Scientific Specific Units

[luisphysics]

This is aimed primarily at Tim.

Scientist like to use certain units because they are useful. I'd like to define my own unit and have it registered to the Unit Table menu, and have CONVERT() know about it.
One such unit is Mass-Energy.
I Tried to do this CONVERT(1_(u),1_MeV/c^2)
but I know I can't do that. I would like a unit called _MeV/c^2 where
1_u = 931.50 _MeV/c^2

This is useful when calculating binding energies in Nuclear physics which is what I do.

Not to take a jab, but my hateful Ti-Inspire CAS can handle this. I can define my own units and the unit converter can do the right thing. Which is about the only time I use it.
If there is a way to make such custom units I haven't been able to find it.

An Idea for notation for such a relation would be
_u==931.50_MeV/_c^2

where == doesn't actually need to be used, but it should represent that these units are related in this way.

[towe]

I'm pretty sure we won't see another update for the Prime.

Why don't you just define a function for that?
Shift-Define, name the functionn "convMeV" and put that equation in there.

[DrD]

(12-05-2015 09:23 AM)towe Wrote:  I'm pretty sure we won't see another update for the Prime ...

Your quote doesn't inspire a lot of confidence. I sure hope it's just a way of trying to coerce an update soon, and not based on privileged knowledge. I know that, here in the Museum, the authors have mentioned a lot of 'fixes' made, since the last update. It seems like Tim, Parisse, and others, closely watch over this forum, always at the ready to help where they can, and capture bug issues, when reported here.

Based on that, I have to think there probably will be future releases, and each will, probably, represent a very conscientious effort by the team, otherwise, why bother?

As a side question, what would an ideal trigger be to send out official updates? Once a year? Every time something as minor as a spelling/grammar change occurs?

As a product targeted for education, consistency of performance must be quite important, so that differing outcomes, (from updates), don't add further confusion within a classroom setting. Keep in mind that the update process itself, always seems to require special steps, due to the range of hardware, knowledge, skills, and abilities of hp prime users. Hewlett Packard recently made operational changes in it's overall business model, and how could working through that help move the update process forward?

-Dale-

[ndzied1]

It's not for the Prime but if you have a 50g the Unitman program is awsome:
http://www.hpcalc.org/details.php?id=4275

It does exactly what you want to do. Plus, I find converting units on the 50g much much easier than on the Prime.

[luisphysics]

(12-05-2015 09:23 AM)towe Wrote:  I'm pretty sure we won't see another update for the Prime.

Why don't you just define a function for that?
Shift-Define, name the functionn "convMeV" and put that equation in there.

Well seeing as I expect myself doing physics for the next 50 years I will be a patient man. So I'm sure we will see an update.

You have to look long term.
The MeV/c^2 is not going anywhere. Physics has been using it for a long time. So that means that this unit and others like it should be included in the calculators built in functions.
I understand there are hardware and software limitations. But if there isn't a user interface to define units or if it isn't possible due to design limitations, then adding them slowly to a standard place would be ideal.

[luisphysics]

(12-05-2015 03:32 PM)ndzied1 Wrote:  It's not for the Prime but if you have a 50g the Unitman program is awsome:
http://www.hpcalc.org/details.php?id=4275

It does exactly what you want to do. Plus, I find converting units on the 50g much much easier than on the Prime.

Hold on hold on you got me all excited now. Is it true that on the 50g, it has click Shift-and-Hold it, then press "Some key", and this defines a user-key?

[ndzied1]

You have to load the program on the page I linked to.

Once you download the file, unzip it and there are instructions in .htm files.
Unitman.htm is in english. Looks like there is french, german and spanish files available as well.

I haven't used it in a while so I'm not sure I'll be much help but I'm sure others here have.

[Tim Wessman]

(12-05-2015 09:23 AM)towe Wrote:  I'm pretty sure we won't see another update for the Prime.

Interesting. What has given you cause to feel that?

[Tim Wessman]

(12-05-2015 03:51 PM)luisphysics Wrote:  

(12-05-2015 03:32 PM)ndzied1 Wrote:  It's not for the Prime but if you have a 50g the Unitman program is awsome:
http://www.hpcalc.org/details.php?id=4275

It does exactly what you want to do. Plus, I find converting units on the 50g much much easier than on the Prime.

Hold on hold on you got me all excited now. Is it true that on the 50g, it has click Shift-and-Hold it, then press "Some key", and this defines a user-key?

I'm not certain for sure, but I think it was limited to a single set of "custom" related units. I was never able to figure it out myself when I tried back in the day...

[Tim Wessman]

(12-05-2015 03:10 AM)luisphysics Wrote:  One such unit is Mass-Energy.
I Tried to do this CONVERT(1_(u),1_MeV/c^2)
but I know I can't do that. I would like a unit called _MeV/c^2 where
1_u = 931.50 _MeV/c^2

It is true there is no way to define custom units. One problem we have is that the "unit" object is basically full in terms on the low level implementation at this point. We have discussed custom units (and "true" custom units not just the "make a synonym for another set of built in units - e.g. you could define _$, _€ that have no relation to any other group of base units) but that is definitely below the list of "fixing" the unit conversion UI. After that, some creative thinking to avoid backwards compatibility issues will be needed.


However, I'm perfectly happy to receive feedback/suggestions for new "built in" units that are well established in useage with actual names and (most importantly) are based on the existing base unit groups.

[luisphysics]

(12-05-2015 03:32 PM)ndzied1 Wrote:  It's not for the Prime but if you have a 50g the Unitman program is awsome:
http://www.hpcalc.org/details.php?id=4275

It does exactly what you want to do. Plus, I find converting units on the 50g much much easier than on the Prime.

I want to be clear I don't own a 50g. And it is likely I will never buy one. I'm looking more into the future of the HP Prime.

I notice a lot of people love the 50g. If I had to guess the reason HP didn't just upgrade the 50g hardware and roll over the software, was likely because the firmware/software/hardware are actually one fused unit. So if that is the case, then the 50g will be stuck in the era in which it was created.

I hope the HP Prime is not in bed with the hardware, but that it is a software, placed on top of the hardware, so that looking into the future when hardware needs to be upgraded we can ditch the current HP Prime hardware, port all your software to a new machine with a better display, and processor..ect. And that I hope they add support for my custom units along the way.

[luisphysics]

(12-05-2015 07:01 PM)Tim Wessman Wrote:  It is true there is no way to define custom units. One problem we have is that the "unit" object is basically full in terms on the low level implementation at this point. We have discussed custom units (and "true" custom units not just the "make a synonym for another set of built in units - e.g. you could define _$, _€ that have no relation to any other group of base units) but that is definitely below the list of "fixing" the unit conversion UI. After that, some creative thinking to avoid backwards compatibility issues will be needed.


However, I'm perfectly happy to receive feedback/suggestions for new "built in" units that are well established in useage with actual names and (most importantly) are based on the existing base unit groups.

Well Honestly looking at most of my books the Prime team has a lot of good useful units like a barn which I use for nuclear cross sections, which is great.

But E=mc^2 => The units to Einstein's Energy to mass equation is MeV = kg * (m/s)^2

But no one ever divides out the c^2 numerically. The reason is that in most cases it just cancels out.

So the mass of the Proton is 1.67*10^-27kg eWWW no one wants to see that.
So what we do is just say the proton has 938 MeV/c^2. which is the same units as kilograms.

What we don't do is
((938E6_(eV))/(3E8_(m/s))^2 =1.04E-14 MeV/(m/s)^2
mainly because the answer is ugly.

And simplifying all the way is also ugly.
((938E6_(eV))/((3E8_(m/s))^2 )*(1.6E-19_(J)/_(eV)) = 1.67E-27 kg if you simplify J/(m/s)^2

I'd like to note that a good example and a familiar one, is the Higgs boson. Check out wiki https://en.wikipedia.org/wiki/Higgs_boson on the right side of the page for the mass. It is in GeV/c^2 because no one wants to write kg.

If you need a more convincing argument I can find one.

[towe]

(12-05-2015 06:51 PM)Tim Wessman Wrote:  

(12-05-2015 09:23 AM)towe Wrote:  I'm pretty sure we won't see another update for the Prime.

Interesting. What has given you cause to feel that?

There's still a lot of bugs 2 years after release.
The translation quality is still horrible, 2 years after release.
The user keyboard doesn't work properly for all keys - 2 years after release.

You'd think such rather massive issues would've been fixed already, at least in the months since the last update.

[DrD]

So, three little things ... and you're predicting there won't be any further updates? What other sage advice do you offer, stock market, lottery picks?

[luisphysics]

(12-06-2015 10:57 AM)DrD Wrote:  So, three little things ... and you're predicting there won't be any further updates? What other sage advice do you offer, stock market, lottery picks?

I agree.
I would like to know more about your prediction.

To Towe:

Honestly quoting product updates is interesting. BUT you have to convince me that the contribution to the halt in innovation in the HP Prime is first order.

Example:
As if the HP exec woke up and said, "we are going to give the world a hint that we are going to stop making the HP Prime and fire all the developers. We are going to ensure that the hardest and most difficult bugs do not get fixed...AND we are going to make sure we make the interval between firmware updates longer than 1 month so people realize (if they are smart enough) that we are going to get rid of the HP Prime.

[Dirk.]

Is this really the right place to discuss HP's update policy?

However, it might be the right place to think about the units-problem.
MeV is often used in the context of nuclear and particle physics. Usually a system of natural units is used, where c=1 (without any unit). So E=mc² becomes E=m (or E² - p² = m² instead of E²-p²c²=m²c^4). This means giving up the feasibility of dimensional analysis.

Should a calculator support this? I personally don't think so. This will create a lot of confusion among not so experienced users, whereas physicists should be able to convert and do the calculations correctly.

-> Therefore: Keep the system as it is - consistent.

But this is only my personal opinion. The people making their living from producing the calculator have to decide.

Dirk.

[luisphysics]

(12-06-2015 05:08 PM)Dirk. Wrote:  Should a calculator support this? I personally don't think so. This will create a lot of confusion among not so experienced users, whereas physicists should be able to convert and do the calculations correctly.

Dirk.

If they are confused over some unit existing on their calculator, then it is because they haven't been exposed to it.

When I was learning units conversions I was usually alone with my calculator. I was able to check to make sure I was doing things correctly. That built confidence and trust in my abilities, that allowed me to move on and learn the next thing(make note that not everyone learns in the same way). Hence my calculator became a learning tool.

Humans are not perfect we make mistakes, we are lazy and we hate repeating ourselves. Why do you think the computer was invented?
I'm not saying you shouldn't know how unit conversions work. The principle behind conversions are super simple. But when you have a long calculation it is easy to make a mistake.

Real world Examples.
Geant4 a physics simulation software has built in units, reason, get rid of mistakes.

Interesting fact, there is a push from the physics community to get a units library into the C++ ISO standard(an International Treaty).

If making units available in computation wasn't important, then no one would want to develop a standard.

Think back to the mistake that happened on Mars when they object that was meant to orbit mars crashed on the surface. Why? Because of unit conversions. Here is the clip. Time index 11:45 https://www.youtube.com/watch?v=7P536lci5po
You can watch the whole thing, and rewind a bit to see when he introduces himself.

So the lesson here is, just because you are a scientist, it doesn't mean you are going to catch all the mistakes. Because you should be able to do things right, it doesn't mean you will.

[Tim Wessman]

(12-05-2015 07:32 PM)luisphysics Wrote:  But no one ever divides out the c^2 numerically. The reason is that in most cases it just cancels out.

So if I am understanding you correctly, the problem is that "by convention" the c^2 is just left in. Here is where it gets very problematic. How would the calculator possibly be able to know when a unit value that cancels with another "should not by convention have anything done with it"?

[Tim Wessman]

(12-06-2015 10:33 AM)towe Wrote:  There's still a lot of bugs 2 years after release.
The translation quality is still horrible, 2 years after release.
The user keyboard doesn't work properly for all keys - 2 years after release.

You'd think such rather massive issues would've been fixed already, at least in the months since the last update.

Thanks for your reply. I understand where you are coming from now and will try to address what I can.

I'd like to go in reverse order and start first with your comment of "You'd think such rather massive issues would've been fixed already". To that, I would respond with a question right back - "Massive according to whom?"

The problem is that we as humans tend to only focus on issues that we see, or things that directly impact us. Do that vast majority of users think there is a massive problem with the user keys? Nope. Unfortunately, the vast majority of users of Prime have never touched user definable anything, nor would ever do so even if it were heavily promoted and 2/3 of the manual were devoted to its use. You and I are not normal users, nor in essence the main target customer for Prime.

I *personally* am on your side in agreement that the user keys have problems right now and I'd love nothing better then to spend the many hours to fix it (not a simple task unfortunately due to some lower level architectural changes that would be required first before being able to even start touching what you see as the problems there).

However, every issue must be weighed against many factors and has many different groups providing feedback.

One of the most important factor is "How many people are being impacted by <issue X>?". If the answer is "very few" then it will become much less likely to be addresses over another issue that might impact 50% or more of users.

The next most important question is "What *other* things may not be done if we take the time to resolve <issue X>?"

For example, let's say that it were possible to give this choice:

1. Fix and make the unit conversion UI better then any prior calculator.
2. Make base conversions work even better then any prior calculator.
3. Make RPN work in the CAS.
4. Return 10MB of RAM to the end user for use in programs (giving you double what you have right now)
5. Implement a 3D grapher.
6. Fix the user keyboard.

Assume all these items take the same amount of time to implement. Which one is the correct choice? I suspect you might have a hard time deciding. I suspect whichever answer you pick might differ from mine, and probably differ from a large number of other people's choices.

The point I am trying to get at here is that it is NOT an easy process to decide what to resolve, what new to implement, and in many cases what *is* a major problem. For example, you cite horrible translations. I'm not certain we've ever received huge major complaints from distributors, educators, support lines, emails, customers, etc on translations. Sure, a few people have commented on some individual bad translations here or there, and we've corrected those when they come in. There has been no "massive outcry" about horrible translations though. In fact, we've received quite a bit of feedback generally that they are pretty good.

Sometimes however, the suggested translations provided as "what the translation should be" *aren't* good translations and the HP translators come back with extremely solid reasoning why they have used the word/phrase they have used. For example, a math term that is common in one region, but in every other country speaking that language it is a different word. Translation is a very difficult thing unfortunately.

If you have specific areas of feedback, please send them to one of the support paths available - phone, email, etc. Sending things to me directly, or posting on a discussion group like this (which is not an official HP operated support site) is NOT going to get the visibility that you'd get by emailing a calculator distributor, calling the support lines, or sending an email to support.

Now for your first and my last comment, I don't think anyone can reasonably say that Prime is worse now then when it started. True, one may argue plenty about whether some of the issues that have been resolved never should have happened. True, one can argue that it was a mistake to switch away from the 50g codebase (it wasn't ), and there are plenty of other things to argue about.

I'd argue though, that there are much fewer issues then there were before, and the issues being found are generally less and less critical. I'd also argue there there is much more capability is in Prime now then existed two years ago. I would even argue that despite the initial roughness at launch, it was the correct time to launch. It takes AGES for new calculator to be accepted (speaking from a product marketing lifecycle time frame) and move into the place where it begins to sell well. Waiting an additional 2 years till now would have done nothing to speed that up unfortunately.

There are plenty of good signs that overall Prime is moving in a good direction, and being well received generally. I mean, just recently THE MAJOR TI USER SITE IN EUROPE (tiplanet.org) placed the HP Prime as tied for top choice for the calculator to purchase! (It would have been #1 except for a problem with battery life when testing for an exam starting requirement in 2017). I don't think anyone would have seen that recommendation coming just 2 years back...

Anyway, I hope I have been able to explain a bit of the difficult reasoning that goes into figuring out what to work on and when.

[luisphysics]

(12-06-2015 10:21 PM)Tim Wessman Wrote:  So if I am understanding you correctly, the problem is that "by convention" the c^2 is just left in. Here is where it gets very problematic. How would the calculator possibly be able to know when a unit value that cancels with another "should not by convention have anything done with it"?

I don't know how the Prime is units logic is implemented, therefore I can't comment.

I don't expect fast response on this issue. I know it is hard to implement things. Since your response didn't come with an 'aha' moment, it is less likely. At this point I don't feel that I have vetted this "recommendation" enough to pass it on to official channels. It would be not so nice if I try to push something official channels when there isn't even a way to implement it. I will have to give it some though. I have made enough of a stink about the issue that if it was implementable you "may" remember.

Maybe I should start a new Thread but what is wrong with the Units UI?