Tools for other OSs (MAC, Linux)

gardener

The other day I had my tech. support question, but guy from Parallex asked me to pose the question on the forum. We are small consulting/design house and for our new client we intend to propose the solution with Propeller chip, but there is some problem. Few years ago someone decided not to replace computers with new windows computers, but replace old windows with linuxes. Regardless the reason we have no window computers available to load the tools. If we propose Parallax chip, we could end up using it. Apparently changing the computers is out of the question.

Does Parallax have a roadmap to expand their tools to other than window platform?

Just for information, most of our core software are not open source software and many packages cost real money. Parallax does not have to open their tools.

Kevin Wood

You might want to read this thread: http://forums.parallax.com/showthread.php?p=631464

Lawson

I believe that the Propeller tool runs in WINE except that serial communications does not work. (aka it won't program) For Programming a Python script is floating arround this forum that will program in Linux. Also, PropTerminal recently added full programming. (think it's only a windows program at the moment, ask the developer)

The "Good Thread Index" sticky is a good place to start http://forums.parallax.com/forums/default.aspx?f=25&m=148376
huh, the loader is a bit hard to find for some reason. going to search.parallax.com and putting in "python loader" got a good hit on the third try. this => http://forums.parallax.com/forums/default.aspx?f=25&p=2&m=150887 page, second post from the top.

Best of luck,
Marty

gardener

Thank you for your help, but I do not think that we get Parallax support there. What bothers me the most is not a very logical answer given by Paul Baker (seems one of the key person from Parallax):

“I don't believe we ever made such a promise. Something I don't think many people understand is that we have a single person who develops the Propeller tool. And while this consumes the majority of his time, it is far from the only responsibility he has. If we met all our consumer demands we would have to write a Tool, C compiler, BASIC compiler and Forth compiler and have each of these working on Windows, MacOS and linux, not to mention the numerous requests for added features we get every week. To do this we would have to double the number of employees at the company, all for something we charge nothing for.”

I'm sure he knows that you do not have to hire additional person to develop good tools. Third party contract, third party partnership, release some programming info to the public, these are just some of the methods to get what people want without expanding company staff. It seems that, Parallax does not want to spend more money or resources on propeller project. So why they invested in two full page adds in recent Circuit Cellar? These adds get me interested again in the propeller chip – for a few days. It looks like a waste of money.

I remember Parallax – Microchip ~15 years ago. Than, Parallax started to offer excellent tools for PIC controller from Microchip. I do not know anyone who used the org. Microchip tools. I think, Parallax made some good $$$ selling these tools. However, the real winner was Microchip, their PIC become hugely popular. And one should remember. the propeller tools are not free: they are nothing without the device or EVB kit, you have to buy them from Parallax.

Paul Baker

Farming out the IDE has it's own set of issues and we would have to find a proven and reputable company and a mutually agreed plan to impliment it. That Im aware of, to date no company which fits the bill has approached us with a plan to do this for a non-windows IDE.

There has been a release of programming information to the public, a python based programmer has been made based on it. Two of our customers have reverse engineered the bytecodes and have freely released this information. To date no one has used this information to create a non-windows IDE.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer

Parallax, Inc.

Post Edited (Paul Baker (Parallax)) : 6/10/2007 8:25:44 PM GMT

Chris Kraft

I don't think its fair to say the fact that no-one has used the reversed engineered information to build a non-windows IDE is proof that no-one wants one or is willing to do the work.

I want one, and I am willing to do the work, I just don't want to use reverse engineered information to do it.

The back cover of the latest "Nuts and Volts" shows an add that says "Tired of playing by someone else's rules" with a picture of something chained down to a sink with "no peek" and "I agree" on it. While at the same time, when people ask for information or code, the answer is no that you will not share that information. That seems, to me, to be a bit hypocritical.

Now don't get me wrong, its your product, your intellectual property. Your under no obligation to share anything with the public. I just wish you would come right out and say that "no, we will not share code, or any of the internal workings our our system, because we don't want to"

I am sure you have your reasons but blaming lack of staff, then saying open source will lead to all types of trouble, even though other organizations have done this with success, really just masks the truth about the situation.

I appreciate the fact that you guys released the code needed to program the Propeller chip its a great first step. I just wish you would release the code to the compiler, or publish enough specs so that the folks reverse engineering the compiler don't have to reverse engineer it but can know exactly how it works.

--Chris

Paul Baker

Chris Kraft said...

I just wish you would release the code to the compiler, or publish enough specs so that the folks reverse engineering the compiler don't have to reverse engineer it but can know exactly how it works.

--Chris

This is the key statement of your's Chris. To release enough specs would take months to do·properly, meanwhile other projects deemed the highest priority by Parallax would languish. Additionaly·if we release the code or the internals doc (which doesn't exist) we would have to spend further time supporting this released information. It's not that we are stating that it will absolutely never happen, it's all a matter of priority, and right now it's not high enough for us to actively pursue.

We have no easy answer to provide, thats why we came out with the categorical "We are not providing non-windows IDE or libraries" in another thread. Our definition of a·"long term" goal is·not the same as some of our customers, so that·statement was made·so I wouldn't have keep stating the same thing every couple of weeks. Im truely sorry, I know this is a disappointment to some of our customers.

I think you're misconstruing our ad, none of it indicated we were going to publish our code or provide exhaustive documentation on how our compiler works to enable someone to write thier·own IDE for the chip.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer

Parallax, Inc.

Post Edited (Paul Baker (Parallax)) : 6/11/2007 2:41:08 AM GMT

Chris Kraft

Thanks for the reply. I completely understand the time involved in documentation. I always put the time into my project schedules but it seems to be the first thing we start to take time out of when we start falling behind.

I don't mean to keep badgering you guys on this topic, its just that I think the Propeller is the most innovative, exciting product out there right now in this sector. The combination of features and power in such a small package is pretty unique.

Really the only thing missing is the fact that I cannot build the code on an alternative platform. I guess I will have to give the reverse engineered compiler a try.

Anyway, all I ask is that you consider opening up, eventually, enough details so that the community can create their own compiler that is mostly compatible with the official compiler, and that we can do that without having to reverse engineer anything.

I know others may want a full IDE but really all I want is a compiler.

--Chris

Paul Baker

I understand Chris, something that is on the horizon (please don't ask for a timeline because there isn't one yet) is a command line interface to the IDE so you can get a backdoor to the compiler. This should be more amenable to people wraping it up in wine and producing a library and thier own GUI (not sure how it would work, but Ive been told it's possible).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer

Parallax, Inc.

potatohead

Why not just run the prop tool in a VM on Linux or the Apple?

I've been using the virtual machines for what I do. (Pre-sales demo, support, testing....) The VMWare player is free for all OSes. You could load win2k, the prop tool, and share files via SAMBA, NFS, FTP, etc... Once you've built up your VM image, just put it where it is needed and build from there. That does technically mean some win32 licenses, but that's it. Everything else is a no-brainer.

I've not had a chance to try this, but the VMware player would very likely handle the serial communications issues where WINE currently is a problem, due to how the prop tool communicates. Worst case, one builds an EEPROM image and uses one of the loader programs, that does run on other OSes, to put it on the chip.

Surely you have some win2K licenses left over from the older computers that could be used to get a VM up and running. If so, then it's just some work loading and configuring and you should be there.

Chris Kraft

potatohead said...
Why not just run the prop tool in a VM on Linux or the Apple?

I have Paralles with XP installed in it. I have bootcamp with Vista installed in it. I have extra machines around that have Windows on them. I keep a old Laptop around with Windows 98 for some stuff that doesn't even work on XP or 2000.

There are two goals I have:
- To be able to use my own programming tools for writing code. I use TextMate for most of my coding on my OS X machine and I prefer that to anything out there.

- To have a tool that is open source so that, if the vendor decides they don't want to support an old platform, I can still program the chips I have around because I have the code and can always make the changes I need.

Its about choice and freedom, not about just making it work. If that was the case I wouldn't have purchased the Propeller starter kit, the Hydra and several prop chips. Yes I can program them, develop for them, I just would rather do that nativly on my OS X machine.

There are other advantages too. For example I can program my Atmel chips in C because I have gcc setup to cross-compile and create binaries for me. If the compiler was open-sourced, I realize thats not likly to happen any time soon, then who knows what might happen. Maybe someone would build a C compiler, or come up with a way of doing something no-one expected.

Also I haven't had the best luck programming the Hydra using XP running in Parallels. I have to un-plug and re-plug the USB connection about 20 times before it will recognize that the chip is there. Oddly enough I don't remember it being a huge pain with the starter kit. I just upgraded to Parallels 3 so I am hoping that this problem goes away. If I could consistently program the Hydra using Parallels I would be much happier.

--Chris
·

potatohead

Understandable. I live in a half open, half closed software world. For those few pieces of software, I must use that are closed, I've licensed them inside a VM. This does give me platform freedom, and a great deal of choice and longevity, in that the VM can be supported long after the fact.

One other thing about VM's under Linux is they can be remoted to where ever they are needed, via the X window system. I realize this is somewhat problematic with this particular scenario, but it's totally viable for most others where you've got input and stored data. One VM, served to any number of users, running on a nice solid box will serve for a very long time without huge expense in either licenses or support, once configured.

Again I understand though. Respect it too. No worries there. IMHO, this will happen for the prop, but it's just gonna take time.

gardener

Hi Chris and Mr. Potato,
You guys hit a bulls eye, the name of the game is freedom. My company was held hostage by chipmaker, also they had to spend a lot of money to change the cpu. I will not seriously touch the chip without certain level of security. Historically only the chips with open architecture survive. There is nothing like prop., but ARM running at 280 MIPS for $11 ea. can go a long way.

Kevin Wood

It's unfortunate that you had a problem with a chip manufacturer, but open source _anything_ doesn't necessarily solve the problem. Yes, it can help, but unless you are in control of _every_ aspect of your production, from mining your own silicon to collecting the cash, there is always some point in the supply chain that can fail.

I think there are a couple of things that should (my opinion) happen here, that would go a long way to resolving some of these concerns:

1. Parallax should come up with a viable long-term strategy to build cross-platform tools.

2. People should stop insisting on the tools being open sourced.

Honestly, the majority of the end users aren't concerned about having their Stamp/SX/Propeller IDE open sourced. They just want to program the chips on their favorite platform.

gardener

Kevin,
I think, I was misunderstood, 8051 or ARM does not mean open source. To “open architecture” I meant; instruction set & code are known, also programing protocols are published, so third party can write & make alternative tools. (It would be nice to have a second source for hardware too, but that is not up to Parallax). Prop is interesting piece of hardware, but it is classical example of closed architecture; limited information available, limited & single source for tools, single source for hardware. Using this product for a serious project is like asking for trouble. I do not think Parallax is serious marketing this device. And BTW the best ARM tools are not open source.

Forrest

My 2 cents - most CPU's on the market ARE single sourced, and many of them DON'T offer free programming tools.

Mike Green

As with a lot of things, the issues are not simple. In the case of some potential users, there are company requirements for multiple sourcing of parts and, unless the potential volume is huge, Parallax is simply not going to provide that. If the volume were large enough, I'm sure Parallax could make arrangements with another chip foundry to be able to produce the device and put copies of the necessary information in escrow so that the customer could be assured of an adequate supply if something happened to Parallax.

It would be nice for the Spin bytecodes to be well documented. The native instruction set and the download protocols are well enough documented for anyone with the resources to provide an assembler or compiler. That's already been done and there's a commercial C compiler for the modified native instruction set (large memory model) that will be available around the end of the year.

There's a fair enough document on the Spin bytecode instructions that a simulator has been written that works pretty well. It's been done by a 3rd party and not verified by Chip, but should be enough to produce some kind of compiler whether for Spin or C or some other language. Again, someone with the resources could struggle through the work needed to verify the bytecodes and produce a compiler.

Most of us can manage with the existing Windows XP-only tool until there are more tools available. I (and others) can run Parallels Desktop with Windows XP. Linux is a bit tougher, but CodeWeaver's CrossOver Mac will run the compiler and most of the IDE and I'm pretty sure their Linux version will work as well. If someone were interested enough to work with CodeWeaver and Parallax to isolate the bugs and fix both so they work together, that will take care of the cross support problem. The downloading part may not work because of the tight serial timing, but the Python downloader works fine with Linux and the Mac OS, so that's merely an inconvenience.

gardener

I'm not sure if I'm missing something. I've checked one more time Parallax publications & I couldn't find any information about the opcodes or programming protocol. I would appreciate if someone drop me a link to this data. At this moment it is a very proprietary processor & it's just too much risk to recommend this device to my customer. I'll regularly check Parallax site if anything changes. Saying that, I think that the prop chip is a great educational tool & an excellent toy for hobbyists. If I'm not mistaken, Parallax' primarily targets this market.

Harrison.

Opcodes can be easily found in the Propeller Manual (www.parallax.com/dl/docs/prod/prop/WebPM-v1.01.pdf). The assembly instruction set is extremely well documented in that manual (extremely well as in each bit in each instruction is explained).

If you are asking about the Spin bytecodes then you will have to search around. Asterik documented all the bytecodes that he could find, unfortunately the link somehow died so you will have to beg him to repost them.

The programing protocol is not actually documented in plain text. There is a short description on bits of it written by Chip Gracey, which is supposedly all that exists document wise. The delphi communications unit that the Propeller Tool uses has been posted here many times so you can use that if you wish to implement your own programming tool. The programming protocol has also been used in a Python script.

There is no reason to say that the architecture is so closed that you cannot write your own tools. The instruction set is open enough for someone to write a cross platform assembler in java (its been done). The programming protocol is simple enough to be written in other languages (its been done). So you can use other platforms to program in assembly and load the propeller without touching any of Parallax's tools.

Harrison

ehdyn

I'm wanting to work with the propeller chip using OSX(non-int*l). Am I to understand that VMware has a player that will allow me to do this?
Are there any limitations I would encounter?
Please let me know the best way to go about doing this.

Thanks,
Ryan

SSteve

VMWare Fusion only works on Intel-based Macs.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
OS-X: because making Unix user-friendly was easier than debugging Windows

links:
My band's website
Our album on the iTunes Music Store

Mike Green

There is a PC emulator called Guest PC for PowerPC Macs. It will run Windows XP albeit slowly and I've done some Propeller and Stamp development using it. I was much happier when I got an Intel Mac and Parallels Desktop.

BradC

Just a headsup on the VM issue. I've been using Virtualbox on linux and MacOS with an XP VM and it's behaving flawlessly with the propeller IDE including USB downloads. No hitches, no hassles. I've never had to re-plug the board or do anything fussy, I just installed the FTDI driver in XP (well, the proptool installer did it for me) and plugged the demo board in. Might be worth a try to those running linux or MacOS (intel) anyway.

Brad

JT Cook

I think the thing a lot of people are not realising is that the Prop is just a year out on the market and is a completely new design. Unlike an ARM or PIC where basically the same design has been around for years, it could take awhile for a freeware 3rd party open source compiler to pop up.
·