Open Apple Vol 1 No 01 FEB 1985

Releasing the power to everyone. February 1985, Vol. 1, No. 1 My Two Bits Tom Weshaar “Congratulations on avoiding the ‘Apple Computer Co. = Apple I computer’ trap’ writes Brooke Boering (author of the freeware programs Fire Organ and One-Key DOS) in response to last month’s My Two Bits. Open- Appleis primarily about the Apple Il compu- ter. But its important to keep a watchful eye on our friends at Apple, Inc., as well. As noted here last month, there is far more respect for the Applell outside the company than inside it There is tremendous technological pres- sure inside Apple to “improve” the Apple II beyond recognition, even though those of us ‘on the buying side are only beginning to understand what we can do with the unimproved model. This pressure doesn’t come entirely from within Apple. The February Popular Computing includes a review of the Apple llc that includes statements like these: “1found the licto be justa second facelift for the Apple Il Plus. Cosmetic surgery is fine in some cases, but nothing short of major internal surgery will modemize the Apple II line” “From a design perspective, compatibility is the lIc’s Achilles’ heel. To maintain that compatibility, Apple's engineers couldn't make the lic very different from its predecessors.” “The licisn’ta bad computer. It's just a computer based on aged technology” ‘The technonerd that wrote that review, like his many confederates inside Apple. has precious little understanding of why we users have purchased more than two million of these machines. We want them precisely because the Apple II has evolved. Ask hackers who were weaned on Radio Shack Model Ones, Commodore Pets, or Texas Instruments 99/4s what they eat for dinner nowadays. All the time and energy they spent learning how to use those lizards has lost its worth. Meanwhile, those of us lucky (or wise, as the case may be) enough to start with an Apple have something of appreciating value inside our heads. Itis Open-Apple’seditorial policy to protect that investment. We'll happily make mock of Popular Computing or any other publication thatattemptsto liquidate our investment. And we'll be watching developments at Apple, Inc. for internal attempts to bankruptus, too. Its important to realize that those of us who already own Apple Ils represent a significant proportion of Apple Computer, Inc's. future revenue. Its a fundamental rule ofbusiness that people who have boughta companys products in the pastare more likely than anybody else to buy again, ‘According to press reports, Apple sold about 750,000 Iles and lcs in 1984 (up from 580,000 in 1983). About a third of those went to homes, offices, or schools thatalready had an Apple in the family. That proportion will increase, not decrease, from here on out. Looking at the data another way, 37 per cent of the people who already own a computer expect to buy another within a year, while only 15 per cent ofnon-owners expect to buy, according to a study done last summer by Dataquest and Software Access International. Apple's president, John Sculley, is a marketing man. Marketing men live by the ‘commandment that a company shalt produce goods that customers want to buy rather than goods that engineers want to produce. Unfortunately, itisn’tat all certain that Scully and Apple are taking the desires of current Apple II owners seriously—even though we represent a ton of future business potential. One gets the impression that their market research is totally based on people who have either never seen a computer before or who have used only competitive brands. The latest example of this disregard for current owners is word from reliable sources that future models in the “Apple II family” won't support DOS 3.3. (If it won't support DOS 3.3 it’s not an Apple II is it?) The reason the new machines can’t use DOS 3.3, Apple's engineers say, is that Apple is converting all its computers to the 5 1/2-inch plastic-encased disks used on the Macintosh. These disks hold 409,600 bytes of data. DOS 3.3, we are constantly misinformed (most recently in the January 1985 inCider, page 82), can handle disks no larger than 143,360 bytes. As the owners of 40-track drives know, this upper limit is all myth. DOS 3.5 will work, without modification, on disk drives having as many as 50 tracks and 32 sectors (the equivalent of exactly the Macintosh’s 409,600 bytes). Apple's engi- neers, however, don'tseem to know this. (Quick, name the only company that sells ahard disk for the Apple Ii that doesn’t support DOS 3.3.) Ican’tbelieve a marketing man like Sculley would leta decision like this be made ifhe knew both the true potential of DOS 3.5 and how his best customers feel about it Since his marketing and engineering minions don't seem to be telling him, | guess you and | will have to. Fire up your word processor and let Sculley know what you think. In case you don’t have it handy, his address is: John Sculley, President Poole Computer, Inc. 20525 Mariani Ave Cupertino, Calif. 95014 Lest | be misunderstood, let me make it clear | have nothing against 3 1/2 inch disks. think they're wonderful. I have nothing against ProDOS. | think it's wonder- ful too, But the common investment of time, knowledge, and productive software Apple users have in DOS 3.3 is simply too valuable to walk away from. If Apple's engineers can't figure out how to get DOS 3.5 to work with 400K disks, there are a dozen or more companies with names like Corvus, Rana, MicroSci, and Diversified Software Research who could show them how. Its time for Apple to spend a little more ofits research and development money on its bread and butter product. Enough of my blustering. The theme of this month’s Open-Apple isn’t monitor- ing Apple—our main task, actually, is to investigate the Apple Monitor. First, however, I want to thank all of you who have already signed on as Open-Apple subscribers. The response rate you've given our subscription campaign has been fantastic— and we didn’t even have a sweepstakes. ASong Called the System Monitor One of the greatest death scenes in all movie history, according to a film critic friend of ours, occurs in 2001: A Space Odyssey. HAL, an intelligent computer on board a spaceship bound for Jupiter, decides the biggest risk to his mission’s success is human error and becomes very hostile toward the astronauts accom- panying him. Dave, a human spaceman, eventually works his way into HAL's memory bank and starts removing ROMs and RAMs (which look like small glass slabs). As each slabis removed, HAL loses intelligence. HAL pleads with Dave to stop. As Dave yanks out HAL's last bit of intelligence, HAL is singing Daisy, which was the first thing he ever leamed. 10 © OpenApple Fortunately, Apples are meek and mild-mannered machines. But ifthe day ever comes that you find it necessary to climb inside and start removing intelligence, the last thing your Apple will sing is a song called the Systém Monitor. This is the first thing Apples ever learned. Itincludes software Steve Wozniak wrote in 1976 to impress his friends the Homebrew Computer Club. Cornerstone treasures, The Monitor software is very elementary. But it's the comerstone of the Apple's operating system. . When people talk about the Monitor, they usually mean one of two things (or three things, ifyou count display screens). In one sense, the Monitor consists of all the machine language subroutines found between byte 63488 (in hex that's $F800) and 65535 ($FFFF). These routines do things such as retrieve the mes- sages you type on your keyboard, print text on your screen, and beep your speaker, In addition, these routines include a program that allows you to interact with your Apple. Using this program, you can determine or change the contents of any memory byte, move data around in memory, verify that two ranges of bytes are the same, and do hexadecimal arithmetic and other stuff. This program, although itis, really just a subset of the entire Apple system monitor, is also usually called the Monitor. The Monitor program, like the rest of the system monitor, Is very elementary. Most of the commands consist of a single character. Error messages consist of a single beep. Nonetheless, it provides a way for you to “get inside” your Apple and deal with it on a byte level —at any time. Very few other computers have a built-in program that opens the system up to you as much as this one does. In and out. The Monitor program starts at byte 65385 (SFF69). You execute the program by entering call 65585, or more typically, call -151 (a negative 151 and a positive 65385 appear to Basic as the same number when you're using the call Vol. 1, No. 1 command). When you enter one of these commands, the Applesoft bracket prompt will change to an asterisk, like this: JoALL -152 The asterisk prompt is your signal that you are “in the Monitor’. Now that we're here, let's figure out if there's anything we can do that's useful. First ofall, try entering a familiar Basic command. How about: SPRINT “I~ (beep) No luck. That beep, however, is the Monitor's signal that you made a syntax error. Notice that no error message is printed on the screen. How about a DOS com: mand? Try this: CATALOG DISK VOLUME 254 A 02 HELLO sand so on DOS commands work just fine from within the Monitor. 'vejust shown aDOS3.3 catalog, but ProDOS works too. With DOS 3,3 youll see a string of numbers following the command's display. ProDOS eats these for breakfast. ifthe numbers intrique you, look at the If Then Maybe column in the June 1984 Softalk. On page Reviewer's Corner ‘The Sider. By First Class Peripherals. The Sider isa 10 megabyte hard disk drive for the Apple I-Plus and Ile that sells, by mail order only, for $695. At this eye-opening price I thought we'd better try one. So far I'm impressed. About four weeks after | placed my order, UPS delivered a massive, mostly empty, box. Inside were the drive itself, a peripheral card for slot 7, and some cables and instructions. The thing takes about two hours to install and initialize. After the Sider is installed, your system uses it for booting when you tum on your computer or press open-apple-reset. If you haven't tumed the Sider on, your computer will “lock-up”; butyou can press control-reset to enter Basic and PR¥6to boot from your floppy rive if you want to go that route. When the Sider starts up, a menu appears that gives you your choice of booting CP/M, Pascal, ProDOS, or DOS 3.5. You can also boot a floppy disk in slot 6 (it usually won't be able to use the hard drive, however), run system utilities, or park. heads (quit) from this menu. One small problem with the way the Sider supports all four operating systems is that you have to set aside at least a few sectors on the disk for each of them. The actual share of space each gets is up to you. The system utilities and the main menu program run under DOS 3.3. Its the only operating system that actually comes with the drive—you must buy the master disks for the others from the appropriate source, as usual. Under DOS 3.3, the Sider appears to be a bunch of floppy disks, each with a different volume number, in slot 7, drive 1 If you initialize the Sider with the maximum amount of DOS 3.3 disk space, you'll get 61 standard 140K volumes, numbered 1 through 61 You access these with the standard DOS 3.3 Vparameter, for example, CATALOG, V35. You can switch back and forth between these 61 “disks”, or volumes, simply by specifying different numbers. If you don’t specify a number, you get whatever volume you used last. Under DOS 3.3, the Sider also allows you to have large, 400K volumes. As mentioned in My Two Bits, DOS 3.3 can handle disks of this size without modifica- tion. You can have up to 22 of these larger volumes at one time. As implemented on the Sider, ProDOS must always have two equalsized volumes. They're called /hard1 and /hard2. Ifyou maximize the ProDOS space at initialization, each volume will have 9,076 blocks of space —a little over 4.5 meg- abytes each. The ProDOS system of prefixes, subdirectories, and pathnames makes a good deal more sense when you have large amounts of space like this to work with, Besides their large capacity, hard disks are also famous for their speed. This is because the disk is always spinning—there’s no coming-up-to-speed delay—and because it spins much faster than a floppy can. The following chart compares how long it takes to sequentially load ten high-resolution graphics from a floppy disk, the Sider, and a RAM (memory-based) disk. It's interesting to note that the operat- ing system you use makesa lot of difference. Ifyour primary interestis speed rather than capacity, a speed enhanced DOS—such as ProntoDOS, DiversiDOS or ProDOS—helps more than a hard disk or even a RAM disk. Seconds required to bload ten hi-res graphics Floppy Sider RAE dicle standard 00S 3.3 88 a 24 high-speed 3.3 20 3 a Prob0s 18 6 2 An unusual feature of the Sider is that one periperhal card can handlea two-drive. system. Most hard disks for the Apple don't have that capability. Those of you with uninterruptable power supplies (review upcoming ina future issue) will be happy to learn that the Sider draws only 40 watts of power. Ifthe Sider has a weakness, it's the documentation. Don’t expect the manual to be as slick as the companys advertising. The manual does an adequate job of explaining how to install and initialize the drive and how to troubleshoot any problems. However, it’s quite thin on actual operation. The “park heads" selection on the main menu, for example, isn’t mentioned in the manual. | assume it’s bad news if the lights go out with the heads engaged, but who knows~the manual doesn't say anything about how to turn the unit off. The Sider comes witha one-year warranty. To receive service you must first follow adiagnostic prodecure outlined in the manual. Ifthe unitis defective you then call to get a repair authorization and return the drive in its original packing to First. Class Peripherals. They fix it free while it's in warranty; they charge $150 after that. ‘Are these folks for real? Apple charges $1495 for its 5 megabyte Profile hard drive— half the capacity at twice the price. This is a new company and buyers take the risk they won't be around long. However, they appear to be doing things right. Their strategy is to sell by mail at below retail prices, Their marketis Apple Il users willing to forgo a dealer's handholding in exchange for those prices. (I suspect there's hundreds of thousands of people like that.) The drive itself is manufactured by a company called Xebec. Members of the Apple users group in Miami bought a bunch of these drives a few months ago and our friends down there say they've been happy with thern. I'm certainly happy with mine. The Sider, by First Class Peripherals (P.O. Box 6187, Lehigh Valley, Penn, 18001; 800-538- 1307). One-year warranty. $695 Downloaded from www.Apple2Online.com February 1985. 141 of that issue, Roger Wagner throughly explores where these meaningless numbers come from and why, Ifthese numbers don't intrigue you, justignore them. Now open the door on your disk drive and try the catalog command again. This is a good way to force a disk i/o error. Why do that? Try itand see: *CATALOS 1/0 ERROR (beep) 1 This time you do get an error message. This is because DOS is sending the message to you. But the important thing to notice is the prompt you get ater the error. A DOS error from within the Monitor always lifts you back up into Applesoft. Enter call-151 again, and try out what happens when you press control-reset. Once again the Applesoft prompt will reappear. So we've established that call -151 gets you into the Monitor and that control- reset gets you back out. Basic commands don't work in the Monitor; DOS com- mands do; however, they send you to Applesoft after an error. Hexamining memory. Now let's try just pressing retum a few times. Voila! You should get a display that looks something like this, but probably with different numbers: * (press return) 8900- A2 00 AO 02 88 BL 3E 49 * (press return) BG08- 3E 92 BC 4A 3¢ 09 BC 99 * (press return) 8810- 00 8B EB CO 55 90 CD AZ Not your normal, school-kid, decimal numbers, mind you; but hexadecimal numbers. In our friendly, everyday decimal system, each digit has ten possible values, 0 through 9. The digit on the far right represents units or ones; the next one to the left, tens; the next hundreds (10 * 10); the next thousands (10 * 10* 10) and so on. ‘The hexadecimal system works exactly the same way, except each digit has sixteen possible values, 0 through F (Startat 0, go up to 9, use A for 10, B for 11; F for 15.) Justas in the decimal system, the digit on the far right represents units or ones; the next one to the left, sixteens; the next 256s (16 * 16); the next 4096s (16 * 16 * 16); and so on. A two-digit hexadecimal number can range in value from $00 (0) to $FF (255). dollar sign in front of anumber indicates it is a hexadecimal number. A four-digit hexadecimal number can range in value from $0000 (0) to $FFFF (65535). Look again at the display you get by simply pressing on the retum key. The numbers on the left edge of the screen are four-digit hexadecimal numbers. They represent an addressin your Apple's memory. The bytes that make up computer memory can be considered small electronic cells that can hold any value between 0 and 255. Each of these cells has an address—kind of like a Post Office Box number, From the Monitor, you can examine the contents of these cells by simply pressing return. In the above example, the contents of cell $B800 (47104) is $A2 (162); $B801 holds $00; $8802 holds $A0; and so on. Each line of the display shows you the values in eight consecutive bytes. The second line begins with byte $B808, the third with byte $B810, Fine, but what ifyou want to see the value in byte $03D0, not byte $8800? Just say so: 0300 0300- 4c You can determine the value currently in any of your Apple's memory bytes by simply giving the address of the byte you are interested in (in hexadecimal) and pressing retum. If you specify an address with a number that has less than four digits, leading zeros will be added automatically. If you specify more than four digits, all but the final four will be ignored. Thus: 2 0e2- 04 "20 020-08 ¥20000 200 4c If you are interested in the values in a range of memory bytes, you have two Open-Apple 11 choices. The firstis to give the address of the frst byte you're interested in and then press retum a bunch of times. This gives you a display like this one: "300 0300- 4c * (press return) BF 30 4C 84 90 4c FO . (press return) (0308- AA 4C BS 87 AD OF 90 AC ' (press return) erEe- 0€ 90 68 AD C2 AA AC CL Alternatively, to see a range of bytes you can give the beginning address of the range followed by period and the ending address, like this: *300.363 e302- 4C BF 90 4C 84 90 4c FD @308- AA 4C BS 87 AD OF 9D AC 03€0- 0£ 9D 6 80 If you give a Jong range, such as 2000.3000, the values will quickly scroll by on your screen. You can stop the scrolling by pressing control-S; restart it by pressing control-S a second time (any other alphanumeric key will also restart scrolling). You can send these memory dumps to your printer, if you like. Use the DOS command pr#1 (assuming your printer is connected to slot 1) to pass output to your printer, Use pr#0 to stop it again. (The Monitor also includes its own versions of the pr# and in# commands, but they interfere with DOS, Always use the DOS versions.) Disassembling your Apple. If you have a silcon-based brain, you may find these rows of hexadecimal numbers very interesting and useful. If, on the other hand, you have a carbon-based brain like most humans, you probably find them difficult to work with. For you the Monitor has a command called List. Beginners. won't find its display any more intelligible than the last, but experienced users find it quite useful. To use List, enter an address, put an L immediately after it (don’t includea space), and press return: "300 0200- 0303 0306- 0303- o30c- o30F- 032- 03€3- oxE5- 0x€3- oxa- oxe0- ote o3eF- o3F2- 03F3- ox6- oF. O3F8- AC 65 FF INP SFFES exe. 4c 65 FF HP SFFGS This range of memory (which you'll only get if you booted with a DOS 3.3 disk) begins at the same byte as our earlier row-by-row display. Again, the numbers on the left indicate the addresses of the bytes revealed on the right. The numbers in the three columns in the middle of the display are the actual values in these memory cells. you compare them to the row-by-row display you'llfind theyarethe same. The difference is that the List command assumes these numbers area machine language program. List organizes the numbers into a meaningful display. On the right side of the display is an assembly language translation of the machine language instructions. For example, the machine language code for jump (equivalentto Basic’s goto) is $4C. The address ofthe machine language program that is to be jumped to always follows the jump command byte. In machine language, addresses are usually written backwards. (No, | don’t know why.) Thus the 4C BF 9Din the first line of our display translates into JMP $9DBF, MP $908F np $9084 ne s6FD JM $8785 LOR $900F Loy $900¢ RIS Loa sence Lov sAACL RTS HP $0851 Nop NOP IMP SFASS 0 STA $4038,x cu om 12 Open-Apple The List command isn’t always able to give a true translation of a range of machine language instructions. For example, look at the question marks near the bottom of the listing, The stuff stored in bytes $3F2 through $3F4 is data, not a command. But the List command has no way to know that It insists on translating the data as if it were a meaningful instruction and gets nowhere. If you have a Ilc, incidentally, the List command will correctly list the additional codes of the 65C02 instruction set, something the List command in other versions of the Il can’t do. Now you know how to list the contents of your computer's memory as raw hexadecimal values and asa machine language program. There is one other useful format for examining memory—as ASCII values. Many bytes hold a number that actually represents a letter or other character. The Monitor can’t show you those letters and numbers directly, but we'll show you how to get a peek at them ina minute. Warming; when examining your Apple's memory with any of these techniques, avoid the area from $C000 to $COFF. This range of memory addresses doesn't contain true memory bytes. Instead, it's full of soft switches that control your Apple. and the peripherals connected to it. Throwing these switches accidentally can mess things up. Non-believers may enter C055 (return) and rejoin us after using controlreset to recover. Making changes. We've looked at how you can enter the Monitor with the call -151. command. There are also other ways to enter the Monitor. You may be familiar with “crashing” into the Monitor, which happens when a machine language pro- gram executes a break instruction. The reset key can also be used get into the Monitor if you make some adjustments to a few special memory cells. Scan back and look at that “data” we saw earlier at bytes $3F2 through $3F4 when we used the List command. What that data actually represents is an address the Monitor wll jump to when you press the reset key, and a check byte that makes sure the reset address is current. If you look at bytes $3F2 and $3F3 in the above display, you'll find they say BF 9D. Tur them around, since they are in the back- wards machine language notation, and you have $9DBF. This is the same address the first line of our display jumped to. Under DOS 3.3, $9DBF is the beginning of the DOS "warmstart” routine. (Under ProDOS, those bytes point to $BE00, which is Basic-system’s warmstart address.) Thus, pressing reset normally warmstarts DOS, which in tum warmstarts Basic, which in turns explains why we go from the Monitor to Applesoft when we press the reset key. If instead of Applesoft you'd like control-reset to put you in the Monitor, all you have to do is change the address at $3F2 so it points to the beginning of the Monitor program. As we leamed earlier, the Monitor begins at $FF69. To make it all work, youalso have to puta $5A in byte $3F4. Ifyou wantto know why, read the June 1985 DOStalk, which examined all this in detail. You might also like to look at Trapping Reset on page 16 this letter. So now we know we want to put 69 FF 5A in bytes $3F2 through $3F4. But how? In addition to allowing you to examine memory, the Monitor makes it quite easy to change the values in memory cells. As before, you give the address you want to startat, follow itimmediately with a colon, and then enter the values you wantin the specified bytes. You must separate the values with spaces, like this: 1372269 FF SA You can change up to 85 consecutive values in one command line —just make sure you puta space between all the values. You can also enter a fewvaluues on one line and continue where you left off in the next line by simply entering a colon with no leading address. ‘After changing some bytes, you can confirm the presence of your new values by pressing return as many times as necessary. The returns will produce a memory display that begins at the last complete address you entered. Watch: *2000:11 22 33 #244 55 56 1:77 68 95 AA 8 CC * (press return) 2000- 11 22 33 44 55 66 77 88 . (press return) 2008- 99 9A BB CC 02 00 00 e@ Let's assume you've made the 3F2:69 FF 5A change to fix your Apple so that control-reset will always take you to the Monitor (until the next time you re-boot, anyhow). Now how do you get back to Basic? The first line of the disassembly display we looked at earlier also contained a jump to the DOS warmstart address. The Monitor has a command, called Go, that allows you to execute the instructions at any memory address. To get back to Basic Vol. 1. No.1 now that we've changed how control-reset works, we can do this: #3006 1 In this command, 3D0 (note that's a zero, not a big 0) is the address we want to jump to, the G tells the Monitor to jump. This address, $3D0, is known as the DOS warmstart vector, and isa worthwhile address to remember if you use the Monitor much, Let me see letters. Control-Y is known as the Monitor's user command. Assembly language programmers can write a special routine, place its address at $5F9-$3FA, and execute it directly from the Monitor by entering controF. Here's a special routine that you can use to change the Monitor’s memory dumps from hexadecimal numbers to their equivalent ASCII characters: #35200 03 (put address of routine at $3°8) *3@0:AS 3C 25 @7 00 @3 20.92 (put routine at $300) #30 BL 3C 48 29 7F 9 20 #260 08 68 09 08 48 A OE #200 02 9 AO 20 £0 FO 6B #:20 [0 FO 20 6A FC 90 08 1:60 After typing all this in, check it carefully. If you heard any beeps as you were entering the line, you made a mistake— probably you substituted an 0 for zero. Now you can examine memory in ASCII by entering the examine memory com- mands just as before but also pressing controFY before pressing retum. Charac- ters preceeded by an ‘in the display are control characters: *690.AB00(contral-Y, return) nege-")"€ 2 INIT seg8- LOADS AVE This monthyou've leamed how to examine the contents of your Apple's memory as rawhexadecimal values, asan assembly language program listing andas ASCII characters. You've also learned how to enter new hexadecimal values into memory. Next month we'll examine some of the fancier things the Monitor can do, such as move and compare values. For now, however, let's look at some Monitor tricks involving Applesoft, DOS, and Logo. Picking Up Applesoft f The Basic connection. Getting the Monitor to dump memory in ASCII as we've Justdone is a nice feature, but typingin all those hexadecimal numbers is a pain in the chips. If we were dealing with Basic, we could write a program and save it, but there's no way to do that from the Monitor. We could save those two ranges of memory we've typed in as binary files and write a Basic program that would reload them, but then you wouldn't leam anything new. Several years ago a fellow named S.H. Lam wrote a letter to Call -A.PPL.E. in which he showed how to execute Monitor commands from Basic. This technique has since been used all over the place. | even used it in DOStalk in March (DOS 3.3 disk free space patch) and May (DOS 3.3 type command). The technique provides for entering editing, and saving machine language programs and requires relatively few characters to be typed, which cuts time, effort, and mistakes. Unfortunately, people often have trouble getting all the Is dotted and Ts crossed when using this technique — especially when typos sneak into program listings, as, happened in the May DOStalk. The technique works like this, You write an Applesoft program that simply defines whatyou would normally type into the Monitor asa string. After defining the. string, your program jumps to a subroutine that gets Applesoft to type the string into the Monitor for you. Here's the magic subroutine: February 1985, 500 CS=C$ + “ N DSCEG” : REM space required before and after N 510 FOR I=1 TO LEN (C$) S12 + POKE SLL+T, ASC(NIDS(CE,1,1))+128 S14 NEXT 520 POKE 72,0: CAL -144 530 RETURN The first line of this routine tacks IY D9C6G onto the end of your command string. Note that there must be a space before and after theN. More than 95 per cent ofall reported problems with this techniqueare caused by a missing space between the opening quote and the N. if your command string (C$) is 500:0, then after line 500 is executed it wll be 500:0 N D9C6G. Lines 510-514 poke this command string into your Apple's key- board input buffer. Line 520 calls a machine language subroutine inside the Monitor that executes whatever's in that buffer. Inthis case the Monitor would puta zero at byte $300, set display to normal, and Jump to $D9C6. $D9C6 is a charmed spot inside Applesoft that continues execu- tion of our Applesoft program. In this case we continue at line 5350's return from the Lam subroutine. ‘One problem with the Lam technique is that the shortcut we used earlier of entering colon without an address doesn't work. Here's how you would set up an ASCII memory dump loader using the Lam technique: 39:00 03° : GOSUB S20 -300:AS 3C 29 07 D0 03 20 92" 128 0$="308:FD 81 3C 48 29 7F C8 20° 130 C$="310:80 08 68 09 CO 48 AB DE” 140 C$="318:08 @2 A AB 20 ED FD 6B” 320:20 ED FD 20 BA FC 90 08" 160 C$="328:60" : GOSUB Seo 170 END + GUSUE 500 GOSUB S00 G0SUB 500 GOSUB 500 GOSUE 500 500 C$=C$ + “ N O9CGG™ : REM space required before and after N 520 FOR T=1 TO LEN (C8) 512 + POKE SL1+1, ASC(MIDS(CS,1,1)}+128 S14 NEXT 520 POKE 72,0 : CALL -144 530 RETURN Tosave even more typing time, the lines from 100 to 170 could be combined. But there's an even better way, and its a DOS trick. Digging Into DOS Writing Monitor programs. Earlier in this letter, | pointed out that DOS com- mands work from within the Monitor. This flexibility leads to an interesting trick. When considering DOS commands that you can use inside the Monitor, you will think right away to try catalog, but how about fp or bload myfile or -/users.disk/ convert? Here's the real beauty, though, exec ascil.dump. The exec command works just fine from inside the Monitor. What you put in the text file you'll exec is bare Monitor commands. Instead of typing in a whole Basic program, all you really have to do is sit down at your word processor and enter: CALL -151 (optional, not needed if you exec From within the Monitor.) 3F3:00 03 00:95 3C 29 07 09 @3 20 92 308:F0 81 30 48 29 7F C9 20 0 08 68 09 CA 4B AS OE @2 AB AO 20 ED FO G8 32:20 ED FD 20 8A FC 90 08 328360 (a5 with the Lam technique, you must use @ loading adsress on each Li Use a word processor to create an exec file filled with Monitor commands. This gives you the ability to edit your mistakes and to save your work. Now we can write “Monitor programs”. It works with both DOS 3.3 and ProDOS. Ifyou get into this, you might want to use it to edit our control-Y program a little. Change byte $303 to $0F and you'll geta display 16 columns wide. A$1F at that spot will give you a 32-column display —anice use for your 80-column screen (which, of. course, you can tum on from within the Monitor with a pr#3). It's not the kind of thing they teach your children in school, but the Apple Monitor is readily available from Logo just as itis from Basic. With most Logos, the ‘command call -151 or its equivalent is probably the most convenient way to get into the Monitor. With Apple's 128K version, however, getting into the Monitor requiresa few tricks. You can't simply call the Monitor's starting address, because Logo tums on and uses all 128K of RAM memory. The Monitor ROM is tuned off and is invisible when you're inside Logo. On Apple's 128K computers, pressing Reset automatically turns on the ROM, makingit momentarily accessible. Normally Logo trapsa press of the Reset key, as we did earlier in this letter. It quickly tums the ROM back off, the RAM back on, and informs you that you've pressed the Reset key and should reboot. The reboot message seems a little heavy handed, however, since your program and data are still intact and ready to run. To getinto the Monitor, we can change the contents of the soft-entry vector. Then Reset will provide our passageway in. This is easy enough. The problem then becomes one of getting back out again. Apple's 128K Logo doesn’t support the standard $3D0G warmstart address. Tosolve this, let's move the contents of Logo’s soft-entry vector into the control-Y vector. Then when we press control-Y from inside the Monitor, we'll go back to Logo Via its Reset trap, and all necessary bank switching will be handled for us. Logo commands will look like gibberish to those of you who have never encoun- tered them before, so we'll explain them a little. In Logo you write programs by creating a series of subroutines, called procedures. Each procedure is given a name and you execute the procedure by simply saying its name. You might, for example, write a routine that draws a house on the high-res screen and call it house. Every time you enter the word another house will appear on the screen. Let's write a short program, called mon, that will poke the changes we need into the soft-entry and contro-¥ vectors. Apple's 128K Logo does haveall the primitives (built-in procedures) we need for this. Ifyou are using a different version of Logo, your version may have different names for these primitives. .DEPOSIT adr valis like a Basic poke. .EXAMINE adris like a Basic peek. First, let's write the procedure that will do the actual poking, TO DD.NON.POKES = -DEPOSIT 1017 .ExAMINE 1018 DEPOSIT 1018 EXAMINE 1011 DEPOSIT 1010 105 -DEPOSIT 1011 255 DEPOSIT 1012 90 ND The to tells Logo we are writing a procedure. This causes what we type to be stored, rather than executed immediately; the same effect line numbers have in Basic. The effect lasts until we type end, which tells Logo we are done with that procedure. After the to you put the name of your procedure. On the next and following lines. ‘come what you want to do. The first two lines of our procedure move the Reset-trap address that Logo keeps in the soft-entry vector into the control-Y vector. The final three lines poke the address of the Monitor program into the soft-entry vector and correct the power-up byte. One problem with do.mon.pokes is that if you run it twice, by accident or otherwise, the control-Y vector ends up with the Monitor's address in it and you can't get back to Logo. Let's write another procedure that will call do.mon.pokes, but only if it hasn't been run already. We can tell by looking at byte 1018, which always equals 255 before running do.mon.pokesand some other value afterward. TO HON IF 255 = .EXAMINE 1018 [00.MON.POKES) 0 That's all there is to It. To execute, type mon, then press Reset. Detecting NOT SELECTED Is there an easy way to detect that a printer is off before the NOT SELECTED message appears? Harold D. Portnoy Bloomfield Hills, Mich. (For those of you who don't have the Grappler+ interface card, the Grappler places a beep and the words “NOT SELECTED in the middle of the screen when you try to print something and the printer is tured off or deselected.) The standard answer to this question is no. How- ever, if you are writing software for your own system and thus know in advance what printer and interface card your program will use, it can be done quite easily. Enter the Monitor with acall-151 and take a look at the softswitches associated with your printer. There are sixteen of them; they begin at $C0B0+s0, where = your slot number, If your printer interface card isin slot 1 for example, look at the switches $C090.CO9F, ‘Sart with your printer tured off. e° off) 51 51 S1 SI S151 51 51 51 Si S151 SI your printer on: leave unselected) 2.CeSF 50 $0 SO SD 5D SD 50 50 SD SO SO 50 SD 50 er on; selected) 52 $3 53 $3 93 53 53 $3 53 53 52 5253 53 The values shown here are what you'll see if you have a Grappler*, Other cards will display other talues. All you have to do now Is peek at one of these locations and see if the expected value is present. For example: NOT" PEEK ( 45260+5. NT NS;~ SELECTED RUN this program while selecting and deselecting your printer. If you have a lic, you'll find that it can’t tell the difference between a printer that is selected and one that's turned off (though it can tell selected from deselected ifthe printer is turned on). I suspect this will be true of al serial printer interfaces. ProDOS and compilers Can | use a DOS 3.3 Applesoft compiler (Einstein, ‘TASC, Expediter), transfer the object code to ProDOS 16) = 83 THEN NS=""=REM 63-553 2 GOTO 20 using Convert, then patch the 1/O hooks to get Pro- DOS to recognize DOS commands? ' Michael Couvillon Ft Walton Beach, Fla. T haven't tested it, but theory tells me it will never work, ProDOS uses a different scheme for recogniz- ing DOS commands than DOS3.3. Anybody out there know of an Applesoft compiler that can produce ProDOS-compatible object code? FIND:/user group/ Doyou know of any Apple Il user groups in my area (northeast comer of Oklahoma)? Claude E. Williams Jay, Okla. You're surrounded by them—but they're all about 60 mites away. I don'thave anything on these groups except their addresses, but you could write for more information: Tulsa Computer Society's Aople Users °.0. Box 1133 Tulsa, 0K 74101 Joplin Apple Users Group 1903 £. 36th Joplin, MO 64801 Razorback H.8.U.6. 3000 Rogers Ave. Fort Snith, A 72301 Hfyowhave a modem, the closest groups to youare probably the Micronet Apple User Group on Comput Serve and the Source Apple User Group. Incidentally we're trying to develop a data base on Apple user groups here at Open-Apple. We also exchange subscriptions with groups that have news- letters. Msg from a parallel universe > CEREAL INTERFACE SE SOUPER CEREAL. /NTERFACE Itried the programs dealing with the onerrroutines in the January issue (page 2) and put the "empty’ file on the ProDOS /ram volume. This speeds things up greatly and makes life easy on the old disk drives. It also executed the onerr routine 8,400 times without fail while | watched the late night news. What else are people doing with /ram? TS. Goodfellow Raleigh, NC. Mousetrapped There seems to bea lack of good information on the lic. | hope your publication can supply that need as well as others. For instance, | feel suckered now for buyingMousePaint. as [have been unable tofind any other programs that use the IIc mouse. Help. Tom Milford St Louis, Mo. Vol. 1, No.1 More mouse software should be upcoming soon. Apple has provided software developers with a Macintosh-like interface for the Apple Il series that uses the 80-column text screen, the mouse, and the infamous "Mousetext” characters that occasionally ‘appear on your screen where inverse capitals should be. Software using this interface will likely begin to appear by summer. Tm not completely sold on the mouse—we have a lot of trouble around here with the ball sliding, rather than rolling, across the desktop; this will drive you nuts with great speed. I like the features of Apple's human-computer mouseface, however. Stalking the I/O Error I recently bought Terrapin Logo for the Apple and the disk wouldn't boot. Just for the heck of it. | went under the hood and reversed my drives. To my sur- prise, the disk booted ok on my former drive 2 (then the door hinge broke). At the same time, along the same lines, I got an unprotected demo disk thatwouldn'tload. Using Disk Recovery | checked it out and got a report that both sides were full of bad sectors, but, strangely, the pro- gram couldn't seem to make up its mind about which sectors were bad—they changed every time I ran the program. After my second drive was fixed, the front side breezed through the testwith no errors reported. also tested a couple of unprotected commercial disks that were perfectly functional. However, one con- tained many mis-assigned sectors and the other had an invalid track/sector list. In both cases | suspect that the producers resorted to subterfuge to conceala very substantial part of their programming and the combi- nation could cause the unwary to delete a large chunk of an otherwise valid program. Gardyloo, Ralph Moredock ‘Campbell, CA Thave several floppy drives; the oldest has been my most-used drive for almost four years. None of them have ever been adjusted, aligned, or (blush) even cleaned. | did take one apart once to see what was inside, but | didn't tueek anything. Based on this experience, I suspect many disk drive problems start as software problems and become hardware problems onlyafter the drive gets “adjusted” inanattemptto fixthings. Those of you oho have had real hardware problems will probably vehemently disagree with me. The Number One cause of disk problems is bad centering. Disks must be perfectly centered in the drive or they won't work. There isa mechanism inside the drive that is supposed to center the disk, but it doesn't work reliably. Hub rings on your disks help solve the centering Problem. Most disks sold nowadays come with these; I recommend you refuse to buy any that don't have them. The hub ring stiffens the magnetic material, consequently the centering device does its job better. The original purpose of hub rings was to keep the center of the disk from wearing out, but this hasn't beena problem on 5-1/4 inch disks. ‘The best tip for centering disks, however, is to wai until the drive is spinning before closing the driv door. Yes, | know all the books say not to do that. Jus never open the door while the drive is running—o more exactly writing on the disk. That can destrot disks, Butclosing the door with the drive on not only doesn't hurt anything, it actually helps, Adisk that is accidentally initialized while off-cente February 1985 isparticularly useless. As soon as you remove it from the drive you've lostany data you've saved oni. Youll never get it back in the drive in just exactly the same off-center orientation youstarted with. Thus, itis very important to delay door-closing when initializing disks. Whenever you have a commercial disk that won't boot, but just sits in the drive and spins, open and close the drive door. This will center the disk and allow it to boot about 90 per cent of the time. And while we're talking about doors, let's mention that it may not help, but it certainly doesn't hurt to openthe doors on your drives before you turn off your compu- ter, That little flicker of the in-use light could mean the drive head is flickering too. If your disk is centered property and still doesn't work, the next most common problem is electrical interference from your display screen. Monitors and television sets work by guiding an electron beam across the screen with electromagnetic radiation. If your disk drive is close enough to the screen, this radiation wil! overwhelm the tiny signals the disk head is picking up from the magnetic fields recorded on the disk. To solve the problem, simply move the disk drive farther away from the display screen. Magnetic inter- ference occurs quite often with the IIc because the drive is naturally close to the screen. And it occurs almost every time youtry to usea Ilc while it's sitting ‘on top ofa television set. If the disk still won't boot, try changing to a differ- ent model of Apple. | was about to throw out two commercial, copy-protected disks that I thought had shot craps; however, I finally figured out that both boot just fine on my II-Plus. It's only on Hes and lcs that they won't work. This ts one of the many joys of copy protection. Finally, if your disk still won't boot, get out your DOS 3.3 System Master and BRUN BOOTIS5. | really have purchased a commercial program within the last year that used DOS 3.2. These disks will not boot on today's Apples unless you boot via BOOTIS. Ifthat doesn't work either, you've covered about 98 percent of the probable problems. If we split up the 2 per cent that’s left, there's a L9 per cent chance the disk is bad, a .09 per cent chance of some other problem (try cleaning the edge contacts on your disk drive card), and a.01 per cent chance your disk drive is bad. The only clear sign that a drive needs to be adjusted ts that disks written with one drive can’t be read by another and vice-versa. One of those suckers needs fixing. All you have to do then is figure out which one, As to the unusual things Disk Recovery found on unprotected commercial disks, | suspect an attempt to enhance performance rather than subterfugeis the root cause. For example, GPLE uses. very strange file structure. You can't successfully copy it by BLOADING it, finding its loading address and length, and BSAV- ING it with those parameters, The reason for thisisn’t subterfuge, however, but to get it to load quickly. AppleWorks vs Oki Thave an Apple tle, Grappler, and Okidata 92. Any idea how | can get AppleWorks to work with my prin- ter? As usual, my Apple dealer doesn’t have the foggiest Arlene Samowich Nashville, Tenn There is an answer to your problem. I just don't know what it is, But here's a note | found in the Princeton (NJ) Apple Users Group newsletter, “Clancy- Paul (a dealer in Princeton) has ‘found’ the Apple- Works configuration disk that Bart Thomas donated to them and spoke about at the last meeting. Anyone who has other than an Apple Parallel Interface Card, ‘or who's using an Okidata 92 printer, will probably need this fix to print properly” From yet another note in the same newsletter, | know that Bart Thomas is an alternate system opera- tor for the Micronet Apple User Group on Compu- Serve. Thus, I suspect Bart could put the fix you (and many others) need on MAUG (if t's not already there) and you could download it. if the fics available from other sources I'm sure one of our readers wil tell us about it. Dump 80 Explain how to get an 80-column screen dump to printer and Il sign up for 100 months. Steve Madigan Culver City, Calif To do a screen dump, you have to peek directly at screen memory to get each character. Screen memory has a Byzantine organization, so begin by setting up anarray that holds the starting address of each of the 24 lines on the screen, like this: 108 IM LADR(23) 110 FOR I= TO 23 : READ X + LADR(I)=X : NEXT 120 DATA 1024, 1152, 1280, 1408, 1536, 1664, 1792, 1920, 1064, 1192, 1320, 1448, 1876, 1704, 1832, 1960, 1104, 1232, 1360, 1488, 1616, 1744, 1872, 2000 Our program will use one loop to advance from tine toline. Asecond loop inside the first will peek ateach of the 80 characters on a line. The trick is that there aren't really 80 different addresses for each line. There are only 40 different addresses —but half of them (odd columns if you consider the first column number 0) are in main memory and half of them are in auxiliary memory When the 80-column mode is active, you can switch back and forth between the two memory banks by hitting softswitches at 49237 (even columns) and 49236 (odd columns). Thus, the rest of our program looks like this (don’t forget to include the lines 100 to 120 above!): 200 IF PEEK(78)=198 THEN 250 : REN 3.3 or ProOOS? 210 POKE 43602,0 : REM brute force <cr> for 005 3.3 250 PRINT CHRS(4);“PRHL” : REM printer on 260 PRINT CHRS(9);°9ON" : REM dont mess up screen 300 FOR LINE = @ 10 23 305: ue 320 330 348 350 360: DO + NEXT 380 : PRINT LS 390 NEXT OR CHR=@ TO 39 OR=LADR((LINE) +CHR POKE 49237.@ : REM peel even coluans (S=_ 5+CHRS (PEEK(ADR) ) POKE 492360 : REM peek odd columns + LS=LB+CHRS (PEEK(ADR) ) 400 REM Rehook 80-col screen without erasing it. 410 REM See tip on page 6 of January Open-Apple. 420 IF PEEK(978)=190 THEN 450 : REM 3.3 or Prod0S? 430 POKE S4,? + POKE 55,195 : CALL 1002 : REM 005 3.3 449 SOTO 508 450 PRINT CHRS (4); “PRHASCIO7” > REN ProdOS 500 END To get this program to work I had to resort to a couple of tricks you may not have seen before. Inline 200 we determine if our program is running under DOS3.3 or ProDOS. If 3.3, we need to print aretumso Open-Apple 15 DOS will recognize the DOS command in line 250— but a retum may mess up our screen display. Thus we resort to the brute force technique in line 210. In lines 250 and 260 we tum on the printer. Line 260 has two functions. First, it tells your printer inter- face card that we don't want what we print sent tothe screen. It would mess up the very display we are trying to dump if it did. Secondly, it tells the card to. expect 90 characters per line. Since we will be print- ing 80 characters plus a carriage retum for each line, the card will actually see 81 characters perline. If you use the standard CHRS(9);"80N" in line 260, your dump will be double-spaced because both your pro- gram and your interface card will send carriage returns after 80 characters. Lines 300 through 390 scan the display memory and place the characters found in the string variable LS. After gathering a line, L$ is printed in line 380. The routine would be a lot faster if we could print the characters as we gather them in lines 340 and 360 rather than saving them in a string for printing later. However, for reasons! don't understand, the system I tested this program on (a IIc/Imagewriter combina- tion) consistently garbled things when I tried to PRINT while the even columns were tumed on. Can any of ‘you wizards out there explain that? Finally lines 400 to 450 tum the printer off and returnto the 80-column screen without erasing it. For the details on how this works look on page 6 of last month's newsletter. That will be $200, please. Problem too complex Here's a little program you may find interesting (1 find it perplexing). LIST 10 ONERR GOTO 100 20 PRINT “HELLO FROM LINE 20” 30 Ae" 40 END 102 I=1+1 : PRINT I 11@ CALL -3268 : GOTO 20 RUN HELLO FRON LINE 20 1 HELLO FROY LINE 20 2 HELLO FROW LINE 28 (program hangs) (you press Reset) just FORMULA TOD COMPLEX ERROR ] ‘The call -5288 that yourecommended in the January issue (page 2) doesn’t work if the error is type mis- match or string too long. The computer hangs and you have to hit Reset. Then, when you try to list the program, you get ?formula too complex. A second attempt to list the program usually works, What's going on here? I'm extremely frustrated at trying to bullet proof my programs. Brent Thorwall Crystal Lake, Ill. T haven't been able to figure this one out—maybe somebody else can help. The stack appears to be getting messed up, but in looking at the Applesoft code I can’t see why it would. ! have figured out that the error isn’t absolutely related to type mismatch — try replacing line 30 withA$=A, also a type mismatch error, and it runs just like it’s supposed to. I'm sure there's an explanation for this, anybody out there got an easy answer? 16 Open-Apple Trapping Reset I'l be happy to contribute $24 a year to keep you productive and off the streets. How can I direct a program to line x or line y after Reset has been hit? Dennis Mah Federal Way, Wash. My kids thank you. Your question is short, to the point, and theoreti- cally very easy to answer. So why hasittaken me two days to figure it all out? In the process I've made an incredible, new, useless discovery about the 6502 microprocessor. Please, no more questions like this one! Part of the reason your question was so hard was that | was sure | could answer it witha few PEEKS and POKES. | wanted to do this completely from Basic. | didn’t want to have to stick any little assembly lan- guage routines anywhere. The reason | had trouble is that pressing Reset, unbeknownst to me, scrambles the contents of the 6502's stack register. Applesoft uses a small section of the stack to parse command lines. If the stack register accidentally points at this area after a Reset, Dr. Basic trips over his own feet and tumbles to the ground. In addition to scrambling the stack pointer, press- ing Reset disables interrupts and causes the 6502 to ‘etch the address stored at SFFFC and jump there, On ail but the earliest Apples SFFFC points to $FA62. The routine there is called, appropriately enough, reset. That routine puts the 6502 in binary (rather than 4ecimal) mode; tums on normal (rather than inverse) characters; turns off graphics and tums on full- cindou, 40-column text with the cursor at the bottom: ‘ofthe screen; connects the keyboard for inputand the display screen for output (disconnecting DOS from the | Ohooks in the process); turns offannunciators 0 and 1 and tums on annunciators 2 and 3; clears the keyboard strobe; tums off any active peripheral card ROM: and rings the Apple's bell. Inthe He and llc, special hardware circuits also tum ‘on the motherboard ROM and the main bank of RAM. Auxiliary RAM is disabled. On earlier Apples with 64K of memory on the other hand, Reset does not change the memory configuration. After doing all of the above, the reset routine takes a look at byte 1012 (93F4). Ifthis byte, called the power- up byte, contains the proper value, the routine next jumps to the address stored in a place called the soft-entry vector, whichis at 1010 ($3F2). Ifthe power- up byte contains can other value, as it always does Open-Apple © Cope 1985 ‘Tom Weishaar Published monthly. World-wide price: US$24/year ‘Send all ism wagerar of Tom Computer and {wo diferent unrelated. independent companies Path aveyene'n ne woidhad en Apple fom Weishaar Apple Open immediately after you tum on your computer, reset will attempt to boot a disk. ‘The address usually stored in the soft-entry vector is the DOS warmstart address — 40383 ($9DBF) for DOS 3.5; 48640 ($BE00) for ProDOS. Thus, the next thing Reset normally does is warm- start DOS. For DOS 5.3, this means the language card is switched back to whatever language had been in use when Reset was pressed; DOS is reconnected to the I/O hooks; a NOMONICO is executed; and ajump occurs to Basic’s warmstart address, 54332 ($D43C), The Applesoft prompt and cursor appear on your screen. Your program and data are intact, but your program is no longer running. ProDOS‘s Basic.system, on the other hand, handles warmstarts a little differently, It first jumps to the CLEAR command inside Applesoft at 54885 ($D665), which empties out all yourvariables and leaves them filled with zeros. This isn’t real handy if you'd like to recover from a Reset without losing any data. The CLEAR routine also fixes the stack register, however —something DOS 3.3 neglects to do. Next ProDOS reconnects itself to the 1/O links, pushes the cursor cover to the left edge of the screen, and warmstarts Basic at 54335 ($D43F). Again the Applesoft prompt and cursor appear on your screen. Your program is intact but no longer running: your data is gone. (Note the two Basic warmstart addresses are differ- ent by three bytes. ProDOS skips the first warmstart instruction, which sends a carriage return to the screen. AProDOS reset doesn't mal the screen scroll. ADOS 3.5 reset does. This is why. It’s also why Pro- DOS has to push the cursor to the left edge of the screen by hand.) Now you know everything there is to know (and much more than you probably wanted to know) about Reset. Your original question remains, however —how do you keep a Basic program running and send control to a special error routine when Reset is pressed? Intheory all you have to do is change DOS so that Instead of warmstarting Basic, it jumps to a Basic error handler. Applesoft doesn't have a “Reset” error, thus we'll have to use another one. There’sa good one at 58123 ($£30B)—illegal direct. This error occurs when you try to do an INPUT, GET, or a few other commands from the keyboard. It never occurs in a running program. So let's modify DOS a little; a sim- ple task, DOS 3.3 keeps the Basic warmstart address in an internal vector table; it's at $9DSE (40286) when DOS is at its normal 48K location. ProDOS, on the other hand, doesn’t use a vector table for this. In fact, the bytes you have to change move around from one ProDOS version to another. Luckily, they're easy to find: first look at the ProDOS warmstart vector at 48640 ($BE00) to determine where the actual warm- Start code lies; the bytes that must be changed are the ‘th and 12th bytes of the warmstart code (at least that's where they've been inall Basic.system versions to date). After we make these modifications, Reset, in con- junction with an ONERRGOTO statement in your pro- ‘gram, will pass control to ONERR with the code foran illegal direct error (149) stored where you can fish it out with the standard PEEK(222). That's the theory anyhow. But Basic still occasion- ally crashes under DOS 3.3 because of stack prob- lems and under ProDOS we lose all our data. Let's fic these problems, too. With ProDOS, we'll change the _jump to the CLEAR subroutine into ajump tojust the Stack initialization subroutine, which lives at byte Vol. 1, No.1 54915 ($D683). With DOS 3.5, we'll put the jump to the stack initialization subroutine over the top of the automatic NOMONICO (good riddance!). Whew. Alittle program that doesall this stuff follows: 10 REM Reset trap 123: 91S + REM address of FP illegal direct error REM address of FP stack initialization 30 IF PEEK(978}=157 THEN $0 : REM 005 3.3 active 35 IF PEEK(978)=190 THEN 60 : REM ProO0S active 40 PRINT “ACTIVE 005 NOT RECOGNIZED." : END 50 Al=40286 : REM 48K DOS 3.3 Basic uarnstart vector 52 A2=40401 : REN auto NONONICO+1 53. IF PEEK(40409)=0 THEN 58 S4 FOR 1=40403 TO 40400 STEP -1 : +: POKE 1,PEEK{(I-3) 56 NEXT REM 58 GOTO 70 + REM shuffle sore REM 00S 3.3 stuff 2 REM around (First RUN only!} to make it work. 60 A2=11PEEK (49641 ]+PEEK( 48642] #255 62 AL=AZ+10 = REM fish out ProDUS adrs REN (58E20) +1 70 REM 2-nyte pokes adrjvalue 72 POKE Al+L,VI/256 = POKE AL,V1-(PEEK(AL+1] #256) 74 POKE A2+1,VZ/256 > POKE A2,V2-(PEEK(AZ+1] #256) 80 ONERR GOTO 200 100 GOTO 160 : REM lockup 200 PRINT “ERROR W; PEEK(222); ~ IN LINE “; PEEK(218) + PEEK(219)#256 Run the program (nothing will seem to happen—it locks up in tine 100) and press Reset. Line 200's mes- sage should appear on your screen. Since the illegal direct error won't ever occur naturally inside a pro- gram, you can always be sure its appearance means somebody pressed the Reset key. Delete line 80's ONERRand RUN the program again. Now Applesoft's built-in message center will print illegal direct error in 100. If you haven't saved the program to disk yet, do it now. Put line 80 back in and add 210 RESUME. RUN and press reset. You'll find yourself in an eternal cycle oferrors. Too many things get changed when Resets pressed to simply resume execution. Instead, warm- Start your program with a GOTO. Inthis case, use 210 GOTO 100. RUN again and try pressing Control-C as well as Reset. That, Dennis, is how you direct a program to line xafter Reset has been hit. The only way to get out of our sample program now is with a hard Reset (control-open-apple-reset or turn the computer off). Sure hope you saved it like you were supposed to. Remember that Reset tums off graphics and 80- column displays, reconfigures memory, and makes other subtle system changes. Your ONERR routine (or your warmstart) may have to take special steps to set these things aright again. The modifications we made here should never be permanently added to DOS. They cause a lot of trou: ble if you don't want Reset trapped. DOS 3.3's INT command could jam and 5D0G does strange stuff. Since this seems to have turned into a definitive explanation of Reset, let's mention a couple more things. If you ignore my advice and try to change the soft-entry vector (instead of leaving it pointing to the DOS warmstart routines), CALL 64367 (SFBGF) after the change to correct the power-up byte. If youdon't, the disk will boot when you press Reset. On the other hand, if that’s what you want to happen, just poke a zero into the power-up byte (POKE 1012.0).