Programmable Calculators: TI or HP?
Back to main page If this is useful to you, Email me at:
As cheap and common as calculators are today, it's hard for some people to imagine that they haven't always been that way. Back in the early days, individuals needing high-end scientific or programmable calculators often had to shell out a fortune to buy the machine they needed. For example, back in 1974, Hewlett Packard introduced the first programmable handheld calculator (the HP-65) at a price of $795. That's the equivalent of over $2500 today! Texas Instruments introduced their first programmable handheld (the SR-52) in 1975 at a relatively inexpensive $295!

Who bought these machines? People who realized that they could have a computer in their hands. Many were tired of waiting all day to do some numerical computation on the school or company mainframe in the form of a large stack of punched cards. By having one of these programmables, they could often write programs of their own to do the same task previously tackled by the mainframe. With this little machine sitting on their desk, they could run many different calculations in much less time. These two calculators gave users the ability to store programs on tiny little magnetic cards that could be reused. Indeed, it is quite thrilling to see on of these machines still read a program card over 20 years later. This was truly a personal "computer."

Richard Nelson was an early user of HP calculators. Early on, he realized the benefit of sharing predetermined sequences for solving common problems. For example, if there were a way to solve a quadratic equation using only 10 keystrokes on the calculator, that sequence should be shared so that someone isn't using a 15 keystroke procedure and wasting alot of time, especially if they had many to solve. He formed the world's first user group, eventually named PPC, in the summer of 1974. Richard hoped to support and spread information about using the HP-65 and making the most out of it's 9 data memories and 100 program steps. Richard began publishing HP-65 Notes (later changed to the PPC Journal) 10 times a year to a small but growing group of users. Each issue contained programs, games, tips, routines and "bug" reports that could be helpful to the members. Over the next 10 years, this group would grow to over 10,000 users in many countries around the world. Indeed, Richard is still a "father" to many current HP users, although in some respects, that role is now played by Wlodek over in England.

When TI introduced the SR-52, many members of the HP group bought it to learn about it's different capabilities. It appeared that TI had leapfrogged HP, since their programmable had 20 data memories and 224 program steps as well as new features like indirect addressing (kinda like using subscripts with arrays) that allowed for new problems to be tackled.

This "leapfrogging" would continue for a while, with HP introducing the HP-67/97 in 1976 with 224 steps of merged memory (effectively about twice that of the SR-52's "unmerged" 224 steps). The HP-67 was the handheld version while the HP-97 had a built-in printer. Programs written for one machine could be read and used on the other. The HP-67 was eventually used on the Apollo/Soyuz mission for in-flight calculations.

TI countered HP's move by introducing the TI-59 in 1977 with almost 1000 program steps and the ability to use plug-in ROM modules containing an additional 5000 program steps! The TI-58 was it's little brother, having about half the program steps and memory, but no card reader, although a version was later introduced that had "Constant" memory, retaining it's program/data memory after being turned off (a feature HP had introduced with the HP-25C, but that's another story!).

Within a short time of the SR-52's introduction, a second user group was founded to support the TI programmables. This group had several publications over the years. The last publication of that user group was TI PPC Notes, edited by Palmer Hanson, an IACC member! (If you own a TI-58/59 and want more information about them, contact Palmer as he still has copies of some of these publications). Palmer had succeeded Richard Vanderberg and Maurice Swinnen as previous editors. The group found many interesting tidbits for the TI machines.

For example, the SR-52 has two extra memories (Try storing a number into memories 98 and 99). If you don't need program mode on the SR-52, you can store numbers into memories 70-92. Amazingly, that meant that this 1975 calculator had 54 data memories available, a huge amount for that day! Users found ways of using these commands to allow storage of data memory values on magnetic cards, a function not available "off the shelf" for this calculator. Fun and games were also had with a "fractured" digits quirk on the SR-52.

A fast mode was found for the TI-58/59 that sped up program execution between 10-50% as well as a high resolution graphics mode for the TI-58/59 and its printer. These "modes" were entered by putting a "hex" code as an instruction into program memory. This quirk was found by hours of experimentation. Once well documented, however, they were easy to use in any machine. These were an early form of the "synthetic" programming that was popular on the later HP-41.

The TI-58/59 also had another advantage over the HP-41 for several years. Usually, to access a subroutine in a plug-in module, you accessed it in your program by using the key sequence: 2nd Pgm 15 A, which would call the subroutine at Label A in program 15 of the current plug-in module. However, since subroutines on these machines could be accessed by using an actual step address in normal program mode, users rightly reasoned that this should work on subroutines used from plug-in modules. So, they discovered that any sequence of steps in a plug-in module could be accessed as 2nd Pgm 02 Subr 0239, which would begin execution at step 239 of program 02 of the master library module. If one analyzed the contents of the plug in module carefully, you could find many places to use as subroutines to save steps.

Users even found interesting quirks in the lower end TI-57 (a 50 step programmable introduced along the with TI-58/59 in 1977). These included the ability to put the unit into such a low-power consumption mode that it might as well have had constant memory! Users also found ways to display the digits a - f on this model.

User loyalties began to be formed to one or the other manufacturer. You tended to become either an HP or TI person. Each manufacturer had different strengths. TI's models tended to have more memory for data and programs, calculated using more digits for greater accuracy, and cost less. This price difference was very important. It meant that even students might be able to afford one. HP's models tended to be more reliable physically and electronically, whereas the TI models had exposed electronics showing inside the battery compartment!

In 1979, HP introduced the HP-41 series of calculators, the first to handle alphabetic characters as well as numbers, in the display. The TI-58/59 had been able to print out alphabetic characters but could not display them. The HP-41 stayed on sale for over 10 years! HP gave the HP-41 four plug-in ROM slots capable of holding 8000 program steps each, to leapfrog the TI-58/59's one slot. Users could plug in a card reader, a thermal printer, extra program memory and eventually even cassette and disk drives!

Apparently, this caught TI off-guard, without an upgraded TI-59 in the pipeline. A couple of years later, TI announced a similar machine (the TI-88) but decided not to introduce it for some reason. I even recall having a press release for this machine mailed to my house.

In the spirit of things, "friendly" competitions were designed. Programs were written on both machines by the user groups to see which one could run the fastest. Programs such as a "Speedy Prime Factor Finder", "5x5 Matrix Determinant and Inverse", and a "One Year Calendar" were refined, optimized and published. The HP group even published a programs to compute the first 1,000 digits of PI in 11 hours and 2500 digits of "e"! Later programs published for the TI-59 computed 1287 digits of PI, filling 99 of the 100 available data registers with full 13 digit numbers. This amazing program ran for over 20 days! Often one group would pass the other only to be left behind a month or two later. Efficiency paid benefits with reduced program run time as well as taking up less space of that ever scare commodity: memory!

Palmer Hanson was in fact one of the players in the "Great Calendar Race". Using the TI-59 and it's printer, he had battled the HP group to print a one-year calendar in as short a time as possible. Over several years, each group created successively faster programs. Palmer had created a program that printed one year in about a minute and a half. Palmer's program took advantage of the TI-59 fast mode.

The HP group responded, however, using the HP-41's much larger memory capacity and alpha handling abilities to crank a year out in one minute and 14 seconds, just 1 second more than merely advancing the printer paper that many rows would have taken. Truly, the program was as speed optimized as it was ever going to get.

Each side, of course, benefited from these "contests" by developing better programming habits and learning more about their respective machines.

PPC members discovered how to create "synthetic" combinations of instructions on the HP-41 that functioned in ways not foreseen by HP to gain even greater control over their machines. This included a number of extra tones, the ability to finally call any portion of a plug-in module as a subroutine, and access to a number of additional display characters. PPC even developed their own custom ROM cartridge to be plugged into the HP-41 calculators. Copies of this ROM (and it's 500+ page manual!) are still sold by Educalc in California, where Richard Nelson presently works. If only manufacturers wrote documentation that extensive!

Which brand was best? That's still a loaded question. I started off as a "TI" person, due to their lower cost and greater memory capacity. I became a traitor (or saw the light) when HP introduced the HP-41. Today, the higher priced HP's still seem to work more often than the less expensive TI's do when found in garage sales or at thrift shops. The TI-58/59's seem especially prone to mechanical and electronic oddities. Most card readers from both TI and HP have quit functioning unless repaired. My HP-65, however, has been repaired and still reads and writes cards well today!

TI, by having lower priced machines, put programming into the hands of a larger audience. HP continued to develop programmable calculators well after TI had largely abandoned the market, culminating in the first handheld machine capable of performing symbolic calculus.

However, there was no "best" brand. Each played to different strengths. Collectors have cast their votes, since HP's are much more valuable on the market today than the TI models.

The competition between the manufacturers actually helped everyone by bringing out more advanced models to compete against the enemy.

This was the cast of characters:

HP-65		7/74, 100 program steps, 9 data memories, card reader
SR-52		1/75, 224 program steps, 20 (22) data memories, card reader, indirect addressing
HP-67/97	7/76, 224 merged program steps, 26 data memories, card reader, printer(97)
TI-57		1/77, 50 program steps, 8 memories, TI's first "fully" merged step programmable
TI-58		1/77, 480 program steps or 60 data memories, ROM plug in slot, alpha printing
TI-58C		1/78, 480 program steps or 60 data memories, ROM plug in slot, constant memory, alpha printing
TI-59		1/77, 960 program steps or 100 data memories, ROM plug in slot, card reader, alpha printing
HP-41		7/79, up to 2200 program steps or 319 data memories, 4 ROM plug in slots, alpha display characters, sound signals

Visitors since 5/15/97