The HP-33E/C that could have been.

[Matt Agajanian]

Given TI’s SR-56 was a direct competitor to the HP-25, it came complete with 100 program steps. So, wouldn’t it have been more relevant to create an enhanced version of the 25C with 100 program steps instead of 49? And yes, I realize the 29C was the equivalent of the SR-56. Although, as its purpose was to be an economical programmable upgrade from the 25, wouldn't it have been much more significant for the 33 to have 98 or 99 program steps?

But was a model with over 100 program steps what the 34C was for?

[brouhaha]

(01-06-2023 04:59 AM)Matt Agajanian Wrote:  as its purpose was to be an economical programmable upgrade from the 25, wouldn't it have been much more significant for the 33 to have 98 or 99 program steps?

Both vendors made calculators to span a range of customer requirements and price points. While there are a few cases where one can infer that a model from one vendor was significantly inspired by a model from the other, much more of the time they were just trying to offer roughly similar capabilities over a range of models.

Bear in mind that a new calculator model often took 24 months or more from concept to mass production. If one vendor introduced a model in 1976, a model introduced by the other vendor in 1978 wasn't necessarily a specific response to that.

[lvt]

I remember when I bought my first pocket PC in the 90 (a Psion), the 256KB version is significantly more expensive than the 128KB version, or you can buy a memory card.

Maybe consumers faced the same problem in the 70's. A calculator with larger memory (higher registers count) would cost more to produce and buyers have to pay more than what they actually need.

[Garth Wilson]

Couldn't the HP do the same thing with about half the number of steps?  STO 1 and RCL 1 for example each took two steps on the TI, but were merged into a single one on the HP.  I was met with many of this kind of thing when I went from the TI-59 to the HP-41.

[Kostas Kritsilas]

Its not a 2:1 difference in all instances. The 2nd and INV keys on the TIs were generally NOT separate keystrokes, or program steps, if you want to think of it that way. There is not doubt that the equivalent programs used fewer program steps (on HPs with merged program steps), but the relationship is a lot more complex than assuming a straight 2:1 or 3:1 relationship; it is very dependent on what the program actually did, and how it was programmed.

[Matt Agajanian]

For the TI-57, even though there were merged operational (memory, unconditional branches), I would suspect that the AOS logic would still grab up more keystrokes than RPN. So, wouldn’t it have been much more resourceful with 100 steps (or more) to at least give it programs least some sort of equivalent in comparison to RPN’s logic?

[brouhaha]

(01-07-2023 06:03 AM)Matt Agajanian Wrote:  For the TI-57, even though there were merged operational (memory, unconditional branches), I would suspect that the AOS logic would still grab up more keystrokes than RPN. So, wouldn’t it have been much more resourceful with 100 steps (or more) to at least give it programs least some sort of equivalent in comparison to RPN’s logic?

Sure. Also, Ford should have put a 462 V8 and four wheel drive in the 1977 Pinto, to enable it to compete more effectively against the Chevy Suburban.

[Kostas Kritsilas]

Going strictly from memory of those days, I believe that the contemporary calculator for the HP-25/25C was the TI-58/58C. and the TI-58/58C had 480 partially merged program steps, maximum (I seem to recall that you could re-allocate program steps to registers). The HP-67 (224 fully merged steps) equivalent was the TI-59 (960 partially merged steps). The later HP-29C had 98 or 100 fully merged steps, and a bunch of added programming features over the HP-25/25C like indirect addressing and subroutines, some of which were in the TI-58/58C already.

The earlier generation programmables were the HP-55 (49 un-merged steps) and HP-65 (100 un-merged(?) steps) vs. the SR-56 (100 un-merged steps) and SR-52 (I don't remember how many steps it had). I also don't recall the register counts for each.

[Matt Agajanian]

(01-07-2023 09:19 AM)Kostas Kritsilas Wrote:  Going strictly from memory of those days, I believe that the contemporary calculator for the HP-25/25C was the TI-58/58C. and the TI-58/58C had 480 partially merged program steps, maximum (I seem to recall that you could re-allocate program steps to registers). The HP-67 (224 fully merged steps) equivalent was the TI-59 (960 partially merged steps). The later HP-29C had 98 or 100 fully merged steps, and a bunch of added programming features over the HP-25/25C like indirect addressing and subroutines, some of which were in the TI-58/58C already.

The earlier generation programmables were the HP-55 (49 un-merged steps) and HP-65 (100 un-merged(?) steps) vs. the SR-56 (100 un-merged steps) and SR-52 (I don't remember how many steps it had). I also don't recall the register counts for each.

The SR-52 had 20 registers with 224 partially merged program steps. Only the operations which needed the 2nd key were merged. For example 2nd if zero. Memory and branching locations weren't not merged. For example, GTO 145 took four program steps- GTO 1 4 5.

[Matt Agajanian]

(01-07-2023 09:19 AM)Kostas Kritsilas Wrote:  Going strictly from memory of those days, I believe that the contemporary calculator for the HP-25/25C was the TI-58/58C. and the TI-58/58C had 480 partially merged program steps, maximum (I seem to recall that you could re-allocate program steps to registers). The HP-67 (224 fully merged steps) equivalent was the TI-59 (960 partially merged steps). The later HP-29C had 98 or 100 fully merged steps, and a bunch of added programming features over the HP-25/25C like indirect addressing and subroutines, some of which were in the TI-58/58C already.

The earlier generation programmables were the HP-55 (49 un-merged steps) and HP-65 (100 un-merged(?) steps) vs. the SR-56 (100 un-merged steps) and SR-52 (I don't remember how many steps it had). I also don't recall the register counts for each.

The SR-52 had 20 registers with 224 partially merged program steps. Only the operations which needed the 2nd key were merged. For example 2nd if zero. Memory and branching locations weren't not merged. For example, GTO 145 took four program steps- GTO 1 4 5.

The SR-56 had the same characteristics. Memory was 10 registers with 100 program steps. 2nd functions were merged, but not the memory nor transfer addresses.

[Steve Simpkin]

(01-07-2023 09:19 AM)Kostas Kritsilas Wrote:  Going strictly from memory of those days, I believe that the contemporary calculator for the HP-25/25C was the TI-58/58C. and the TI-58/58C had 480 partially merged program steps, maximum (I seem to recall that you could re-allocate program steps to registers). The HP-67 (224 fully merged steps) equivalent was the TI-59 (960 partially merged steps). The later HP-29C had 98 or 100 fully merged steps, and a bunch of added programming features over the HP-25/25C like indirect addressing and subroutines, some of which were in the TI-58/58C already.

The earlier generation programmables were the HP-55 (49 un-merged steps) and HP-65 (100 un-merged(?) steps) vs. the SR-56 (100 un-merged steps) and SR-52 (I don't remember how many steps it had). I also don't recall the register counts for each.

It is interesting to look at the U.S. street prices for these TI models when they were introduced compared to HP models at the time.

When the TI-57/58/59 were introduced around May 1977, the TI-58 was the same price ($100) as a HP-25. The TI-58 was much cheaper than the HP-29C ($159) which was introduced around the same time. The following advertisement is from the October 1977 issue of Popular Electronics magazine.
   


Just a few months earlier (June 1977) the following ad from the same magazine shows the SR-56 selling for considerably less ($79) than the HP-25 ($116) at that time.
   

The TI-58C appears to have been introduced sometime in 1979, the same year as the HP-34C and HP-33C. Interestingly TI did not create a CMOS version of the memory chip used in the TI-58. Instead they bought a standard CMOS RAM chip (RCA) and developed a custom chip to interface to it. The result is a substantial redesign of the PCB.
http://www.datamath.org/Sci/WEDGE/TI-58C.htm