swamp First ofall, if you don’t know how already, youhave toleamto solder (pronounced sodder). Soldering isa technique for connecting electronic components and wires together using heat and a tin-lead alloy. Solder- ing provides a good electrical connection that is also physically strong. In addition to its use in electrical work, solder is also used in many other situations. The gutters on your house are probably soldered. However, if you know all about non-electrical soldering, you know just enough to be dangerous. Electrical soldering requires both smaller soldering irons (the device that provides the heat) and a different kind of solder (the tin-tead alloy). Vol. 1, No.0 Heathkit sells a self-instructional kit on soldering. It costs around $20. It includes an excellent book and some electronic parts you can safely destroy while learning. If you're interested in the hard side of the Apple Il, this kit is a real good starting place. If you don't have a soldering iron, don’t buy one until you need it. The first part of Heathkit's book describes the various kinds of irons that are available and what kind you need. Radio Shack sells adequate ones. Here at Open-Apple we'll help you out with a column called Bus School. It wil be about Apple hardware and techniques for connecting your Apple to the world. Watch for it. Replacement reading material Now that Softalkis gone, what Apple magazines do you read and recommend? Mamselle Hepzibah Okefenokee Swamp InfoWorld, while not an Apple-only magazine, ismy favorite. It's a weekly and concentrates on news in the microcomputer industry. Its editors have a healthy respect for Apples; you won't feel short-changed reading this one. In addition to news, InfoWorld publishes lots of reviews and a couple of very good columnists, Call APPLE. and Apple Assembly Line are my favorite Apple-only publications. When you read these you feel like you're among friends—people who have Apples, use them, and are trying to share their discoveries with all of us. Hardcore Computist isan Apple-only magazine with asingle-minded devotion to one of today’s most pop- ular hobbies — defeating copy-protection schemes. Nibble has always been a whole-heartedly Apple magazine. It is targeted toward new users, however. Forsomereason, | often have the feeling that many of its articles are written by minors. ‘A+ and inCider have always struck me as oppor- tunist publications. Both are late-comers from pub- lshers who first had other computer magazines. Most articles appear to be written by either CP/M holdouts or professional free-lance writers who do articles on Commodores one day and Apples the next. If your main interest is the advertising, A+ now appears to be the thickest Apple magazine. Variable recovery Why is it that the values of all variables are cleared whenever the most minor change is made toa line of an Applesoft program residing in memory? | have worked with other machines that let you change a program line, issue a goto to resume processing, and have all variable values still intact. This is a big aid in debugging a program since you need not start back at the beginning with a run. Is there a poke or some other technique that will restore pointers or values for variables in Applesoft after altering a program line? Harold F. Williams Hutchinson, Minn. The reason Dr. Basic clears all the variables is that your program and its variable tables are adjacent to each other in memory, (see Figure 1 on page 4 of this newsletter.) When you edit a program, the memory space used by the program listing Itself gets either bigger or smaller. To keep the variables intact, they would all have to be moved to keep them adjacent to the program. The other Basics you've worked with probably use a different memory allocation scheme so that the problem doesn’t occur. If you are using ProDOS, however, there is a solution. Before editing your program, issue thestore command. This will save the current values of all January 1985 variables in a file on your disk. After editing, issue a restore command to bring the variables back into memory. Goto (notrun!) will then resume processing with all variables intact. Putting RWTS in solitary Is there any way to edit DOS 3.3 so only the RWTS subroutine and boot up procedure are left? Would this free up space on the disk? Also, on some disks that | boot up, the cursor flashes, but on others, it doesn't. I have searched the DOS manual, but have found no information on either subject. | would be grateful ifyou would answer these questions. Albert Ting Bellevue, Wash. ‘A flashing cursor indicates you are using Apple's 40-column firmware. A non-flashing cursor indicates you are using Apple's 80-column firmware (activated by programs on some of your disks via pr#3). Note that the 40-column firmware cannot display 80 columns, but the 80-column firmware can display either 40 columns (press esc 4) or 80 columns (press esc 8), To edit DOS down to just RWTS, put aLDA $C088,X. UMP $FF69 at $B70E and initialize a new disk (from the Monitor enter B70E:BD 88 CO 4C 69 FF). When you boot this new disk, RWTS will be loaded normally, but then our new instructions will turn off the drive and jump to the Monitor rather than continue loading the rest of DOS. This effectively frees up sectors 10 though 15 on track 0 and all of tracks 1 and 2, however, you will have to unmark the VTOCS free- space bit map on the newly-initialized disk yourself. If you don't know how to do this, see the March 1984 DOStalk. Running without filenames Ineed to create a DOS system that would allow me to simply boot up a disk and run a machine language program at a certain sector on a given track. I'd also like to be able to load other programs, given a track and sector, I wish to do nothing more. No filenames, no file directory, just what I have said. | have tried this before, but soon found | just didn’t have the necessary know- how to doit Jeff Biggus Glen Ellyn, Ii. The easy way to run a machine language program on start-up is to modify the DOS boot routine at $B700-$B749 so that it loads your program instead of the DOS command interpreter and file manager (see previous letter—but don't do what it says there), To do this, start by storing your main program ona disk in ascending sectors. For example, if your program is $1200 bytes long, use a disk zap utility to store the first 256 bytes in track 1/sector 0; the second 256 bytes in track 1/sector 1;andsoon. When you'e filled track 1/sector 15, go to track 2/sector 0 and continue until your entire program ison the disk. Now, to get this program off the disk and back into memory, use the disk zapper to change the number at $B7E0 (track 0/sector 1, byte $E0) to the number of sectors used to store your program; the number at $B7E7 (track 0,/sector 1, byte $E7) to the high-byte address where your program's final sector should be loaded (low-byte of address must be zero); the numbers at $B715 and $B71A (track 0/sector 1, bytes $15 and $1A) to the numbers of the track and sector where that final sector is stored. At $B75B-$B75D (track 0/sector 1, bytes $3B-$3D) put ajump ($4C) to your program's starting address. For example, if your program starts and loads at $2000 and is $1200 bytes tong, put the stuff from $2000-20FF at track 1/sector 0; $2100-$21FF at track J/sector 1; $2F00-$2FFF at track 1/sector F: $5000- $30FF at track 2/sector 0; $3100-$31FF (the end of the Program) at track 2/sector 1 Then put $12 (the number of sectors) at $B7E0; $31 (the high-byte address of the program's final sector) at $B7E7; $02 and $01 (the track and sector of the programs final sector) at $B715 and $B71A. At $B73B put 4C 00 20 (imp $2000) The program will be loaded and run automatically when you boot. Your program can load and run other programs, if they are stored on the disk using the same ascending-sector scheme, by making the above changes in memory, loading the X register with what's at $B7E9 (the slot number of the disk drive times 16), and jumping to $B700. HeDaP heaven All this talk in the computer magazines about the “sony” state of computer documentation is disgust- ing, Users act as if they have a Constitutional right to information about how computers and software work. Butas any student of the history and traditions of data processing knows, this information has always be- longed solely to the Heros of Data Processing (HeDaps). There is nothing new about the notion that an elite, closed group of people should control the technology of the time. That's how witch doctors did it and what's good enough for them is good enough for us. Just imagine the sorry state the world would be in today ifusers had had access to witch doctor informa- tion 10,000 years ago. A careless user might have riddena spirit hard, put itaway wet, and contaminated all the spirits. If anyone could be a witch doctor, then everyone could be a witch doctor; and then no one would really be a witch doctor; and then where would we be? lam nota HeDaP myself. But I do acknowledge that theirs is the power and the glory. Regretfully, | must admit | didn’t always understand that. There was a time when it seemed to me that computers couldn't be any harder to start than lawnmowers. | argued constantly with one big HeDaP in particular about who should be allowed to use the power of our company’s computer. He patiently explained to me many times how damaging us common bunglers can be to big computer systems. Unconvinced and unrepentant, I snuck into an IBM seminar for HeDaPs. The seminar was about a machine called the DisplayWriter. Around the seminar room stood unique beings—half from the world of technology and half from the world of commerce — called Sales Engineers. The Sales Engineers sat each HeDaP down in front of a machine. A little stick man made of letters popped up on the screen and started to dance. Curious, I pressed a key on the machine's keyboard. The man stopped dancing. A Sales Engi- neer approached but neither he nor any of his colleagues could make the man dance again. My machine was dead. I had, with one uninformed key- stroke,killed alittle dancing man —to say nothing ofa computer priced at 10 grand. I was embarrassed and sorry and knew from that moment on that computers must be protected from the masses. Perhaps the greatest weapon we have in this fight is obscure documentation. The early materials that accompanyed CP/M, for example, were so good that one HeDaP told me CP/M stood for Conspiracy to Protect the Ministry. | believed him. Oh, the users gota little uppity there for awhile when those two garage engineers developed the Apple Il. 1 must admit itwas frightening to see tens of thousands of people learning how to use and program a com- puter with the books that came with the machine. Fortunately, it was only Basic they were learning. It is my pleasure to assure you that the Heros of Data Processing have now gained control even at Apple. Those books are no longer standard equip- ment. They cost $50 extra. That should keep the users from nosing around too much. OpenApple 7 Even better, however, is Apple's policy with its MacIntosh. You have to bea certified HeDaP to get that computer's $150 programming manual. And the great joke is that even if users get a copy it won't help them; theyneed a Lisa to write programs — that kills me. The Macintosh is a shoe-in for the HeDaP machine of the year. Believe me, | know howserious the situation iswhen HeDaPs lose control. I've seen War Games; I've seen Tron; | know what happens when you let untrained individuals loose around a computer. I sleep soundly at night knowing that the HeDaPs have the users on the run. However, | must admit | would sleep even better if documention writers would write more like scholars and a little less like news- paper hacks. Even politicans know more about proper English than most technical writers do. For good examples of scholarly writing we need tum. no further than the platform-writing committees of last summer's political conventions. The Democrats had a gem that went like this, “The Democratic party ‘opposes quotas which are inconsistent with the prin- ciples of our country” Isn't that beautiful language? It is absolutely impossible to tell whether the Democrats think all quotas are un-American or if they think some are and some aren't and they oppose only the ones thatare. For their part, the Republicans wrote, "We....oppose any attempts to increase taxes which would harm the recovery...” Do they oppose all taxes or only those that are harmful? You can't tell. This is what scholarly writing is all about. Obscurity is the foundation on which elitism is built. According to a report in the Wall Street Journal, a journalism professor from the University of Kansas told a group of newspaper editors this summer, “Outside this room, there are only three or four people left who know the difference between which’ and ‘that’ Don‘tyou agree it's scary to think that one or two of them might be documentation writers? (I don’t think we have to worry that any of themare politicians.) This unbearably clear-minded professor, whose name is John Bremner, insists writers should always use “that” to introduce a clause that is essential to the meaning of a sentence. “Which”, which should be preceeded by a comma, is reserved for clauses that add incidental information, according to Bremner. | recommend that any technical writers who know the difference between which and that tum them- selves in for reprogramming, Then we'll only have a few newspaper editors and this guy Bremner to worry about. Deacon Muskrat Okefenokee Swamp Indeed, documentation that always has commas in front of whiches is a danger to HeDaPdom, which suits us just fine. Open-Apple © Copyright 1985 b y Tom Weishaar Published monthly. World-wide price: US824/year Send all correspondence to: Overiand Park, Kans. 86207 Usa CompuServe: 120,202 8 Open-Apple Vol. 1, No.0 Become a Charter Subscriber Right Now! Ll Yes, sign meu for 12 monthly tesuea of . . Open-App/e and bill me $24 when you Ordering Information send my first issue. Open-Apple is published monthly. A one-year sub- I 0 No, and I've written why below. scription costs $24. All subscriptions are fully guaran- teed—upon cancellation, unsatisified subscribers will reer ‘the unused portion oftheir subscription payment in full. Special Offer for New Subscribers: Do Not Send Money Now. We will bill you with your first issue. If you decide Open-Apple doesn't interest you, simply write “cancel” on our bill and return. You will owe us nothing, Intemational Subscribers: Open-Apple is sent air mail to all subscribers outside North America. No addi- tional subscription fees are required. We offer a single world-wide price because we want to be the gathering place of the world-wide Apple community. Here‘s Howto Order: Simply fill out our form and mail it. Youwill receive our current issue and a bill for your first prp = $ 3 ® OOODOOOOODODOOOOOOOOI00 year's subscription by retum mail. Comments/Questions/Jokes: Cut out or photocopy our coupon and mail today to: Open-Apple 10026 Roe Overland Park, Kans. 66207 USA. oe a Open-Apple Rare 10026 Roe poor Overland Park, Kans. 66207 Open-Appte U.S.A. Releasing the power to everyone.