Linux Propeller Tool
Ken Peterson
Posts: 806
I just ordered a Propeller Demo Kit, but I prefer not to use Windows. I did a quick search of this forum for linux, but I didn't happen to notice any results regarding the Propeller Tool. Does anyone know if a port to Linux is in the works yet?
Comments
Releasing it to the GPL if it were ported may sound like a good idea but it would include Parallax releasing all their secrets about things like how Spin is compiled, and other things like that.
While it would be great, and I hope for it as well, I don't think it would be as "highly popular" as you think. In the Linux users here, yes; but that number is small and slow-growing.
EDIT: I forgot to say- I myself would also be ecstatic if there was a Linux release here.
Post Edited (LarryHedgehog) : 3/28/2007 7:47:31 PM GMT
I understand that parallax is a business and that it may not make sense to devote limited resources to a small segment of their user base. If there are no plans then I'd like to have that info as part of my decision process when making new purchases - (maybe not a deal breaker but would be nice to know)
The stamp serial software works well with Wine but I didn't have any luck with the propeller software...
Post Edited (rePete) : 4/10/2007 12:32:30 AM GMT
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer
Parallax, Inc.
This would let us generate intermediate languages that
translate to spin and assembler. Even the ability to generate
code using external macros would be helpful.
Scott
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer
Parallax, Inc.
wait for the next version of the chip.
To say that Linux users represent a small fraction of the population suggests
a significant misunderstanding regarding Linux users and Linux itself. Only a totally
Windows-centric individual would make such a demeaning and misinformed comment.
I for one have totally dropped my Windows/XP partition and will not be upgrading to Vista.
There may not be alot of high quality games available but, unlike Windows, there's an
enormous variety of design and development packages/libraries covering graphics, threads,
security, etc. Linux users are far more typical of the demographic that would want to try
Parallax products.
For those with either little engineering experience and or just cold feet I suggest you try out the Ubuntu
version of Linux. You can pick up a live DVD/CD at Borders or Barnes/Noble news stands <end of advertisement>
Actually what I said was that Linux represents a small fraction of the parallax user base - this is an assumption - it's based on the fact that it is not easy/possible to program the Propeller from the Linux environment. And that Parallax does not devote significant resources to Stamp development on Linux but they do on Windows. I shouldn't pretend to understand the Parallax business model but it seems to me that they focus on Windows.
Actually I'm Linux user with no Windows partition at all. I would like to use the propeller but I can not program it from my PC and have to borrow my wifes computer (not as easy as it sounds). I'm trying to get it running under Wine - but haven't had any success yet...
If you are a professional engineer, who is not doing any mechanical design, then you are more likely to be using Linux on the job. I work with these people on a regular basis. (CAD, PDM, PLM, etc...)
A whole lot of research / sciences types are using either Linux or Mac though. There are a fair number of GIS types moving there as well. Unix has deep roots in these niches as these folks where early adopters of powerful workstations. Those ran UNIX first. Engineering, mechanical in particular, was the same way, but the need for integration and a strong push toward reducing costs and people dependancies (one engineer is as good as another means and methods) drove win32 development and adoption in ways not seen to the same degree in other niches.
The biggie with engineering, from a "why don't we see more Linux tools?" point of view happens to surround Microsoft Office and enterprise systems (ERP, MRP), that leverage it. Integrated business systems now involve engineering at a far closer level than used to be the case. Outsourcing and just distributed round the clock processes are also contributing to this in a big way. Linux is totally capable of these things. In fact, at a systems level, it's far more robust in general than the win32 platform is. However, the core software elements necessary for this to happen in an enterprise wide manner are lacking features on Linux at this time.
Because of these factors, nearly all the software that matters for many aspects of design and engineering, runs on Windows, or commercial Unixes. There is one solid MCAD port to Linux and that is the UGS CAD package. I'm actually gonna be setting up a mechanical engineering team with Linux and Mac computers for the first time ever this year! (I'll be doing systems work in tandem with the admin staff, training on the software and some consulting on processes, data management, etc...) Since SGI fell by the wayside, I've not dealt with a new installation of engineering tools on anything other than win32 machines in about 4 years. A whole lot of these guys want Linux, use it personally, but can't yet operate on the same level on the job with it.
I'm sure the number of active Linux users in the embedded space is higher. (A big part of why I'm here learning actually!) Fewer dependancies in general allow this to occur. It's a niche similar to the research, science and visualization one in this regard.
All told, targeting a win32 release at this time is still a very safe bet where end users are concerned. Parallax is doing nothing out of the norm here. I'm sure the win32 focus comes from the tool sets they use and the familiar environment keeping overhead low. If you practice smart computing, a win32 environment is not bad at all.
We need to put a nicely equipped Ubuntu machine in Chip's hands! Seems to me, Parallax could transition to using one of the nicer GUI toolkits that also runs on Linux to build their IDE. That would still be windows at first, but ports at some point would become a lot easier...
I don't think Parallax would be wise to GPL their software tool. The above mentioned library will be good enough to integrate into one of the existing Linux IDE's, or for a new one to be created. This will cramp the style of the GPL purists, but everybody else will get along just fine.
The last few years I have been evaluating the Linux distros to see if it is possible to get a divorce from this present shotgun marriage. There was never any honeymoon and no kids and no consummation though it did consume a great deal of my time and money (if only I could bill Bill). All my machines are at present dual booted to XP and one or more Linux distros such as Mepis, Ubuntu, etc. I can handle a few odd Windows programs with Wine and most everything that I need is now available under Linux. The drivers are great considering that many of them had to be reversed engineered but no thanks to the mainstream mentality that prevails.
What many of us engineering types need now is for application and driver software to be written so that it can easily be ported to Linux as well (or just to Linux). It is not uncommon to discover that other engineering firms are going the Linux way. Vista is blowing out the minimum resources required and 1GB RAM is only really the starting point (as was 64/128M for XP). Linux is now my preferred direction (for many many reasons) and I have very little reason to remain with Windows. My office app is OpenOffice and I use Mozilla Firefox and Thunderbird as I find these programs superior. Linux is convincing me more and more each day. BTW, the ISP I'm with hosts the latest Linux distros which I am able to download and try, it only takes a few minutes and the D/L usage is free.
Come on Parallax, we wouldn't even bother asking some of these other companies. You guys are trailblazers and you know once the trail is blazed that there are many that follow and Lo!, behold the new standard! Me too!, Me too!, they shout over each other.
Anyways, at least consider it and give Linux a look-in.
*Peter*
If one of you have the right stuff to implement a Linux version of Spin, I am sure Parallax is interested. But the main dilemma with Linux is that the main market for Parallax is among learners, not engineers.
Simply put, there are far more begineers with cash that enjoy using the Windows-based PC that is sitting in front of them. I've run into a similar barrier in teaching English in Taiwan. Though my expertise in language and teaching grow; I am still mostly teaching fundamentals to rank beginners. Once they get further along, they find their own way to pursue excellence.
On the other hand, I dearly love Linux for the ability to provide an important source for educating those of us who go beyond the basics. I certainly have learned most of what I know about the software for internet from Linux documents and the open source community. WindowsXP is quite obscure. I have no idea what all the updates have done to my computer or whether it is really that good.
Of course, if I really want to operate Linux; I have 40 gigabytes of emty harddisk that I can impliment as a dual boot system. Since upgrading to larger harddisks is so reasonable, that may provide you with a feasible option.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"If you want more fiber, eat the package.· Not enough?· Eat the manual."········
Well said .. I can't add anything to that ....
Kramer,
>>"the main market for Parallax is among learners, not engineers" << - Perhaps - but I think its is fair to say that with the Propeller it has already proven to be a lot more than that ... and has the ability to beat the pants off many of the bigger names out there..
In general:
I noted that the ImageCraft 'Propeller C' project is initially only going to be a windows deployment .... c'mon with Linux closely related to Unix and many Unix apps written in C and most real Linux/Unix users proficient in C ... surely it has to be a logical step to have a Linux Propeller Tool.
My Ideal world - [noparse][[/noparse]Dream] 1) a Linux Propeller Tool (C/Spin/Asm).... 2) an IP stack for the propeller !!! Currently looking at enc28j60 and IP/Stack is wrecking my head - only at initial stages though .... there is a C based IP stack for the PIC written by Microchip - wouldn't it be so much easier to port that to 'Propeller C' rather than convert to Spin/Asm ...[noparse][[/noparse]/Dream]
Quattro
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
'Necessity is the mother of invention'
My point is that it is possible to develop complex, high performance, easy to develop, install and run software that is multiplatform. It is no longer a technical issue, but a management one and one that will be very very hard to "crack" because you have to learn different tools. If your software will only run under Windows and you are the only one that supplies the software and you have a unique enough product with no real competition, then your potential customers will have to support Windows ... and the cycle continues. Apple has bypassed the issue by providing a platform that will run pretty much anything (as long as you spend the money for Windows too) including most Linux software.
http://marketshare.hitslink.com/report.aspx?qprid=2
Linux is so low it falls into the "Other" catagory (the source reports all Linux distros account for 0.57%).
We have exactly 1 person that works on the IDEs, he can either be working on improvements to the Windows versions or work on developing the IDE for other platforms. We receive more requests for addtional features than we do for other platforms (though the platform advocates are more vocal).·Which makes more sense: developing a handfull of new features that will work on 88.28% of OSs or an IDE that will work on 0.57% (assuming the IDE would work on all of the countless distros of Linux) and subsequently double the development time for all new features?
What you will (eventually) see from us is what we did with the BASIC Stamp IDE, we'll release a library for other people to develop thier own tool. But the effort requires coordination between two people whose time is tied up for a very long time. It may take until the new chip comes out before we can devote the resources to it.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer
Parallax, Inc.
Post Edited (Paul Baker (Parallax)) : 4/22/2007 8:29:51 PM GMT
I'm ok with the wait Paul. No worries on that score. On the what makes more sense question, it is worth considering you only need to target one Linux. Users of other Linuxes can either run that one, or modify the Linux they are using. No need for any Parallax support for anything other than the target Linux. Pick one and that's it. It's also worth noting that the cross-platform issue only need to be done once. After that, things progress as usual, with some small additional overhead for the multi-platform stuff that will be ongoing.
Multi-platform advocates are vocal because the return is extremely high for them! Again, the learning only has to happen once. After that it's all about choices and their return. Lots of people have time, they don't always have license dollars...
There is a fair amount of Linux running in IT departments. I know some folks, who run multiple win32 VM's on a single 64 bit Linux server. This, combined with applications like the X Window system and VNC, make for a very powerful and easily accessable computing environment. One can lose the KVM, administer the machine when doing low level tasks with the attached keyboard and mouse. Do the rest from a nice X desktop, where ever it makes sense! The VM's, once running, can be accessed via VNC, or RDP. One nicely configured machine (lots of RAM) can replace many of them for an overall savings. Server VM's can be moved pretty much at will, backed up, state preserved, etc... thus eliminating a lot of risk and manual configuration changes. Want to test out some new config, software, or perhaps just that potentially nasty update? Great! Save the state of the VM, do your thing and evaluate it. Sucks? Restore the old VM state and it never happened. This savings is huge. Contrast that with the usual, back up, apply changes, test, restore, etc... cycle.
None of this kind of thing is in the marketing numbers. Neither are the various firewalls, data sharing servers, web servers, application servers, mail, etc... A single well equipped machine can serve an application to many users, via the X window system, keeping administrative costs way down. (no distributed software configuration and installation) This kind of application serving can also keep data out of users hands, forcing them to use the application to manupulate it. (no direct access to shared files, everything on server, managed by application only.) Almost none of the people learning on Linux register there either, because they download.
IMHO, the core logic stuff is multi-platform almost by nature, leaving user interaction the sticking point. If it were me, I would build on win32 using one of the multi-platform tool kits avaliable. QT, Swing (though this one is annoying), GTK, etc...
Having passed that step, a full port then becomes a lot easier. The learning transition can be done one step at a time, keeping win32 the focus until the state of things evolves well enough to make Linux an option. Having done this, interested others with passion can help as easily as they can now because they can get the same dev tools with no worries.
Of the engineering and design people I work with, perhaps as many as 20 percent are active Linux users. A small few are able to use it professionally. The biggest niche is embedded / electrical. The smallest is mechanical and industrial design.
Like I wrote above, I'll be setting up an MCAD system, with CAM for a group for the first time this year. The company they work for, interestingly, is an all Apple shop! The big gap for them has been mechanical design software. This year, we will have full on, high end, mechanical design software running on both Mac OS and Linux. Because of hardware issues and some delay on the Apple side of things, 64 Bit Linux will be in use until Mac OS comes on line and is production ready.
The nut is cracking, but very slowly. If 64 bit computing is indicated, Linux has the lead right now. It's feature complete, stable and fast. 32 bit stuff will be win32 for a very long time to come because the software stack is entrenched very deeply. The sciences are adopting Linux quickly. Some of that will trickle down to those engineering groups where it makes sense, some of it won't, depending on the target computing needs.
I find the learning comments very interesting. A lot of these types know what Linux is and how to use it, largely because it's feature set, in terms of development and applications in general, cannot be beat on a cost basis! In the mid 90's, I did what everybody else, trying to learn did; namely, copy cracked software. Sometime around 2000, I switched to OSS for this and have never looked back. Essentially, I can grab most any hardware that matters, bring up an OSS environment and then do lots of stuff. Sure beats paying for licenses, tracking down student versions, sharing with others, etc... I copy nothing today. If it's cost is too high, I simply don't need to learn on it.
Coupla things that would be possible on Linux:
-The complete Parallax development environment. Purchase the hardware goodies, get the books and a bootable CD. Take essentially any recent vintage PC, boot the environment and find all the dev tools, documentation, supplementary applications all nicely installed, configured, etc... I don't know that this makes sense, given the sub-option below, but it is possible! I wonder if the robotics people wouldn't be all over this!
-License cost free distribution of app collections. Google for the Open CD, for an excellent example of how this works. This is a collection of the better OSS applications, compiled for win32 and packaged in a friendly way. I've demonstrated this at my local school. (It does have a Linux lab, but they are still not getting it...) Students, who need to perform class work electronically, simply load the Open CD (or perhaps the Parallax CD), onto their machine and have everything necessary to perform, distribute and collaborate on their work. I've had teachers request assignments in Microsoft formats. Of course, I went right down to the school and put the nix on that! Given the volume and overall quality of applications available for the download, there is zero reason for students of any kind to actually pay for anything to learn, given the body of OSS software that exists and it's powerful feature set.
Anyone building their skills on OSS is then free to use them, share them, distribute the work created by them and repurpose their hardware at will. I went through this discovery process and learning around 2000. Today, I can grab a machine out of the thrifties, combine that with a net connection and be doing serious work later that day. The cost is only my time and bandwidth, and perhaps a coupla blank CD's. The very worst case is having to purchase some minor bit of hardware to get OS support.
One biggie is not being forced to upgrade. Look at the coming Vista mess. The DRM stuff really messes with the low level task scheduling and consumes a lot of resources that would otherwise be freed for a lot of users. Nobody has a say in that however. OpenGL has been removed from the OS, replaced by a buggy and slow wrapper that essentially forces applications to render through Direct X, no matter what. Most of the MCAD software that matters is OpenGL. OpenGL is cross platform and powerful. It sees heavy use in the sciences for visualization, it is the only 3D renderer for Apple and Linux. I can't imagine that being an accident... It does make one wonder if features are not added to increase revenue instead of actually add value these days.
I will not ever purchase a Vista OS. I will, sadly, be running it on somebody elses dime though. The support issues are high and the return on that investment is pretty low, given the end result will be getting the same applications to run with few hassles on a more resource intensive environment. Ugh.
If somebody does this kind of thing in OSS land, it's forked and everybody goes on with their business with few worries. Worst case, bunch of work gets done to smooth out the bumps, then everybody goes on about their business. Those not wanting to deal with the hassle, simply don't, until such time as they are ready to do so. Later this year, we all will be buying either a Vista PC, or a blank PC. Getting an XP PC will no longer be an option to us.
Gonna run a lab?
Configure machines to boot from network. Install one server machine that contains the Parallax dev environment. Boot server, then boot workstations. Students store data on USB key, or other removable media. Make mistake? Reboot and no harm done, no distributed loading of applications and data, etc... (This is how my local High School runs it's Linux lab. LTSP is the project name)
I totally understand the resource issue. It carries a lot of weight and costs. I'm ok with what's happening right now, just thought if we are gonna talk Linux, we might as well really talk about it, that's all. Consider this post advocacy that might someday justify some efforts in that direction.
Professionally, I've migrated to nearly all OSS stuff, and run it on win32 fairly often. I author documents in Open Office, and will often include it with distributions intended for the end consumer. I run it on win32 because the engineering software runs there. That is the only reason however. Should things change, I've not lost my apps, other than the engineering software, and can continue using my skills on whatever hardware happens to be at hand.
The only programming I do on win32 happens to be for the prop. Everything else is on Linux because I've no license worries. I'm atypical where professionals are concerned, but I'm in a growing segment where people who are learning are concerned.
Post Edited (potatohead) : 4/22/2007 8:39:10 PM GMT
The best hope for IDEs on other platforms is if Intel based Macs and Linux allows an encapsulation of 386 assembly language code into a library. This way we can roll the existing compiler into a library, otherwise it would require Chip's involvement and that's not likely to happen until the next chip is out.
As he has already mentioned, Chip wants to eliminate the entire OS issue altogether and put the entire development tools on the Propeller itself. Whether this will happen on the next chip hasn't been determined yet.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer
Parallax, Inc.
Post Edited (Paul Baker (Parallax)) : 4/23/2007 2:46:03 AM GMT
The fly in the ointment is that the Spin interpretive code is proprietary to Parallax and there's no official documentation on it. We're already seeing early releases of an open-source assembler for the Propeller native instruction set. There will be a Windows-only C compiler for the Large Code Model and, once the Large Code Model is "finished" and stable, there could be other OS independent versions. There is a fairly large set of documentation on the Spin interpretive code, but it's been done independently, is incomplete, and, most importantly, is unverified and "unblessed" by Parallax. It may be some time as a result before any of us will see any 3rd party Spin byte code compiler for the Propeller. Given the resources available within Parallax, it will be some significant time before the library is made available.
Mike
My question is: why on earth develop an IDE and toolchain in a platform-specific language and tool, when there are plenty of cross-platform languages and GUI toolkits out there? It doesn't take significantly more development time and even if development and testing is done only on Windows, at least it leaves a door open.
-- Remy
Paul,
is the existing compiler written in 386 Assembler?
Mike,
where can i find this "fairly large set of documentation on the Spin interpretive code"? I can't find anything about that.
Andy
Look at the forum messages from Cliff Biffle and at his website here: www.cliff.biffle.org/software/propeller/
I have three OSes running here (on different PCs). MacOS, Win2k and Linux. I don't care that much what OS it is (well, I have my favorite) as long as the software running on it does its job. I (alomost) gave up trying to advocate my beloved OS.
I see more advantage if Parallax invests their manpower into new chips than waist (biased word!) it with OS-related tool-issues.
As soon as they have spare time (will that ever happen?) or earned enough to source the tools out, I'd be happy to use a Linux-version. But for now, I'm happy with the Win-version.
Nick
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Never use force, just go for a bigger hammer!
The DIY Digital-Readout for mills, lathes etc.:
YADRO
Yes it is.
These both represent our sentiments.
But like I said we will eventually release a library for those that want to develop thier own tool chain.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer
Parallax, Inc.
Post Edited (Paul Baker (Parallax)) : 4/23/2007 5:42:13 PM GMT
Until Microsoft came out with Vista, I would have said Linux tools would be nice but not essential. But if XP should disappear from the map entirely, there will be a lot of folks who refuse Vista's bloat and control and who will be taking long — and longing — looks at Linux and OS/X.
It's too bad Borland abandoned their Kylix project. It would have permitted a straight-across port from Windows/Delphi to Linux.
-Phil
Anyway, thanks for a great chip and a good tool, and yes, improving it has definitely more value than porting it.
-- Remy