HP Forums

Full Version: Open discussion: should newRPL change target hardware?
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
I'd like to hear opinions on a hardware target change for newrpl. Since the 50g hardware will be discontinued, I wonder if the project should change its target hardware to the Prime or stick with the 50g hardware.

The CPU is very similar, and it has been demonstrated that the Prime ROM can be replaced freely, so besides technical difficulty in achieving it I'd like to hear reasons for one or the other choice.

If we establish that the main objective of newRPL is "to be used":
* Which target audience is larger now: current 50g's out there or current Prime users?
* Which target audience will be larger in 2 years (using a mature newRPL)? 50g users will stay steady, Prime users will increase relatively, but total calculator users will probably fall down drastically. Can the Prime future sales offset the 50g head-start?
* Would Prime users be willing to flash an alternative ROM to have RPL?
* Which users will benefit more from newRPL, 50g or Prime? 50g users already have proper RPL, but they will get great gains in speed. Prime users already have speed, so they would have less incentive to run newRPL.
* Will 50g users finally switch to Prime if they can run newRPL on it?

I'm all ears.
With the demise of the 50g hardware and HP's commitment to the Prime it might be an option to offer your code base as a replacement for the Prime's less then perfect RPN implementation, in the same way as Bernard Parisse's work is included in the package. XCAS continues to live as an independent project but feeds the development of the Prime CAS. Maybe newRPL can be part of the Prime in a similar fashion.

As an option, if HP discloses the application API and the tools to build an App, you can offer a stand alone newRPL App which appears side by side with the built-ins.
I'd be happy of a Prime calculator (including some hardware upgrades to improve keyboard legibility etc.) that fully mimics the RPL operating system on the HP 50G. Imagine this would also be a cost effective path for HP since it means the hardware development cost can be spread over 2 end-user markets & hence greater overall production numbers.

From a marketing perspective, it may be better to offer 2 different calculators though rather than just a firmware update.

Cheers,

Michael
It seems both posts above point to HP coming to help, either by providing an SDK, or by integrating newRPL into their system, or by launching a new calculator more adequate for newRPL, or even with newRPL preinstalled.
I don't believe any of that is possible. newRPL is not on HP's map, I bet you nobody at HP downloaded the demo to see what it does.
The discussion I was proposing was more oriented towards what I (a.k.a. the newRPL team) should do for the good of the newRPL project, regardless of what HP plans are or what HP (can, should or could) do for this project, and what would be best for the prospective users of newRPL.
I would be happy with both offerings, with a slight edge for Prime availability.

If it becomes available on the 50G I would install it on my 50G and keep my 49G+ as is in order to have both OS.

If it becomes available on the Prime, I will have to buy a second Prime (rev C hopefully) and install newRPL on my current rev A Prime.

François

P.S. I downloaded the demo a few days ago. Will try to find time to play more with it.
If you're only developing for one device: Prime

If not: Prime and 50g

I always thought of launching it as an 50g application and then the natural progression would be moving on to the prime. However as someone above points out: It might be harder to convince people with a prime to try it out as the calculator is newer and very different, many people who has a prime maybe has a 50g they can spare etc. On the other hand, newRPL on the Prime is much more attractive as it does things that the Prime cannot whereas on the 50g it would result in a much faster application.
I'd target the Prime -- it will be available for the foreseeable future. The 50g will quickly become a collectable Sad Moreover, the 50g has RPL already so a second implementation isn't a big change. The Prime doesn't so you'd be doing something new.

If it were me, I'd ask HP nicely about the possibility of developing a RPL app for the Prime. I don't know what the official stance would be but asking can't hurt.


- Pauli
I don't think newRPL is looking to become an app on the Prime, it is looking to become the host of apps written in newRPL and as an interpreter for old RPL applications. People aren't disappointed with the Prime for not including a RPN app, rather they were let down because the value of having the whole platform written in a stack-based language that supports everything to be in RPN is key to the consistency we see on the HP 50g. So in order for newRPL to make a change, it has to be the host software and I don't think there's an easy way around that right?
The Prime's got enough horsepower and storage that you could just run newRPL on top of its OS, though, if it's capable of fitting onto a 50g.

I'd say the best bet is to make a platform-independent RPL core, really - low RAM and storage usage, written in a portable language like C (or, write a SysRPL compiler for ARM targets) - and then port it to platforms as needed, with unique front ends for the different platforms. But, that's obviously going to be more work.
(03-08-2015 04:15 PM)bhtooefr Wrote: [ -> ]I'd say the best bet is to make a platform-independent RPL core, really - low RAM and storage usage, written in a portable language like C (or, write a SysRPL compiler for ARM targets) - and then port it to platforms as needed, with unique front ends for the different platforms. But, that's obviously going to be more work.

If you read the documentation, it clearly states that newRPL is written in C/C++ and that the aim is to make the software pretty portable:

"This project proposes to reimplement from scratch the RPL language made popular by HP programmable calculators, in portable C/C++."

And

"Hardware abstraction layer: This module will run on bare metal hardware and provide the basic services needed for the calculator to function. It will be kept to a minimum, to make porting to a different hardware easier."
OK, then, so it's already in that direction.

I can see arguments for doing both the 50g and Prime. I'd say that keeping the 50g as a target means that resource usage will be kept down, making things much more capable on the Prime, though (or supporting cheaper new designs).
(03-08-2015 08:19 AM)Paul Dale Wrote: [ -> ]I'd target the Prime -- it will be available for the foreseeable future. The 50g will quickly become a collectable Sad Moreover, the 50g has RPL already so a second implementation isn't a big change. The Prime doesn't so you'd be doing something new.

If it were me, I'd ask HP nicely about the possibility of developing a RPL app for the Prime. I don't know what the official stance would be but asking can't hurt.


- Pauli


An SDK from HP? Won't happen, opening the platform could jeopardize their precious exam approvals. TI puts a lot of effort to block Ndless on the NSpire for that reason. We should expect something similar from HP and wouldn't blame them, they have to protect their business.

Currently I'm leaning more towards the 50g. Why?
People buying Primes today could care less about having RPL, so they are very unlikely to flash an alternative OS.
The only real reason to target the Prime would be so people using 50g's today can "switch" to a Prime with newRPL. But one thing is to reflash an inexpensive 30b and a different one is to do that with a US $100+ device. I see some resistance there.
Many people with 50g's today are happy with their machines, so there will be resistance to change too.
On the other hand, the same people with a 50g on their hands (that has already been given the death sentence), being offered speed+grayscale+more RAM available+SDHC compatibility... will at least give it a try.
But nothing is set in stone, the discussion is still open.

Availability of hardware is a very good point, tough. [...Thinking out loud...] With the 30b and 20b gone, and also with Eric and team taking so long on the 43S hardware (and huge cost associated with it) perhaps we should all join forces and create an alternative OS to target the Prime. Then the Prime could be THE calculator hardware, and the user can flash a rom with classic RPN mode 34s/43s, or newRPL at their choice (switchable from the UI, not flashing 2 separate roms).
You've always said you needed more space for features: There's plenty of room for everybody's project in the Prime hardware.
If this coordinated effort were to happen, then I'd definitely choose the Prime and go with the flow. The extra effort would be spread across more people and would be definitely worth it.

What are the future plans for your projects? How is the 34S team going to handle the lack of hardware? You have one finished project on dead hardware and a prospective project on prospective hardware. I have a half-baked project, on hardware with its days numbered.
Perhaps we should all jump to a bigger boat to stay afloat? (yes... ok, battleship)


Claudio
(03-09-2015 04:23 PM)Claudio L. Wrote: [ -> ]What are the future plans for your projects? How is the 34S team going to handle the lack of hardware? You have one finished project on dead hardware and a prospective project on prospective hardware. I have a half-baked project, on hardware with its days numbered.
Perhaps we should all jump to a bigger boat to stay afloat? (yes... ok, battleship)

A concise answer regarding me: I simply don't have much time for any hobby project, including the 43S. If the hardware ever reaches my desk my interest will rise but most probably not to the extent of the 34S development. (I was kind of "laid off" at the time.)
I wasn't suggesting a reflash of the prime, rather a RPL application running on it. That would be more palatable all around I'm sure.


- Pauli
(03-09-2015 09:08 PM)Paul Dale Wrote: [ -> ]I wasn't suggesting a reflash of the prime, rather a RPL application running on it. That would be more palatable all around I'm sure.

Yes, I got that. But to do that we need some sort of SDK from HP, hence the beginning of my previous post. I just don't see that happening, they won't shoot themselves in the foot.
Has anyone asked about the possibility of developing third party applications?

NDAs exist for situations like this.


- Pauli
New RPL on the Prime would be great !
(03-09-2015 09:52 PM)Paul Dale Wrote: [ -> ]Has anyone asked about the possibility of developing third party applications?

NDAs exist for situations like this.


- Pauli

I personally tried to contact Tim Wessman via personal email back when they released the Prime (I think one or two months after the release). I never got an answer, perhaps it wasn't the right email address, or the right time to ask, or the right person to ask, or the right person asking the question.
I signed an NDA with HP in 2006, and as far as I recall there was no expiration date on it, so who knows, it could still be valid.
But still, under NDA it can't be open source, which means nobody can find or fix bugs, and only a couple of people will have access to the code.

Claudio
Another + for the 50g is that the keyboard has lots of dedicated buttons for programming in RPL. Maybe that won't be that big of a deal as I suspect lot's of people prefer writing programs on a computer anyway, I for instance am planning on writing a newRPL mode for Emacs.

Claudio, have you thought about making an newRPL interface for iOS or Android like there is for WP34s? If you're afraid that it wouldn't be approved on the App store, take a look at Pythonista for instance which implements Python on top of iOS. I don't think any calculator application on those devices are really comparable to the completeness of a high-end calculator, which is too bad since it would certainly be possible. I don't mean this as the first target device but maybe that's something to aim for in the long run.

I'm all in for the joined-forces, "Battleship" project and you should call it exactly that.
(03-10-2015 08:30 AM)Ailurus Wrote: [ -> ]Another + for the 50g is that the keyboard has lots of dedicated buttons for programming in RPL. Maybe that won't be that big of a deal as I suspect lot's of people prefer writing programs on a computer anyway, I for instance am planning on writing a newRPL mode for Emacs.
This sounds really cool!
And yes, the 50g keyboard has a better layout, though with an overlay it can probably be made comfortable.

(03-10-2015 08:30 AM)Ailurus Wrote: [ -> ]Claudio, have you thought about making an newRPL interface for iOS or Android like there is for WP34s? If you're afraid that it wouldn't be approved on the App store, take a look at Pythonista for instance which implements Python on top of iOS. I don't think any calculator application on those devices are really comparable to the completeness of a high-end calculator, which is too bad since it would certainly be possible. I don't mean this as the first target device but maybe that's something to aim for in the long run.
I did, that's why the simulator is Qt-based. I never actually used Qt for any device other than a PC, but it should reduce the porting effort to Android at least.
Reference URL's