Asking HP to open source HP calculator software and patents

[acoto]

Hi

I has been asking myself about HP-28, HP-48 and HP-50 software.

Each day RPN/RPL seems less and less relevant for HP, and HP-50 seems to be on a dead end.

Do you think this is a right moment, and apropriate to ask HP to open source RPN/RPL calculator software, and give open permissions on related patents?

I think calculator enthusiast will be quite happy to have access to software and help to evolving it. I’ll like to see more saturn routines converted to native ARM for instance.

For world/people it may bring cheap RPN alternatives to casios ans TIs.

For HP, it may get interesting if global interest reignites in some way RPN calculator market. They may try to navigate in premium calcs market.

Has something like that been attempted before?

[cyrille de brébisson]

Hello,

I can ask management what they think about it...

I actually had a look in the archives, but unfortunately, I do not seem to have the 28S source code, nor the 48 on...

I however have the 50 series source code...

You talked about converting Saturn routines to native ARM... Well, this is what has been done for the 20b series, and is still used in the Prime... Saturn ASM code, rewritten in C!

Cyrille

[Thomas Radtke]

(05-15-2018 04:50 AM)cyrille de brébisson Wrote:  You talked about converting Saturn routines to native ARM... Well, this is what has been done for the 20b series, and is still used in the Prime... Saturn ASM code, rewritten in C!

Is this where the 20b/30b RPL like stack lifting comes from?

[teenix]

Does HP still have older source, like for HP-46 and HP-95C?

Would be good to complete all the LED emulators :-)

cheers

Tony

[Giuseppe Donnini]

I sincerely hope that Cyrille's statement doesn't extend to HP as a whole, namely that they didn't keep the source code for the HP-28C/S and HP-48S/G series. To me, that would be the most shocking moment of my "calculator life". And it would be high time for the community to try to salvage this outstandingly great piece of software, by getting in touch with pepole, like Bill Wickes, who surely must still have it -- provided HP doesn't object, of course. Maybe people who were granted access to parts of the source code in the 90's (through HP's BBS developer's conference) should then come forward, too.

[pdo]

It would also be good to re-license the HP Tools software suite (http://www.hpcalc.org/details/4263) so that it can be effectively maintained.

The current license forbids redistribution of modified copies, so it is currently not possible to build the tools on a modern computer without patching them. I have a patch to fix some of the problems (http://www.cloudycat.com, also at http://www.hpcalc.org/details/7860), but it would be nice to host the code on, say, github, to ensure its survival going into the future.

Paul

[Tim Wessman]

(05-15-2018 10:51 AM)Giuseppe Donnini Wrote:  I sincerely hope that Cyrille's statement doesn't extend to HP as a whole, namely that they didn't keep the source code for the HP-28C/S and HP-48S/G series. To me, that would be the most shocking moment of my "calculator life". And it would be high time for the community to try to salvage this outstandingly great piece of software, by getting in touch with pepole, like Bill Wickes, who surely must still have it -- provided HP doesn't object, of course. Maybe people who were granted access to parts of the source code in the 90's (through HP's BBS developer's conference) should then come forward, too.

Losing code/designs for old products is extremely common - not just large businesses but small as well. You move to a new system that doesn't directly support the old one, and no plans in moving forward the punch cards or whatever medium it is on takes place since the people who generally care about that are not the ones controlling the purse or the time.

Its just how it is unfortunately.

[Maximilian Hohmann]

(05-15-2018 04:21 PM)Tim Wessman Wrote:  Its just how it is unfortunately.

It was the same in the companies for which I worked as contracted engineer. But as engineers we are innovators, not historians, and the hourly rate and the project timing is such that it leaves no time for archiving duties. The only option would have been to take the stuff home and keep it there but that is usually forbidden by the terms of the contract.

[Gene]

I believe the best example of this is the loss (?) of the 12c source code. It required building an emulator to run the compiled binary, I think.

[Tim Wessman]

That's why we've learned to just stash all the marketing stuff in our code repositories as well. Code repositories seem to be moved forward quite regularly without issue since the mid 90s or so, but anything prior is quite iffy.

You'd laugh at the number of times one of the business/marketing people has come to me asking if I chance to have an old flier someone made 5 years ago, or a specific marketing image or presentation and I am just able to browse our archive directory and pull it for them usually. They just look at me like I'm a wizard or something.

[Jim Horn]

When the HP-110 Portable was on the market, Sam Chau, a fellow enthusiast/HP engineer modified his to have nearly twice the standard RAM. We asked for an in-house unsupported copy of the firmware source so we could modify it to use the increased RAM (the -110 let you partition it between user RAM and a RAM disk). After a month we heard back that the source code disks had been inadvertently used for another project and erased so the source code no longer existed.

Sam ended up writing a driver to treat the added RAM as another RAM disk. But the loss of source was an unpleasant surprise.

[pier4r]

(05-15-2018 04:21 PM)Tim Wessman Wrote:  Losing code/designs for old products is extremely common - not just large businesses but small as well. You move to a new system that doesn't directly support the old one, and no plans in moving forward the punch cards or whatever medium it is on takes place since the people who generally care about that are not the ones controlling the purse or the time.

Its just how it is unfortunately.

Thanks for the info. Another point for a sort of duty for archives that should be instituted by the society. People cry over the library of Alexandria but we lose tons of contributions and solutions every moment.

[hth]

(05-15-2018 04:50 AM)cyrille de brébisson Wrote:  Hello,

I can ask management what they think about it...

Cyrille

If you do ask, could you kindly ask what (open source) license they want to have on it? In my case "it" means the HP-41 source code that was released as list files back in the eighties without any proper license statement/agreement, just a "do not contact us for support".

Håkan

[cyrille de brébisson]

Hello,

Multiple points here...
1) Yes, big companies do loose stuff, all the time. In the case of the 28/48 source code, when the calculator operation was created in australia, we where told that all the data was on the "Scoobidoo" server... althrough said server might have existed in the past (actually, we could track no less than 7 servers named like this from the lates 80's to the mid 90's!), we were unable to recover the source codes as, even if these things were backedup, the backups could not be restored on more modern systems (backup are notoriously OS dependent)
This taught me one thing about backup systems. It tought me that you do NOT need backup systems. You need RESTORE systems! And that pure files works much better than any type of "backup" systems!

2) For the story, we ended up getting the 48/38 source code from an ex HP employee that had departed HP 4 or 5 years prior (for a competitor in addition!) and who had had the good idea to keep a personal copy of the code!
But this did not include any of the hptools (which we got from the publicly released version. Probably from Joe horn's disks!)

3) The Saturn ASM routines that I transposed to C are all the math functions. I have not done the same work for the RPL core system.

4) Can you confirm my understanding of what you said: you have the HP41 source code? and would like to be able to publish it? If yes, I would apreciate a copy, and I can add this to the list of things to publish...

Cheers,
Cyrille

[ndzied1]

Here is the patent I think the multiple equation solver in the 50g is based on. Granted in 1992 so it should be expired now. I would love this to be implemented on the prime.

https://patents.google.com/patent/US5175700

[Giuseppe Donnini]

Hi Cyrille,

Thanks a lot for answering this kind of questions!

(05-16-2018 04:58 AM)cyrille de brébisson Wrote:  1) Yes, big companies do loose stuff, all the time. In the case of the 28/48 source code, when the calculator operation was created in australia, we where told that all the data was on the "Scoobidoo" server... althrough said server might have existed in the past (actually, we could track no less than 7 servers named like this from the lates 80's to the mid 90's!), we were unable to recover the source codes as, even if these things were backedup, the backups could not be restored on more modern systems (backup are notoriously OS dependent)

If the server still exists, and if HP management doesn't object, it would be invaluable to recover these data, even in raw, binary form. A temporarily unreadable source is better than no source at all. And there might be people out there with more time at their hands than the company's employees, who might make it readable again one day.

Quote:2) For the story, we ended up getting the 48/38 source code from an ex HP employee that had departed HP 4 or 5 years prior (for a competitor in addition!) and who had had the good idea to keep a personal copy of the code!
But this did not include any of the hptools (which we got from the publicly released version. Probably from Joe horn's disks!)

Could you then confirm that you yourself, or the ACO team in general, never used the original development tools as seen in the following picture (actually a screenshot from a video taken by Jake Schwartz during one of Bill Wickes' presentations in the early 90's --- Thanks again, Jake, for all your efforts in preserving HP calculator culture for posterity!).



And what about the so-called ERS (External Reference Specification) for the HP-28/HP-48? Bill Wickes used to consult it whenever he had to face a question about a subtle, intricate detail he had forgotten. The RPL48 package by Raymond Del Tondo and Detlef Müller contained a small excerpt (with permission by HP Corvallis) from the HP-48 kernel ERS (dated August 7, 1986), namely chapter 11 about the "RPL Reader and Parser Tools", and this is surely one of the most interesting reads I came across, and well worth preserving.

Quote:3) The Saturn ASM routines that I transposed to C are all the math functions. I have not done the same work for the RPL core system.

Is there any chance that the source code of the HP-71B Math ROM might have survived? The implementation of the IEEE floating point math exceptions would be of particular interest here. A careful comparison of the source code given in the HP-71B Software IDS (vol.III) with a HP-48 ROM disassembly shows that - amazingly enough - the whole core of the IEEE math was preserved in the HP-48 code, albeit without the user shell around it. So, I think it would be an interesting project to implement a full-blown, separate IEEE user mode on the HP-48 (maybe in the style of Jake Schwartz' and Rick Grevelle's HP-16C emulator mode, with appropriate status line information for the exception flags, warnings, errors, etc.). Unfortunately, the HP-71B Software IDS only covers the main math routines, whereas the HP-48 code also includes the whole wider functionality of the Math ROM.

[Jake Schwartz]

(05-16-2018 04:58 AM)cyrille de brébisson Wrote:  Hello,

Multiple points here...

2) For the story, we ended up getting the 48/38 source code from an ex HP employee that had departed HP 4 or 5 years prior (for a competitor in addition!) and who had had the good idea to keep a personal copy of the code!

Something tells me that you are speaking of Diana Byrne, who left HP in the late 1990s and ended up working (remotely from Corvallis) for TI at that time (and even attended the 1998 HP Calc Conference in Vancouver, Washington as a TI employee).

Jake

[acoto]

(05-15-2018 04:50 AM)cyrille de brébisson Wrote:  I can ask management what they think about it...

Cyrille

Wow!

Thank you very much for taking up the idea. I think that is awesome, even if at the end they say no.

[David Hayden]

(05-15-2018 04:50 PM)Gene Wrote:  I believe the best example of this is the loss (?) of the 12c source code. It required building an emulator to run the compiled binary, I think.

A few years ago at HHC we had one of the 12C developers there to answer questions. I don't recall his name. I asked him if he had a copy of the source code and he said yes.

So the source is out there somewhere.

[David Hayden]

(05-15-2018 10:57 PM)pier4r Wrote:  Another point for a sort of duty for archives that should be instituted by the society. People cry over the library of Alexandria but we lose tons of contributions and solutions every moment.

It's a great irony of digital information. When the digital age started, people hailed it for the ability to keep exact replicas of information. "No more fuzzy reproductions of reproductions of reproductions!" Digital information has higher fidelity, but it's much more fragile than analog.