What are vector tables? You remember that a vector is a pair of memory locations that hold the address of a routine. A vector table is, quite simply, a table of vectors. Thus, locations 58368 through 58533 hold the addresses of various routines, mostly having to do with I/O or interrupts.

58368-58383         E400-E40F

This is the vector table for the screen editor handler. For a description of its contents, along with the contents of the next four vector tables, see HATABS at locations 794 through 831 (where we called it a "handler address table").

58384-58399         E410-E41F

The vector table for the display handler. See the note at EDITRV.

58400-58415         E420-E42F

The vector table for the keyboard handler. See the note at EDITRV.

58416-58431         E430-E43F

The vector table for the printer handler. See the note at EDITRV.

58432-58447         E440-E44F

The vector table for the cassette handler. See the note at EDITRV.

You will notice that the following 16 vectors are three bytes long rather than two. Why the extra byte? The first byte of each vector is a 6502 JMP instruction, while the address is in the second two bytes.

The purpose of these vectors may not be obvious to you (they weren't to me). Atari knew that it would probably need to make changes to the OS at some point. It also wanted to make sure that old programs would still be able to work with these newer versions of the OS, even though some of the addresses would be different. The solution was to use vectors. That way, even though the addresses in the vectors would change, the addresses of the vectors would remain the same, and programs using these addresses would still work. The reason that some programs don't work with version "B" of the OS is that these programs didn't use the vectors.

58448-58450         E450-E452

DISKIV is the initialization vector for the disk handler. It points to location 60906.

58451-58453         E453-E455

This is the entry vector for the disk handler. It points to location 60912.

48454-48456         E456-E458

CIOV is the entry vector for CIO (Central Input/Output). See Appendix Seven on I/O for an explanation of what CIO does.

You can use CIO yourself by first setting up an IOCB (see locations 832 through 959), and then using the following routine:

100 DIM ML$(7)
110 GOSUB 10000
130 END
10000 FOR BYT=1 TO 7
10030 NEXT BYT
10040 RETURN
10050 DATA 104,104,104,170,32,86,228

The data is for this machine language routine:

68     PLA
68     PLA
68     PLA
AA     TAX
2056E4 JSR $E456

CIO expects the number of the IOCB you want to use, times 16, in the X register. That's why we have IOCB* 16 in the preceding program. You should substitute the IOCB number you are using for IOCB. Remember to OPEN the IOCB first.

CIOV points to 58564

58457-58459         E459-E45B

This is the entry vector for SIO (Serial Input/Output). Again, check Appendix Seven on I/O for an explanation of SIO's function. SIOV points to 59737.

58460-58462        E45C-E45E

SETVBV is the entry vector for a routine that serves two purposes. First of all, as we saw at VVBLKD (548,549), it will set up VVBLKI and VVBLKD for us. Second, as we saw at CDTMA1 (550,551), it will also set up the vectors for the system timers. See VVBLKD and CDTMA1 for more information.

SETVBV points to 59666 in version "A" of the OS, 59629 in version "B."

58463-58465         E45F-E461

This is the entry vector for the stage one VBLANK routine. Unless you have your own routine, VVBLKI (546,547) normally points here. See VVBLKI and VVBLKD (548,549) for more information on VBLANK.

SYSVBV points to 59345 in version "A" of the OS, 59310 in version "B."

58466-58468         E462-E464

XITVBV is the exit vector for the VBLANK routine. This is what VVBLKD points to unless you've changed it.

Use XITVBV to return to where the computer left off from when the VBLANK interrupt occurred. It points to 59710 in version "A" of the OS, 59653 in version "B."

The following four vectors are designed for use by the OS only.

58469-58471         E465-E467

This is the initialization vector for SIO.

58472-58474         E468-E46A

SENDEV is the vector for the "send-enable" routine.

58475-58477         E46B-E46D

This is the initialization vector for the interrupt handler routine.

58478-58480         E46E-E470

CIOINV is the initialization vector for CIO.

58481-58483         E471-E473


This is the entry vector for the blackboard mode, which is more commonly known as the "ATARI MEMO PAD" mode. Type "BYE" from BASIC, or turn on the computer with no cartridges or disk drives to see what I mean. This mode lets you type things on the screen without anybody caring what you type. In other words, you can press RETURN and nothing will happen. To get back to BASIC, press SYSTEM RESET (this won't erase your BASIC program).

BLKBDV points to 61987.

58484-58486         E474-E476

WARMSV is the entry vector for the warmstart routine. The OS jumps through here when SYSTEM RESET is pressed.

WARMSV points to 61723.

In case these locations don't seen useful to you, try this:

X = USR(58484)

What you have just done is told the computer to go to 58484, which contains a machine language instruction to go to the address in the next two memory locations. Since this routine is for what's called warmstart, the computer will now act just like you pressed SYSTEM RESET. You use the other locations in this section just like this. TRY IT!

58487-58489         E477-E479

This, appropriately, is the entry vector for the coldstart routine. Whereas going through WARMSV only initializes the OS RAM, going through COLDSV initializes all RAM, meaning that any programs in memory will be erased. See COLDST at location 580 for a way to hook COLDSV up to SYSTEM RESET rather than WARMSV.

COLDSV points to 61733

The following two vectors are designed for use by the OS only.

58490-58492         E47A-E47C

RBLOKV is the entry vector for the cassette "read-block" routine.

58493-58495         E47D-E47F

This is the vector for the cassette "OPEN-for-input" routine.

58496-58533         E480-E4A5

VCTABL is a table of the initial values for the OS RAM vectors.

Now we're into the final part of the OS, which consists mostly of the various built-in handlers, interrupt routines, and so forth. What follows is a list of addresses for some of these routines, which can be useful to you in one of several ways. If you're a beginner, the list will provide you with an idea of exactly what the OS does. If you're a machine language programmer then, along with the OS listing, the list will help you find the various routines so that you can see exactly how things are done. By studying the routines, you can also pick up on programming techniques (don't be afraid of the OS listing; it's really not that difficult to understand). Finally, if you really know what you're doing, you can rewrite the routines and put them in your own programs, customizing them to your own needs.

Most of the routines will not work without some kind of previous setup, so make sure you check the OS listing before you attempt to use them.


Please note that all the following addresses are for the original OS only. Some of them may be different in the newer versions. At the time of this writing, however, the OS listing is for the original version, and that is why these addresses are used.

Return to Table of Contents | Previous Chapter | Next Chapter