Future of PBP Compiler

[Ioannis]

Hi Charles.

We 've been using PBP for many years now, but the product seems to have bogged down.

No updates, no announcments of any kind, no hope for any improovments.

OK, I know it is a difficult task, but others (Compiler developers) have moved on. PBP is stuck on the trusted old now version.

The wish list is large, but a new compiler is a must now. Chips like 24 or 32 series should be supported for a viable compiler.

I see no point of surprise from your side. So if there is really a plan from your company to give people a new product, the time has absolutely come (or maybe pass already?) to announce it.

Is there any future of the PBP?

Ioannis

[Charles Leo]

Just wanted everyone to know that I'm not ignoring this. I'm trying to poll everybody involved in order to formulate a meaningful answer.

[DavidP]

Looking forward to what you find....

[Ioannis]

I'm trying to poll everybody involved in order to formulate a meaningful answer.

Thanks for the reply, but the above phrase is not very promissing... After many years of waiting for a serious update/upgrade we are looking for a ... meaningful answer?

I'd expect something like: "OK, our next compiler is going to support 16/32 bit MCU's and also the long waiting Color graphics!"

Ioannis

[Henrik Olsson]

I know from previous history that MeLabs "policy" - official or not, I don't know, has been to never say or promise anything about what's coming or not coming in the future so I was a bit surprised to see Charles saying that he will "formulate a meaningful answer" - I'm looking forward to it though I fear what it's going to be.

PBP3 works great for me and I love using it. It's simple, powerful and IME very very reliable. I started with PBC ~15 years ago (or more) and soon upgraded to PBP and have stuck to it since then. I would hate to see it die.

With that said, even the 8 bit devices has become extremly powerful compared to old days of the 16F84 etc and with more flash and more RAM managing programs written in PBP for these devices is getting harder. In the "old days" proper functions and local variables might not have been "needed" and strictly speaking it is of course totally possible to do anything without it but it's getting harder and harder. And Writing code for sharing and re-using is difficult without it, or actually USING already written code is difficult.

First came the addition of LONGs and then in PBP3 the improved CONFIG handling, the conditional comlilation stuff and the user command stuff which is still lacking official documentation. Apart from that not much has happend during 20 or so years - only more and more devices has been added - which of course is good.

/Henrik.

[Charles Leo]

I've stirred the pot and the discussions are taking place. While that is ongoing, I would like to clarify what we're talking about.

The issue that I've thrown in front of the financiers and developers is PBP for PIC24. That's the only issue on the table at the moment. It's important not to conflate other wish-list items, as they may be subject to different circumstances.

PIC32 support can only be considered after we have a stable PIC24 compiler. These two architectures require completely different compilers. We are not contemplating PIC32 support at this time.

The age-old wish-list items of local variables and user-defined functions are not going to happen in PBP Gold. The compiler would have to be rewritten from the ground up and the result would be something very different from PBP. In my opinion, it would do damage, negating the very things that I love about PBP Gold.

New commands are not part of this discussion. The criteria for new commands are very different than the obstacles we face in developing support for new architectures.

I reject the notion that PBP is doomed if it doesn't support new architectures. PBP is an awesome development tool for 8-bit micros and will continue to be that. I 'll go even further to say that PBP is perfectly matched to the 8-bit PIC microcontrollers - which brings me to the nature of a PIC24 compiler.

Allow me to describe a hypothetical product. I'm not trying to discourage anyone, but the reality of a new compiler might be different from many users' expectation. Please discard the idea that we will simply add the PIC24 devices to PBP, charge $50 for an upgrade and continue our practice of free technical support. That is simply not possible.

If we add support for PIC24 devices, it will be released as PBP Platinum Edition.

PBP Platinum retail price will fall in the range from $600 to $1000 US. Owners of PBP Gold would be offered a discount of $270 to obtain the new edition. This is new-compiler development. Some of our methods and concepts will carry over from PBP Gold, but the end result is a new executable and new assembly libraries.

The purchase price would not include technical support beyond installation assistance. Ongoing support would be a separate subscription or per-incident purchase.

PBP Platinum would require more technical expertise from users. In order to realize the advantages of the PIC24 parts, the user will have to understand fundamentals of the target device. Platinum will only be recommended for advanced users. The option of building a "dumbed-down" compiler to simplify PIC24 development has been rejected. It would defeat the purpose of the endeavor. I want a tool that gives me the same power over PIC24 that I currently have in the PIC18 family.

PBP Platinum would only run as a language tool in MPLABX, at least until further development produces an IDE. I'm not happy about this one at all, but I'm told we need the Microchip infrastructure.

Development will take at least a year.

Perhaps this information will give you an idea of the decisions we face. As far as the real possibility of this happening, I've spoken to all of the necessary players and I haven't been shot down yet. That is, honestly, more progress than I expected to achieve.

[Ioannis]

Future of PBP Compiler

Hi Charles. Thanks for the detailed reply.

I do understand the situation and your plan. Also about PIC32, a different animal, also understood your point.

In a new development I feel there are two ways.

1. Ask people what they want/need

2. Create needs and make people buy your product (What Apple does...)

From your reply, it seems that you decided a third way, of doing a difficult product, for advanced users and at a high price tag.

I am not of course the expert here or the manager, but sure this is not what was expected.

I sure understand that a PIC24 compiler would be a different product, so it would need time and money to develop, so it should be sold accordingly. But $1000 max? Isn't it too high? And with no free support, at least for a period of time?

If you insist on this, then the compiler should be one of a kind. With features not found elsewhere.

Anyway, all these are thoughts on a good will basis. No offense given or taken.

Ioannis

[Charles Leo]

I think your assessment of the situation is accurate, Ioannis. These are precisely the considerations that are being discussed.

[richard]

PBP3 works great for me . It's simple, powerful ,efficient and reliable.. I too would hate to see it die.

As Henrik has said, 8 bit devices have become extremly powerful compared to old days of the 16F84 more flash and more RAM ,managing memory in PBP for these devices is getting harder. It seems to me the most useful feature missing from pbp3 is a C style structure type array. I can just imagine how much easier it would have been to develop my tft and touchscreen code instead of battling along with asm code and ext modifiers.
Just looking at the forum views on the TFT project {11000+} there is obviously some interest out there for such a feature but to my knowledge only two members have successfuly implemented my version of it [probably due to my convoluted code,one of my best efforts was once described as chicken scratchings by a real programmer] I'm sure an upgrade with that feature alone would be well recieved.
I for one am not particularly interested in pic24/32 devices. the enhanced 8 bit ones do mostly everything i want ,what they can't do a raspberypi with python handles easily [no $1000 dollar compiler required]. In fact a rpi with a pic18 attached can conquer nearly anything.
I would rather get a better pbp3 for my $300 upgrade money. I guess it comes down to where is your market now , is it Developer$ or not for profit hobbyists like myself. i have a feeling any serious developers using pic24/32 chips would be well and truly entrenched in C these days anyway .
If i had a vote pbp3+ first donald t last