Programmable Calculators: The First Generation

[Matt Agajanian]

Hi there.

Although they were the only games in town, SR-52 and HP-65, both were handheld geniuses for with their own powerful functionalities.

Was it equitable to compare these two when one was practically over 2-1/2 times more capable?

[Paul Dale]

The 65 was introduced over a year and a half earlier.


Pauli

[Matt Agajanian]

Yeah. I'm aware. And yes,

18 months is long enough to develop and thoroughly test a better mousetrap,

TI is a memory and circuit developer.

[SlideRule]

(05-19-2017 02:45 AM)Matt Agajanian Wrote:  Although they were the only games in town, SR-52 and HP-65, both were handheld geniuses for with their own powerful functionalities.
Was it equitable to compare these two when one was practically over 2-1/2 times more capable?

The marketing gurus certainly did {especially from the TI perspective}! I'm only mildly curious with the assertion the 52 was "over 2-1/2 times more capable" than the 65. In what manner was the 52 More Capable than the 65 with respect to functionality?
For reference, the HP-45 might be considered as less capable than say the 55 or the 65 with respect to programmability: faster for recursive algorithms, certainly: but XXX% more capable, hardly. I sincerely doubt (and mildly challenge) the assertion the 52 was 2½ times more capable than the 65 with respect to functionality.
Kindly request a modest elaboration.

BEST!
SlideRule

[Gene]

Not to retread this topic too much (which has been discussed a lot! - see here: SR-52 thread), here are some of the areas where the SR-52 was ahead. If I have to program something on an HP-65 or SR-52, I will pick up the 65, FWIW. I greatly prefer the HP 65 to the SR-52 in nearly every circumstance, but...

SR-52 vs. HP-65

22 user addressable memories vs. 9 (1)
224 program steps vs. 100
Program line numbers vs. no.
Indirect addressing vs. no
Some merged prefixes vs. nearly none (2)


(1) not all 9 of the HP 65 data memories are always available. SR-52 are 0-19, 98 and 99. Also in non-programming mode, STO 70-97 and RCL 70-97 work perfectly.

(2) 2nd function is merged. F and F-1 and G are not on the HP 65. Yes, I'm aware that STO 22 takes THREE steps on the SR-52, but the HP 65 hurts here too IMO.

[Matt Agajanian]

SlideRule and MoHPC members, to be concise, my choice of words & figures to describe the 52's enhancements was somewhat meant as numerical comparisons (224 vs 100, nine registers vs 20). Sorry if this context upsets folks.

[SlideRule]

(05-19-2017 04:08 PM)Matt Agajanian Wrote:  ... the 52's enhancements was somewhat meant as numerical comparisons (224 vs 100, nine registers vs 20)...

Matt

Thank you for your response and the included elaboration. I NEITHER want NOR intend to INHIBIT honest discussion / comparison of calculator features. Your clarified position is reasonable and verifiable. The SR-52 certainly offered more program and data memory than the HP-65: no argument. I am very interested in the various meta-analysis with respect to measurable correlations, as efficiency, utility etc of handhelds and their respective functions. If you have insight and quantifiable observations, please continue your dialog and share the same with all. I have tendered similar arguments in the past with varying degrees of concurrence and rebuttal. I look forward to the opinions and insights of the members, including yours. Again, thank you for your clarification to my initial response.

BEST!
SlideRule

ps: I occasionally tend to unintended harshness

[Dieter]

(05-19-2017 04:31 PM)SlideRule Wrote:  The SR-52 certainly offered more program and data memory than the HP-65: no argument.

Sure. That's what I would expect from a 1970's calculator that follows more than a year after another. It's just the usual, natural evolution of technical features and capabilities. Simply take a look at the next year and include the HP67:

1974: HP65 has 100 partially merged steps
1975: SR52 has 224 unmerged steps (only prefix merged)
1976: HP67 has 224 fully merged steps (equivalent to approx. 350...400 unmerged steps)

1974: HP65 has 9 data registers
1975: SR52 has 20+2 data registers
1976: HP67 has 26 data registers

1974: HP65 has card reader for recording programs
1975: SR52 has card reader for recording programs and data
1976: HP67 has intelligent card reader that records programs as well as data, automatically reads and writes status information (display, flags, angle mode) and allows merging of different programs.

...etc., etc., etc.

It's just the usual technical evolution. You can't directly compare an HP65 and an SR52, just as you can't do so with an SR52 vs. an HP67.

Dieter

[Don Shepherd]

(05-19-2017 08:17 PM)Dieter Wrote:  

(05-19-2017 04:31 PM)SlideRule Wrote:  The SR-52 certainly offered more program and data memory than the HP-65: no argument.

Sure. That's what I would expect from a 1970's calculator that follows more than a year after another. It's just the usual, natural evolution of technical features and capabilities. Simply take a look at the next year and include the HP67:

1974: HP65 has 100 partially merged steps
1975: SR52 has 224 unmerged steps (only prefix merged)
1976: HP67 has 224 fully merged steps (equivalent to approx. 350...400 unmerged steps)

1974: HP65 has 9 data registers
1975: SR52 has 20+2 data registers
1976: HP67 has 26 data registers

1974: HP65 has card reader for recording programs
1975: SR52 has card reader for recording programs and data
1976: HP67 has intelligent card reader that records programs as well as data, automatically reads and writes status information (display, flags, angle mode) and allows merging of different programs.

...etc., etc., etc.

It's just the usual technical evolution. You can't directly compare an HP65 and an SR52, just as you can't do so with an SR52 vs. an HP67.

Dieter

Excellent points, Dieter, thanks.

[Gene]

(05-19-2017 08:17 PM)Dieter Wrote:  1975: SR52 has 20+2 data registers

1975: SR52 has card reader for recording programs and data

Dieter

Gene: Two tiny points and clarifications.

1) The SR-52 had 20 data registers (00 - 19) and it was publicized that registers 98 and 99 were also available for for regular data use, making up the 22 here. 98 and 99 had the benefit of not being cleared by the CMs command.

In addition, registers 70-97 were also available during RUN mode for STO and RCL, giving 28 more registers.

So if you are operating in RUN mode, you really have 50 data memories available.

Registers 60-69 are used for the AOS stack and under certain circumstances can be used for STO and RCL as well. This means you **could** operate in RUN mode with 60 data memories.

2) The SR-52 card reader cannot record data registers per se.

The information in point 1 above points out the trick that was used to save data registers to a program card.

Using register 98 or 99, a program loop is created that does a RCL for memories 00 - 19 and then does a STO into register 78-97. The program should fit into 64 steps (registers 70-77 are mapped into steps 00 - 63 at 8 steps per register), then you can then write a card after running the program. This card would contain the program and the rest of program memory, which contains the data registers copied into the end of program memory. The write program to a card instruction copies registers 70-97 to the card.

It turns out that the program to loop and copy data registers into program memory was **small** enough that the reverse routine ALSO fit into the program memory space 00 - 63, allowing ONE card that could save or restore data registers to / from a program card.

Very neat trick for its day.

[Matt Agajanian]

Just as the 41 has its hidden talents of synthetic programming, it's cool to see how these 'under the radar' tricks would inspire some genius programming that our calculator clubs brainstormed.

[Dieter]

(05-22-2017 02:26 PM)Gene Wrote:  2) The SR-52 card reader cannot record data registers per se.

The information in point 1 above points out the trick that was used to save data registers to a program card.

Gene, thank you very much for the detailled explanation. I was quite sure the SR-52 did not have a dedicated read/write data feature, but then I read a 1976 TI flyer that says it can do so: "And you can record up to 28 data registers onto blank magnetic cards. Read them back in later." Obviously this claim requires some fine print with the details which have obviously been, err... "forgotten" those days. #-)  BTW, was this way of writing/reading data officially documented in detail in the SR-52 manual?

Anyway, this sets the HP67's card reader even more apart from the SR-52's from the previous year. So the technical progress that could be observed from one generation to the next becomes even more clear.

Dieter

[Gene]

TI gave full acknowledgement and press to those registers 98 and 99 in press and in materials such as the TI PPX Newsletter.

I don't know about the manual (most of mine do not have that info but later revisions might).

For 28 registers, I presume that would require manual RCL and STO into 70, 71, 72, up to 97 and then writing a card.

I believe under **program control** using R98 and R99, all 20 regular data registers were able to be read / written from a card.

Very neat.

[Matt Agajanian]

Speaking of SR-52's extra registers, here's a 52 Notes on the subject.

http://www.claudiolarini.altervista.org/pdf/52NOTES.pdf

**ADDENDUM

The link seems to be all 219 pages of 52 Notes archives.

[Gene]

Matt, that is a great find.

Please hunt around as I have been doing to see if that site has any other great PDF resources. I've wanted a searchable 52Notes for years!

[Matt Agajanian]

You are welcome, Gene. I'm on the lookout.

[matthiaspaul]

(05-27-2017 08:08 PM)Matt Agajanian Wrote:  http://www.claudiolarini.altervista.org/pdf/52NOTES.pdf

A possibly more permanent link to the individual issues:

https://www.rskey.org/CMS/index.php/the-library/14

Greetings,

Matthias

[Matt Agajanian]

See, this is why you're all the experts! Thanks for expanding my initial view! Brilliant!