Asking HP to open source HP calculator software and patents

[pier4r]

I am slowly learning that

"digital information is highly available for short periods". (unless ton of care is given)

[DavidM]

(05-16-2018 02:54 PM)David Hayden Wrote:  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.

Could it have been Dennis Harms?

https://youtu.be/VitFi_KC8cQ?t=1170

[Tim Wessman]

(05-16-2018 02:54 PM)David Hayden Wrote:  

(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.

We spent a lot of time talking with him and he could not track it down...

[cyrille de brébisson]

Hello

> Could you confirm that the ACO never used the original development tools
~Correct, We used the so called "HP Tools" (the RPN + ASM compilers + library tool + linker). At the very begining, we used the unix versions on a hpux computer, but it was very slow and we very quickly switched to PC versions based on the same source code.
I then added the MASD asm syntax to that tool set (as the 49 included a lot of source code from the Meta Kernel which used that syntax).
I then codded a first "debuger" tool (in C), and very quickly a second (debug2, in Pascal) which became the integrated development environment that we used for everything from compiling to debuging, using the emulator or the real calculator (through a special version of the software with RPL hooks)...
We never used the Corvalis "debuger boxes"which were a HW debug sollution.

> And what about the so-called ERS (External Reference Specification) for the HP-28/HP-48?
Nope, never used it. We had our own documentation (mostly comming out of the French hacker scene), and our own knowledge + access to the source code of course...


>Is there any chance that the source code of the HP-71B Math ROM might have survived?
Yes and no....
The 48 math libraries are the childs of the 71 math libraries with dates in the comments starting in early 85. So in lots of ways the work that was done to convert these libs from saturn ASM to C for the Euro series (20/30/40b) did the work that you are looking for.

Cyrille

[David Hayden]

(05-16-2018 03:44 PM)DavidM Wrote:  Could it have been Dennis Harms?

Yes. That's the name.

I see from a later comment by Tim that they followed up and Dennis couldn't find the source code. So sad....

[cyrille de brébisson]

Hello,

I am working on a proposal for management regarding things that we could "open source".
Note that I am not making ANY promisses here (in my name or in HP's) and that I have no idea if/when this would/could lead to anything concreate. However, who does not ask does usually not receive much!

So,
in order to help me, if you have in your possession ANY old HP calculator (related) source code, can you please PM me and send me a copy so that I can include it in the proposal?

Regards,
Cyrille

[pdo]

(05-25-2018 10:18 AM)cyrille de brébisson Wrote:  I am working on a proposal for management regarding things that we could "open source".
...

Hi Cyrille,

Just wondering if this went anywhere? (it's been almost 6 months now).

As I mentioned earlier in this thread, I'm particularly interested in whether the HPTools collection for HP48 programming (https://www.hpcalc.org/details/4263) can be given a proper open-source licence.

Thanks,
Paul

[EugeneNine]

Could host the original and then patches on github.

So I just downloaded these and tried a make and get:

Code:

error: static declaration of 'sp' follows non-static declaration
 static int sp = 1;

[pdo]

(11-18-2018 11:22 PM)EugeneNine Wrote:  So I just downloaded these and tried a make and get:

Code:

error: static declaration of 'sp' follows non-static declaration
 static int sp = 1;

What OS was this on? I tried it last on Fedora 28 64-bit a few days ago, and although it spewed out lots of warnings it did manage to compile okay.

I started looking more closely at the warnings with a view to fixing potential problems, but it's difficult to get motivated knowing that I can't open access to my git repository without violating the licence.

Paul

[EugeneNine]

Slackware x64

looks to be due to the way older vs newer gcc handles these, older it was a warning, newer its an error. According to my searching anyway.

so its dependent on the gcc version.

[Thomas Okken]

I haven't run into this problem myself yet, but a couple of people who maintain Free42 ports have reported builds failing because of warnings being treated as errors where they didn't used to be.

If that is caused by -Werror being the default in the latest gcc, then that's pretty awful. Gcc has no -Wno-error with which to override -Werror completely; it only has -Wno-error=<warning-name>, for overriding -Werror for specific warnings.

Whether it's a change in gcc or in the default build settings enforced by some package builders, it's an awful idea, because it's breaking legacy software left and right. Upgrading your compiler should not cause your builds to fail.

[ijabbott]

(11-20-2018 08:30 PM)Thomas Okken Wrote:  I haven't run into this problem myself yet, but a couple of people who maintain Free42 ports have reported builds failing because of warnings being treated as errors where they didn't used to be.

If that is caused by -Werror being the default in the latest gcc, then that's pretty awful. Gcc has no -Wno-error with which to override -Werror completely; it only has -Wno-error=<warning-name>, for overriding -Werror for specific warnings.

Whether it's a change in gcc or in the default build settings enforced by some package builders, it's an awful idea, because it's breaking legacy software left and right. Upgrading your compiler should not cause your builds to fail.

In the case of "error: static declaration of 'sp' follows non-static declaration", it's undefined behaviour (ref. 6.2.2 Linkages of identifiers, particularly paragraph 7), and I think it makes more sense to treat it as an error rather than a warning, otherwise the linkage of 'sp' would be unclear.

[pdo]

Code:


    error: static declaration of 'sp' follows non-static declaration
     static int sp = 1;

Hi EugeneNine,

Looking at this more closely this weekend to try to understand the problem, I realize that the error you quote is actually what you get when you try to compile the unpatched hptools archive (hptools-3.0.8). So, for whatever reason, it seems that your application of the patch didn't work. Did you see the instructions on my page by any chance?

Also, in the process of doing this I realized that my patch is actually a little old, and not the current state of my git repository, so I'll update it soon. Currently, the warnings I am left with pretty much all relate to possible overflows when doing string copies -- they may or may not be bogus, I'll need to take a closer look sometime.

Paul