TI versus HP



#31

Hello:

I used a TI59 (and still do) back at Yale and used to go head-to-head with the HP41 guys all the time. It seemed the TI trumped the HP every time on quality of build, power, convenience and even panache.

Can you tell me what the attraction of the HP line is? It just seems that TI beat them every time esp. in the high end handhelds.

Dirk in Beverly Hills


#32

http://www.ti59.com/

#33

I can't speak for the high end calculators, but I will talk about the non-graphing scientifics. The reason that these are so superior to TI is because of RPN. RPN is vastly superior to the various forms of algebraic entry. RPN is more powerful, quicker, easier, and therefore more reliable.

Before switching to HP, every TI that I ever had lasted at best a couple of years. The numbers wore off the keys, the keys had a lousy tactile response, etc. My first HP is now 19 years old, and still works perfectly. The numbers will never wear off the keys because the numbers are molded in. The tactile response is as good today as when I first purchased it. It's ability to perform long chains of calculations without having to resort to parenthesis or manually storing temporary results is unmatched. The HP is a serious tool in the hands of an engineer, while the TI is more of a toy or a machine for grade school students. In college, it was always the HP that was highly prized by the engineers, not the TI.

Ultimately, this comes down to a matter of opinion as to which is superior. Those who contend that TI is better are welcome to their opinon, but those who frequent this forum are partisans for HP, especially the older ones rather than the newer debased offerings. Are you looking for a dialogue, or just trying to start a flame war?


#34

Why even ask if someone is trying to start a flame war? That question in itself is inflammatory. Just answer the guy's question and if he responds with a reasonable answer, or not at all, you have the answer to your question.

Quote:
I can't speak for the high end calculators, but I will talk about the non-graphing scientifics. The reason that these are so superior to TI is because of RPN. RPN is vastly superior to the various forms of algebraic entry. RPN is more powerful, quicker, easier, and therefore more reliable.

Before switching to HP, every TI that I ever had lasted at best a couple of years. The numbers wore off the keys, the keys had a lousy tactile response, etc. My first HP is now 19 years old, and still works perfectly. The numbers will never wear off the keys because the numbers are molded in. The tactile response is as good today as when I first purchased it. It's ability to perform long chains of calculations without having to resort to parenthesis or manually storing temporary results is unmatched. The HP is a serious tool in the hands of an engineer, while the TI is more of a toy or a machine for grade school students. In college, it was always the HP that was highly prized by the engineers, not the TI.

Ultimately, this comes down to a matter of opinion as to which is superior. Those who contend that TI is better are welcome to their opinon, but those who frequent this forum are partisans for HP, especially the older ones rather than the newer debased offerings. Are you looking for a dialogue, or just trying to start a flame war?



#35

No, I think it was flame bait. Why? because I own both and there is no way anyone can say with a straight face that the Ti is well built.

Ti's were the cheap calculator solution. You need a CHEAP solution, you buy Ti. The Ti-59 did have lots of horsepower, more so than any LED Hp. But the Hp's were well built and well designed for daily use.

Shortly thereafter Hp released the Hp41c and a bit after that the Hp11c. The Hp41c was so superior to the Ti-59 in EVERY respect. But even the LED Hp's had good to great keyboards, compared with poor to functional for the 1975-83 line of Ti's (nearly a decade of poor keyboards).


#36

Back in January 1981 a complimentary article "The HP-41C: A Literate Calculator" appeared in BYTE. The author did comment "... The most fundamental defect in the architecture of the HP-41C, inadequate numerical precision, is a serious flaw indeed. ..."

It still is.


#37

I agree Palmer. While the HP41c was superior in quite a few respects over the TI59, it was not superior in terms of numeric precision.

A few matrix test routines show that pretty well.

Gene

#38

Palmer says:


Back in January 1981 a complimentary article "The HP-41C: A Literate Calculator" appeared in BYTE. The author did comment "... The most fundamental defect in the architecture of the HP-41C, inadequate numerical precision, is a serious flaw indeed. ..."

It still is.

I dug out this old article and re-read it. I found the paragraph where the author makes the quoted comment. He makes another comment that betrays a common misunderstanding about the behavior of calculators.

Imagine that we type the following 10-digit number into an HP41:

1.414213562

Now press the x^2 key. The exact result is a 19 digit number:

1.999999998944727844

But the HP41 can only store and display a 10-digit mantissa, so we would hope it would display the correctly rounded 10-digit result:

1.999999999

which it does. Isn't this what we would want our 10-digit calculator to do? When given a number whose exact square comprises more than 10 digits, wouldn't we want the calculator to return a result correctly rounded to 10 digits?

Now suppose that the number we started with, 1.414213562, wasn't something we typed in, but resulted from typing in the integer 2 and then pressing the SQRT key. This number is, in fact, the correctly rounded 10-digit version of the square root of 2. When we subsequently square it, should we expect to get back exactly 2? If we did, the HP41 would have made an error in its multiplication followed by rounding, even though it might have made us feel better to get a result of exactly 2.

Should the calculator have remembered whether it got the original 1.414213562 as a result of taking the square root of 2, or if it was just typed in? This would be the only way it could return exactly 2 if it was starting from a calculated square root of two, but 1.999999999 if the starting number had been typed in. The calculator would have to have some way of distinguishing the two cases.

We don't want the calculator to return a result in error just to make us feel good, do we? If a 10-digit calculator returns a result of exactly 2 when starting with 2 and pressing SQRT and then x^2, it has made an error. The only way it can happen without an error having been made is if it's really a calculator with more than 10 digits for calculation results, but only displays 10 digits. Or, it is really a 10 digit calculator that has had its display set to display fewer than 10 digits, which the HP41 can do, of course.

One of the other criticisms of the HP41 the author of the Byte magazine article offers is that the HP41 returned 1.999999999 when 2 was entered and then SQRT and x^2 were pressed. This is a misguided complaint, and shows that the author didn't understand what a HP41 calculator should do (or any other 10-digit calculator with a full 10-digit display and no hidden guard digits).

He also says that the HP41 doesn't have any guard digits. The HP41 does in fact have 3 guard digits. Its internal calculations are done with 13 digits and the result rounded to 10 digits. That's how the HP41 guarantees that the result of +, -, / and * are correct to 10 digits. Its just that these extra guard digits are not accessible once the calculation is done, unlike the TI machines, for example.

He complains that after some calculations, the HP41 might lose as many as 2 or 3 digits in its result. He doesn't give a specific example of what he means, so its not possible to give a comment to a particular result. But, it's well known that catastrophic cancellation can occur with subtraction; ANY calculator is subject to this phenomenon, and more digits can certainly help this problem.

As to whether it's a "serious flaw" for a calculator to have only 10 digits, that topic has been beat to death here and elsewhere. 10 digits is WAYYYY more that the slide rules that preceeded it, and more than even 7-place log tables which were the thing you used if you needed a result with lots of significant digits in the days of slide rules.

But, some of this author's comments are just not correct.


#39

Rodger wrote: "He also says that the HP41 doesn't have any guard digits. The HP41 does in fact have 3 guard digits. Its internal calculations are done with 13 digits and the result rounded to 10 digits. That's how the HP41 guarantees that the result of +, -, / and * are correct to 10 digits. Its just that these extra guard digits are not accessible once the calculation is done, unlike the TI machines, for example"

Gene: And that's probably the rub. On the TI, you have the extra digits hanging around with the number displayed to 10 digits, while on the HP41, those digits are used internally but the result is rounded to 10 digits and those guard digits disappear, which can lead to the results shown when a displayed answer is used as the input for a sequential calculation.

The argument raged for quite some time, IIRC, about the advantages of 10 good rounded digits vs. 13 digits displayed as 10.

IMO, I think the TI method w/guard digits tends to more accurately present a 10 digit result than the HP41 method. The TI method almost always took longer for the result to deteriorate than the HP41 method.

Gene


#40

If you really need to worry (for, say, science or engineering in the real world) about whether your 10 digits are correct or not, you should think a lot about how those 10 digits were derived!

What real-world cases need this accuracy?

ABout the only thing I can think of are timing and frequency, and associated measurements of length, based on timing and the speed of light.

The speed of light is defined as 299 792 458 meters per second (EXACTLY). A second is defined as some number of cycles (I don't remember it and I don't have a handy reference to look it up) produced by a certain energy level transition in a cesium atom (you even have to use the corret isotope), with comparable precision.

The best atomic time standards (hydrogen masers and other masers) have stabilities that can be measured at the parts in 1e14 or 1e15 level - so here you really can worry about the digits produced by your calculator.

The NASA program I used to help support measures the distances between parts of the world at levels of a few mm out of 10000 km, so we needed high precision calculations, too. This was all based on very precise timing measurements.


#41

The situations in which that much accuracy is needed is an entirely different matter!

Although, sometimes in monetary problems dealing with billions of $$, the rounding can add up.

I just meant to say that IF someone needed such accuracy, IMO the TI59 style guard digits seem to have preserved accuracy longer than the HP41 approach.

#42

Hi, Dave:

Dave posted:

"The speed of light is defined as 299 792 458 meters per second (EXACTLY)"

    EXACTLY ? ... Exactly the other way around:

               The meter is the length of the path travelled by light
    in vacuum during a time interval of 1/299 792 458
    of a second.
    The meter is defined from the speed of light, which is a fundamental constant, and a time unit, not the other way around.

    You don't 'define' the speed of light, it's just what it is, a universal constant in the broadest sense, and then, arbitrarily, you define the meter from it in any way which suits you best.

    In this particular case, and for obvious practical reasons, it's been defined so as to closely agree with the previous definition (the length between two marks on a platinum-iridium rod kept at some place at this or that temperature, etc).

Best regards from V.

#43

Valentin,

I'll stick with my original statement about the speed of light being "defined." (Maybe "defined" is too strong - maybe "assigned" is a better term.)

Of course, as you note, it has whatever speed it has, but we then need to assign some value to it.

I defer to the US National Institute of Standards (NIST), at which you can find (http://physics.nist.gov/cgi-bin/cuu/Value?c|search_for=universal_in!) under "Fundamental Physical Constants" that the speed of light has the value I gave and is stated as being "exact."

As you also note, the meter is then defined from that definition plus the value of the second.

Dave

PS Every time I read one of your posts, I tend to say your name mentally, but I'm not sure if the "n" at the end is pronounced or not (i.e. is it nasalized like most final French consonants?).


#44

Hi, Dave:

Dave posted:

"I'll stick with my original statement about the speed of light being "defined." (Maybe "defined" is too strong - maybe "assigned" is a better term.)"

    You can do whatever you like, but it's the meter that is defined from the speed of light, not the other way around, as you can easily google out if you care. And, of course, stating that the speed of light is "defined" as a certain value in meters when the meter itself is defined as the speed of light divided by some value, is what's called a circular reference.
"Of course, as you note, it has whatever speed it has, but we then need to assign some value to it."
    Assigning some value to a physical magnitude isn't called "defining" but "measuring". For instance, that your height is, say, 1.80 meters doesn't mean that we're defining your height to be that value, but that we've measured your height and it comes to that. Similarly, you can measure the speed of light in meter per second, but you can't define it as a certain number of meters per second.

    That the meter's definition has been chosen in such a way that it gives a neat integer value for the speed of light which comes pretty close to the former definition, is nothing but an arbitrary, convenient choice, a "definition". But once that unit (the meter) has been defined, the speed of light is just measured in terms of these previously defined units.

"I defer to the US National Institute of Standards (NIST), at which you can find (http://physics.nist.gov/cgi-bin/cuu/Value?c|search_for=universal_in!) under
"Fundamental Physical Constants" that the speed of light has the value I gave and is stated as being "exact."

    I've made the point clear enough by now. By the way, your reference doesn't mention the word "define" or "definition", did you notice ?
    As stated, that the value is exact is a consequence of the particular definition chosen for the meter.
"PS Every time I read one of your posts, I tend to say your name mentally, but I'm not sure if the "n" at the end is pronounced or not."
    Yes it is. In its original Spanish, "Valentin" is pronounced as "va (as in "bar") - len (as in "length") - TIN (as in "tinkle"), with the emphasis on the final syllable.
Best regards from V.
#45

My recollection of the article is that the reviewer didn't seem to understand that while the taking of the square root of two followed by squaring sequence yielded a 2 in the TI-59 display that did not mean that the value in the display register was exactly two. He also did not seem to understand that a similar effect can be obtained with the HP-41 by simply operating in fix 8 mode.

The reviewer's square root of two test was modified by others. In one modification one starts with 2 and repeats the square root square sequence and sees that the HP-41 alternates between 1.999999999 and 1.414213562, while on the TI-59 the sequence the display alternates between 2 and 1.414213562 but the values in the display register are gradually decreasing. Some HP folks suggest that the HP result is preferred but they fail to note how many iterations would be needed (something like 200 or so as I recall) before the TI-59 errors are as large as the HP-41 errors.

A second modification starts with a small integer in the display and takes five square roots followed by five squares. In this test the end results with the TI-59 is always smaller than the starting values (due to truncation, I think) while the end results for the HP-41 are sometimes smaller and sometimes larger (due to rounding, I think). But the the TI-59 error is always much smaller than that with the HP-41.

If you try to do this test on a Casio fx-7000 or a TI-81 which doesn't have AOS but EOS you need to use parentheses carefully or the machine will essentially remember that you did a square root followed by a square and properly get exactly two. That's sort of akin to the old high school algebra technique of waiting to evaluate until the end of the problem until you can take any cancellations which may be available.

As to what number of digits are needed in the mantissa that clearly depends on the application. I admit that some applications are working with only a few digits. In inertial navigation we were working with at least seven digits as in accelerometer measurements good to a micro-g over a range from a micro-g to ten g's, etc. Slide rules were never good enough for processing that kind of data. In the early sixties we used Friden's a lot. If my memory is right we had eight or ten digits available with the Friden's.


#46

Palmer said:
My recollection of the article is that the reviewer didn't seem to understand that while the taking of the square root of two followed by squaring sequence yielded a 2 in the TI-59 display that did not mean that the value in the display register was exactly two. He also did not seem to understand that a similar effect can be obtained with the HP-41 by simply operating in fix 8 mode.

He never mentioned TI calculators in the entire article except to mention that TI used the same size magnetic cards with some of its calculators as the HP41 used. Let me quote from page 136 of the article:

The most fundamental defect in the architecture of the HP-41C, inadequate numerical precision, is a serious flaw indeed. Numbers are represented, both internally and in the display, with 10 decimal digits; there are no guard digits. As a result, inaccuracies are quite often introduced into the least-significant digit. For example, SQRT(2)^2 is evaluated by the calculator as 1.999999999. For operations on some data, the corruption goes still deeper and 2 or 3 digits become suspect. There is something absurd about the world's fanciest calculator not being able to give results accurate to more than seven or eight decimal places.

I already pointed out that he misunderstood some things. In the next paragraph, he says;

Actually, a subsidiary problem is more serious than that. Conditional tests on data are carried out on the full 10-digit representation. Consequently, a test that effectively asks "Is SQRT(2)^2 equal to 2?" will give a false result, wihch can lead a program astray.

Didn't he realize that this sort of problem, due to finite register size, occurs with every calculator that does floating point arithmetic? It is to be expected, and the calculator user must deal with it.

The author of that article wasn't comparing HP calcs to TI calcs specifically.

---------------------------------------

Palmer says:
As to what number of digits are needed in the mantissa that clearly depends on the application. I admit that some applications are working with only a few digits. In inertial navigation we were working with at least seven digits as in accelerometer measurements good to a micro-g over a range from a micro-g to ten g's, etc. Slide rules were never good enough for processing that kind of data. In the early sixties we used Friden's a lot. If my memory is right we had eight or ten digits available with the Friden's.

In an earlier post, Ron Ross said:

The Hp41c was so superior to the Ti-59 in EVERY respect.

And you responded:

Back in January 1981 a complimentary article "The HP-41C: A Literate Calculator" appeared in BYTE. The author did comment "... The most fundamental defect in the architecture of the HP-41C, inadequate numerical precision, is a serious flaw indeed. ..."

My aim is to point out that the author of that article had a number of misunderstandings, and to suggest that the phrase "serious flaw indeed" is perhaps overstated. Are we to believe that having only 10 digits is a serious flaw, but that 3 more turn a serious flaw into a non-problem?

The nature of ill-conditioned problems is that you lose a certain number of digits, whether your calc has 10, 12, 13 or 14 digits. Any problem that loses 3 digits of accuracy on the HP-41 will almost certainly lose 3 digits on a TI calc too. Losing 3 digits out of 10, which would happen only rarely, still leaves 7 correct. That's enough for almost all practical problems. Hundreds of thousands of engineers and scientists found the 10 digits of the HP41 more than adequate.

I've said before that, all else being equal, having more digits is a good thing; sometimes, though, all else isn't equal. However, I've noticed that the recent TI calculators have gotten MUCH better with their math.

But comparing the HP41 with the TI59, or comparing the HP49 with the TI86 is like putting a welterweight in the ring with a heavyweight. It's just not fair in some respects and the outcome in those respects is just what you might expect. The TI86, which I own, can give a 14-digit square root; the HP48 can only give a 12-digit result. Just as we would expect. That's why I want to evaluate calculators on another basis--do they do what they should do? I plan to post more on this topic.


#47

I used to have a TI-58 (functionally the same as the '59, but with
only half the memory and no cardreader)
I seem to remember that SIN(45°) was NOT equal to COS(45°).
Now, that is a fundamental flaw - that HP neatly
avoided by rounding the result to 10 digits.
BTW this exact same 'error' would show on today's 49 calcs, if
they would show their results with the internally used 15 digits:
SIN(45.) = 7.0710678118656E-1
COS(45.) = 7.07106781186545E-1

Cheers, Werner


#48

Page C-1 of the Persoal Programming manual for the TI-58, TI-58C and TI-59 addresses the problem of sin 45 not equal to cos 45 and states that when doing t register comparisons precautions should be taken to prevent improper evaluation due to the guard digit differences. It notes that the displayed values are equal in FIX 9 mode and suggests the use of the EE-INV-EE sequence to leave only the displayed value for further use. That will work for the 45 degree case. It will not work for the case where the displayed value of sin 39 is 0.6293203910 and the cos 51 is 0.6293203911 . You might think that we can solve the problem by operating in FIX 8 mode but with a little work you can find that the displayed values for sin 32.9999999989 and cos 57.0000000011 will be equal in FIX 9 but not equal in FIX 8.


#49

Common programming sense says that in making equality comparisons of this type, you see if the difference is less than some tolerable amount. Otherwise the critics won't be happy with even a 1000-digit calculator.

#50

I have no problem with the idea that some applications are less demanding than others. I suppose that is why some people use Casio's, some use Sharp's, some use HP's and some use TI's. I suspect that most don't really look hard before they buy -- certainly not as hard as we looked at the computational capbility of airborne computers before selecting them for navigation and guidance applications.

Some use certain devices because of input and pressures from others whether their professors, their high school teachers, their colleagues, or whatever. These days most high school students use TI graphical calculators, not so much because they are better which they are, but almost certainly because back in the mid 80's someone at TI had an idea about the use of graphic calculators, committed to the idea, marketed the idea to the high school community and delivered the necessary support.

I expect to publish some more results on the Albillo problem and your modifications to it in the near future. The winner to date for machines in my inventory is the RREF routine on the HP-49 which is typically an order of magnitude better than the B/A method.

#51

Rodger --

Good, thoughtful post on the philosophy of "guard digits" and number of digits maintained.

The topic being discussed is that of [sqrt(2)]2 equaling 1.999999999 on the HP-41 (and other pre-Saturn calc's,) and 1.99999999999 on Saturn-based calc's.

In an archived post from last month, you also described the performance of the HP-30S for the trigonometric-based "forensic function" that helps reveal roundoff errors in algorithms. You showed some insightful evidence that results of trig calculations on the HP-30S are slightly rounded off to a very-close integer in certain cases.

I would suspect that the philosophy behind these methods are to provide reassurance to novice users that the calc is performing inverse operations correctly. We expect to see exactly 2 as the result of [sqrt(2)]2; we expect to see exactly 3 as the result of cos-1(cos(3o)). "If a result is extremely close to an integer, it probably should be that integer, in most cases." Rounding, and the use of many decimal digits, is facilitated by the use of high-precision floating-point representations in the HP-30S.

In Appendix D of the HP-16C Owner's Handbook, there are several programs for converting floating-point values between the calc's native 56-bit binary-coded decimal representation and a hexidecimal display of the IEEE 32-bit single-precision floating-point standard proposed at the time (1981-82).

According to the Wikipedia entry, it seems that the proposed format was adopted as ANSI/IEEE Std 754-1985 (IEC 60559:1989) without further modification.

The format allows integers up to +/- 16,777,216 (224) to be represented exactly. This is verifiable by running the HP-16C programs (or the converter programs linked in the Wikipedia entry). The IEEE 64-bit double-precision floating-point format can exactly represent integers up to +/- 253 (~= 9.00719925474 x 1015).

Regards,

-- KS


Edited: 26 Feb 2006, 11:48 p.m.

#52

[quote]
Why even ask if someone is trying to start a flame war? That question in itself is inflammatory. Just answer the guy's question and if he responds with a reasonable answer, or not at all, you have the answer to your question.

Thanks for your suggestion. I'll be sure to take it under advisement.

#53

Who's that walking on my bridge?

#54

I have both. I started with the TI-58c because it gave the most for my $100 budget at the time. Next I stepped up to the TI-59 along with the printer and additional modules. I got a lot of use out of those before getting my HP-41cx. The TI-59 did have more digits, and I understand the original HP-41's had quite an inaccuracy in a couple of the trig functions. Those inaccuracies got corrected early on though. To say the TI-59 was better than the HP-41 is like saying that the DC-3 airplane is better than a modern Airbus because the DC-3 gave window seats to a higher percentage of passengers. If you really want a list of things the HP-41 can do that the TI can't, we can put that together for you, but it will be a very long list. There are also issues of speed, battery life, and build quality.


#55

When I bought my first calculator I didn't know anything about them. I heard of RPN and stuff but I didn't really know why it's better. In 1976 I was deciding between the TI-52 and the HP-25, I chose the HP-25 only because the material, the craftmanship and the look of the 25 was much better than the TI-52. I didn't know squat about what each could do. To say that the TI was better in build quality is beyond me.

#56

Quote:
Can you tell me what the attraction of the HP line is? It just seems that TI beat them every time esp. in the high end handhelds.

Sorry, I really don't know where you got that idea. I had both an HP-67 and a TI-59 around the 1978-79 time frame, and they were like chalk and cheese. The TI-59 died prematurely - it had much lower build quality than the HP-67. The keys didn't feel as positive as the HP's, some required much greater pressure than others, and several went dead within a couple of years.

Don't get me wrong - TI tried hard and did a mostly OK job with the '59. But in build quality, programming model, documentation and design they were obviously trailing HP.

Best,

--- Les

[http://www.lesbell.com.au]

#57

My first calc in 75 was a TI SR50. Wealthier students of physics at my university came with programmables soon (TI or HP) - I borrowed them sometimes, so I've experienced both sides. Of course, TI was easier for me to handle - and featurewise, HP and TI were almost equal. Nevertheless, the superior design (logical consistency, keyboard layout) of HP made me to turn to RPN.

In 77, I could pay for a repaired HP25C, and I used this for 5 years and sold it then to buy an HP11C, which disappeared with my bag 9 years later. Both calcs were fully functional when they left me and were looking almost mint (ebay grade ;-)) then. In the same time, I've seen quite some TI calcs worn down or semifunctional. Don't get me wrong, that's just my memory.

You may call this a difference in reliability, i.e. quality. This personal experience is my reason for holding high the flag of (old) HP. However, meanwhile things may have changed.


#58

Quote:
I used a TI59 (and still do) back at Yale and used to go head-to-head with the HP41 guys all the time. It seemed the TI trumped the HP every time on quality of build, power, convenience and even panache.

I know this is a recurrant theme in these replies, but it bears repeating in this instance...It's the keyboard.

When I was in school in the 70's, my friend had a TI-52. The keyboard was horrible!! There were two or three keys that he had to stand on to get to register...there was even one key that took about ten seconds to ooze back up after it was pressed. Even the keys that worked normally (for TI) had bad tactile feel because of the floating, non hinged design.

And don't even get me started on panache...My 29c with it's compact, beige colored case, (as opposed to the boxy black case of the TI), wonderful RPN entry (as opposed to the TI's clumsy algebraic entry (oops, lost count of the parenthisis...I'll have to start over)), color coded double shifted logical keyboard, with large enter key and slanted keyfaces (as opposed to the confusing 2nd and INV keys on the TI, with nothing but monotonous square keys), had major panache...and still does to this day. And when the guy who sat two seats behind me got out his HP67, the envy was almost palpable...
Hal:)

#59

Well with a TI you got what you paid for. The TI-58 was sold for $125 and the TI-59 for $300. Compare this with the $450 that the HP-67 went for or the $750 for the HP-97.

The only calculator that cost less than the TI-58 was the HP-21 ($80) but that one was brain dead (not even LASTx for crying out loud!)

The comparison would be as follows:

TI-59 ($300) vs HP-67 ($450)

TI-58 ($125) vs HP-25 ($125), or HP-29C ($195).

Also, even at a price 1.5 that of the equivalent TI, the HP products had less memory and no ROM packs.

E.g. the HP-29C has 98 steps and 30 registers while the TI-58 had 240 steps with 30 registers.

Similarly the HP-67 had 224 steps and 26 registers versus 720 steps and 30 registers for the TI-59.

Note, of course that the HP machines had merged instructions (e.g. STO 0 was one step vs two steps for the TI-58/59). But even so the larger memory allowed more complex programs to be developed.

In addition, on the TIs being able to jump straight into any ROM program, allowed ROM code to be used as an extension to user programs saving precious program memory.

---------------

From the usability perspective, apart from the RPN vs AOS issue (I won't go into this one :-), the TIs design was horrible. Apart from the 2nd and INV constructs that took a while to get used to (e.g. RTN is INV SBR, obvious once you know it, but I remember searching for the RTN key and cursing), there is that OP thing.

It appears that once TI run out of keys they placed all the remaining commands in the OP nn category (which includes printer commands, statistics, memory partitioning, register operations etc.). All in all a nightmare making the quick reference guide a must.

The HP-67 is exquisitely designed with functions selected with a minimum of keystrokes (remember the A-E keys having function shortcuts if no program is loaded) etc.

The quality is also excellent. Just holding an HP-67 in your hand gives you a feel of the quality and solid built of the device.

**vp

#60

All I say is this. My 1st calculator was a TI 35 (one of the oldest versions). The keyboard lasted about a year, then the bouncing carnaval started. I already knew HP calculators (my dad used them) so I bought one of the last HP 34C's available. It served me well during the last years of highschool (a tough period for a calculator in those days) and through a major part of my studies at university. I still have it, and more important: it still works the same way as it did when I had just bought it. OK, I had to repair it once (broken battery contact) but at least it could be repaired. I would say the old HP's were much better constructed than the early 80 TI's. I know the TI 59, it is I guess quite robust (a colleague of mine used one for a long time) but it has not the solid feel of the HP 67 that my dad used. In addition, the advantage in memory size the TI59 had was over in 1980 when the HP's 41CV came out.


Possibly Related Threads...
Thread Author Replies Views Last Post
  HP Prime vs TI : Factoring ? HP Pioneer 7 642 10-29-2013, 02:00 PM
Last Post: CompSystems
  HP-Prime_CAS versus MAPLE_CAS CompSystems 0 181 08-18-2013, 11:58 PM
Last Post: CompSystems
  Binary versus Decimal prefixes bill platt 22 1,325 04-27-2013, 11:22 AM
Last Post: Walter B
  HP & TI in the 70's revisited Palmer O. Hanson, Jr. 1 276 05-12-2012, 02:17 PM
Last Post: John W Kercheval
  OT--HP & TI in the 70s Matt Agajanian 12 834 05-09-2012, 01:20 AM
Last Post: Matt Agajanian
  HP vs TI's SR-50 Matt Agajanian 15 1,028 03-28-2012, 06:45 PM
Last Post: Bill Carter
  HP-41 Expandability Then and Now… TI? Kerem Kapkin (Silicon Valley, CA) 4 420 02-14-2012, 02:05 PM
Last Post: M. Joury
  HP 28S saw 36V versus 4.5V, damage and repair? Joshua Keena 4 311 09-10-2011, 10:58 PM
Last Post: Joshua Keena
  HP 15c LE - U.S. versus European version Michael Kussmaul 16 852 09-09-2011, 12:29 PM
Last Post: Walter B
  A speed comparison of the 41CL versus the 41CX (and my first YOUTUBE video) Geoff Quickfall 6 477 08-08-2011, 02:33 AM
Last Post: Ángel Martin

Forum Jump: