Why has not ......

richaj45

Hello:

Just some thoughts.

1)
Why has not Parallax continue to develop tools for the Prop? I would think the least they could have done by now is a BST quality tool that works on the three OSX, Linux, Win. If you want to sell silicon these days then you have to give away really good tools.

2)
Why has not Parallax build the next generation of stamp with a Prop as the byte interpreter? That could have a high performance product that has more memory then any of their existing stamps. After all interpreting byte codes is a strong point of the Prop architecture.

3)
Why has not Parallax come out with a Prop 1.5 by now? Not the every think but the kitchen sink Prop II. Parallax's very smart engineers did what so many do. That is, after one success, think they can put every bell and whistle in the next product. A prop 1.5 with more pins and more hub memory and faster clock would have been a great interim product that they could have completed.
I believe a Prop 1.5 is a six to nine month project. How many years has Prop II been in development?

4)
Why has not Parallax put any real-time debug hardware in the next Prop? Do they not know that even the eight bits MCUs have those feature. Without good tools, that include debug hardware hooks, the product with never move past the enthusiastic hobbyist state.

5)
Why has not Parallax realized that the Cogs nano-size memory of 512 instructions will restrict it to an interpreter of one kind or another? Because of that the chip will not be able to compete with any of the other micros except for niche applications were hand optimized code is worth the effort.

Any have any thoughts that are civil enough to write down?

cheers,
rich

Heater.

1) Sadly the Prop tool has it's compiler written in x86 assembler which is not
really conducive to porting around from platform to platform. There is a
dependency on some 3rd party GUI editor widget that does not help here either.

Still, an open source, cross-platform Prop Tool is in the making as we speak.

Dev tool work is in full swing with the GCC C compiler for the Prop. Which has
shaped up very well.


2) No idea about stamps but if you mean a small single core Prop I don't really
see the point.

3) A Prop 1.5 with the planned 64 pins would have been great. In fact I'd still
love it. Perhaps times have just moved on and it is not really worth doing
anymore.

4) Never had the need for debug hardware. I can usually get a COG to spy on
other COGs work if I need to.

5) Expanding the nano-size COG memory is not trivial. That address range is
built into the architecture by the size of the instructions src and dst fields.
Yes it could be expanded with banking or other such ugly cludges. My prefered
expantion is 64 bit COGs with 24 bit source dst fields addressing 32 mega LONGS
in each COG !!

Hand optimizing PASM into COGs is not always needed. With the propgcc C compiler
I can get a complete FFT algorithm written in C into a COG. Or a
FullDuplexSerial driver. There is a lot of scope there even in such a small
space.

Heater.

Of course pretty much all of these questions come down to resources, human, financial, and temporal.
Given their constraints I think Parallax has done an amazing job.

frank freedman

richaj45 wrote: »

Hello:

Just some thoughts.

<clip>



</clip>

Any have any thoughts that are civil enough to write down?

cheers,
rich

Hmm, rather disagreeable way to start a discussion. While a couple of proponents of competing material have gotten into some heated(?) threads, for the most part this forum seems to be among the most civil I have been a member of or regularly reviews. That said, are you for real, or just trolling for flames.

FF

StefanL38

The propeller-chip has a different concept. So things are done differently.

Reducing the price of the propeller-chip from $15? (don't really remember the old price) to $8, an unsigned QFN-package and more than 1.000.000 sold units indicate to me that the product is quite successful.

So from a point of economical success no need to do the things you mentioned.
It might be even that way: Chip Gracey wanted to develop is own highly customised piece of silicon that looks and works EXACTLY the way he personally likes it.
Same thing with the propeller-tool.

From a point of view of what would I like to have?
Of course why isn't there a company that sells a chip for 1 cent you just shout at "go build me a robot that delivers beer from the fridge and don't forget to mill the beer-can-holder for my motorbike!" and the chip knows what do do
and will be back in 2 minutes with everything done. Meaning I have BIG wishes of what a microcontroller should be able to do - well knowing most of these wishes will NOT come true.

best regards
Stefan

Kye

Please see the GCC effort.

Thanks,

Rayman

Regarding point 2, Bean has come up with Prop Basic, which I think fits that bill.
I'm not sure how much Parallax has supported that though...

richaj45

Thanks for the feedback.

I must say that i am amazed at what is accomplished with such limited resources of the Prop.

I think that is a testimony to the innovation of the people using the Prop.

I just think that the company receiving the benefits of the sales should do more work on the development tools than third party people develop.

But then it sound like the business model is ideal. Let others do the work and collect the $$$.

cheers,
rich

Heater.

richajj45,

But then it sound like the business model is ideal. Let others do the work and collect the $$$.

I'm not sure that entirely fair comment. Parallax has provided dev tools for the Prop. For free as well.
Recently they have invested in the development of the GCC C compiler for the Prop. A product that will be given away open-source and free.

frank freedman

richaj45 wrote: »

Thanks for the feedback.

I must say that i am amazed at what is accomplished with such limited resources of the Prop.

I think that is a testimony to the innovation of the people using the Prop.

I just think that the company receiving the benefits of the sales should do more work on the development tools than third party people develop.

But then it sound like the business model is ideal. Let others do the work and collect the $$$.

cheers,
rich

Dude, you seriously need to go back into your tent at OCW or wherever and ask your navel for a more accurate understanding of this product and users. Chip's company gets the $$$ because he created something of value to us and at a price point we are willing to pay. No more, no less. Microchip and intel are not forcing us to be prop users. If you want a freebie design and support it yourself.

Frank freedman

While I realize my answer may not be so civil, your rather scurrilous attack on Chip and Parallax was not exactly civil either. Also, we chose to use the device and contribute back of our own free will. What have you contributed of any value to this forum?

Kye

It's not fair to look at a private company from the outside and say what they should be doing. I understand where you are coming from, however.

Microchip and Intel, for example, are large companies with lots of resources. Parallax has said on their company about page that they are small company ~40 people and probably do not have the same resources as Intel or Microchip.

That said, I used the prop chip to make the CMUcam4, which will be coming out soon, work. The prop tool isn't amazing. I never used any of the pre-made parallax objects. But, I have a working product now that meets all my goals. That is what I care about.

I never at any point had to worry about getting the interrupt logic working with high speed frame grabbing and ntsc/pal video at the same time. ARM chip's did not offer me a solution to my needs.

pedward

richaj45 wrote: »

Hello:

Just some thoughts.

1)
Why has not Parallax continue to develop tools for the Prop? I would think the least they could have done by now is a BST quality tool that works on the three OSX, Linux, Win. If you want to sell silicon these days then you have to give away really good tools.

They are releasing a new PDT in 2012, which includes a new compiler written in C and a new IDE.

2)
Why has not Parallax build the next generation of stamp with a Prop as the byte interpreter? That could have a high performance product that has more memory then any of their existing stamps. After all interpreting byte codes is a strong point of the Prop architecture.

Parallax sells the PropStick USB, which is a DIP-40 form factor and is exactly like the BASICstamp. BASIC stamp users can make the move to this unit. Maybe they will make a BASIC for the prop when they have to refresh the BASICstamp when the SX chips run out.

3)
Why has not Parallax come out with a Prop 1.5 by now? Not the every think but the kitchen sink Prop II. Parallax's very smart engineers did what so many do. That is, after one success, think they can put every bell and whistle in the next product. A prop 1.5 with more pins and more hub memory and faster clock would have been a great interim product that they could have completed.
I believe a Prop 1.5 is a six to nine month project. How many years has Prop II been in development?

If you look back, there is a reference to Chip finishing up work on the Prop 1.5, so they did put some effort into this. However, launching another piece of silicon isn't cheap and they may have focused their resources on the Prop 2 instead of trying to fight a war on 2 fronts.

4)
Why has not Parallax put any real-time debug hardware in the next Prop? Do they not know that even the eight bits MCUs have those feature. Without good tools, that include debug hardware hooks, the product with never move past the enthusiastic hobbyist state.

The interesting thing about the Prop is that you could simulate all of the software in an emulator, maybe Chip has designs towards this? I'm certain that Chip has some ideas on this subject, but as others have pointed out, you can use additional COGs to implement on-chip debugging.

5)
Why has not Parallax realized that the Cogs nano-size memory of 512 instructions will restrict it to an interpreter of one kind or another? Because of that the chip will not be able to compete with any of the other micros except for niche applications were hand optimized code is worth the effort.

This is an interesting question, and it usually arises from those who haven't programmed the Prop. The prop could very well be considered a RISC chip, but it has some elements of CISC instruction sets. It has a very efficient instructions and with some clever, documented, methods you can do overlays or instruction streaming. I'm certain everyone at one time or another wanted more RAM!

Any have any thoughts that are civil enough to write down?

cheers,
rich

Phil Pilgrim (PhiPi)

pedward wrote:

2) Why has not Parallax build the next generation of stamp with a Prop as the byte interpreter? That could have a high performance product that has more memory then any of their existing stamps. After all interpreting byte codes is a strong point of the Prop architecture.

Parallax sells the PropStick USB, which is a DIP-40 form factor and is exactly like the BASICstamp. BASIC stamp users can make the move to this unit. Maybe they will make a BASIC for the prop when they have to refresh the BASICstamp when the SX chips run out.

The closest Propeller parallel to the BASIC Stamp is not the 40-pin PropStick USB, but the 28-pin Spin Stamp, which plugs into any socket that will accept a BASIC Stamp.

-Phil

pedward

Phil Pilgrim (PhiPi) wrote: »

The closest Propeller parallel to the BASIC Stamp is not the 40-pin PropStick USB, but the 28-pin Spin Stamp, which plugs into any socket that will accept a BASIC Stamp.

-Phil

Forgot about that.