Parallax Propeller on Usenet (newsgroups)
Peter Jakacki
Posts: 10,193
Many years ago when the internet was still fairly new I would spend a bit of time on Usenet groups such as comp.arch.embedded or even comp.lang.forth although the later was always bogged down in the language itself rather than the use of it. A few weeks ago I looked through these groups and could not find one mention of the Propeller chip so I put my "two bobs" worth in. Interesting to see how people react and even those who are "informed" have many misconceptions. If you are interested then have a look at these groups, I would love to hear your opinion of the posts and also how we could do better in promoting the Propeller as we are want to do, especially when someone raises a problem and you think "the Prop could do that in a cinch"
BTW, the traditional way of accessing Usenet is via a "newsreader" program but you can also access these groups via Google groups if you are so inclined (don't let Usegroupies find out though, tsk tsk)..
https://groups.google.com/forum/?fromgroups#!forum/comp.arch.embedded
https://groups.google.com/forum/?fromgroups#!forum/comp.lang.forth
BTW, the traditional way of accessing Usenet is via a "newsreader" program but you can also access these groups via Google groups if you are so inclined (don't let Usegroupies find out though, tsk tsk)..
https://groups.google.com/forum/?fromgroups#!forum/comp.arch.embedded
https://groups.google.com/forum/?fromgroups#!forum/comp.lang.forth
Comments
From a specs-only perspective, the Prop looks like an odd duck. I admit to having difficulty getting my head around its core concepts. But part of that difficulty was, at the time, a dearth of examples in my field (mainly robotics) where I could visualize the advantages. Now, of course, there's the PropBOE and Andy's new Propeller-based robotics texts. And having written some articles for publication where I used the Propeller, it was nice not having to always worry about running out of external interrupts or internal timers. But until you actually build a project using these features, they're mostly nebulous marketing blurbs.
On groups like c.a.e. there's always one better microcontroller out here. Engineers settle into their favorites, then go to great lengths to defend their choice, and explain why yours is wrong. They even do this in Usenet groups and forums where their employer is not even looking!
I think you did a fine job explaining what the chip could do. Especially now on the cusp of the P2 being readied, the best defense is a good offense, and that's plenty of really neat examples and projects. Some of the success of these will depend on the features and capability of any development board Parallax (or others) come up with. I realize it's somewhat heretical, but one ideal board would be Arduino compatible, capable of driving and using both 3.3V and 5V logic (the 5V logic so it'll plug into existing shields). There are others, like a P2 version of the new Propeller demo board. Someone thought to move the chip out of the center of the breadboard area, and now I can actually use it for something!
I have just read the thread you pointed at.... very interesting perspectives indeed. Clearly some very clever fellows on that site.
The whole issue of mutiple cores and multiple threads and interrupts I find fascinating as that is dear to my heart. Although I do not consider myself an expert programmer, I have spent a lot of time on the whole multiple-thread-per-cog issue, and have managed to cobble together a very credible (at least in my opinion) co-operative multi threading scheduler addressing many of the issues they raise. The supposition of "one-thread-per-cog" is no longer valid, and I wish I could give those folks some actual practical insight on this.
Being that I am not very PC literate, and don't know my way around forums very well, I wonder how I could help them be better informed on the per-cog-multi-tasking advantages of the propeller. Does one need to become a member of that forum to post a response ?
Cheers,
Peter (pjv)
The most striking thing to me was the tendency to get off in the weeds for any reason. Props do great video. Doing that is easy, robust, etc... The example being discussed is at the upper limit for Propeller video capability, and as such it's got some details that would normally be ignored in a macro level discussion of the merits of the chip design, but instead became the focus... Weird.
Truth is P1 is being pushed well outside it's original design scope, and those of us "in the bubble" have a mastery of it that isn't common, and is a significant step up from what we might expect somebody to understand coming in new. Big barrier there, if you ask me. Ideally P2 will land closer to the sweet spot for more general applicability, meaning somebody can put the thing to use with far less mastery thus lowering that barrier considerably. We shall see.
BTW: That same video terminal implemented on TV, or a nice composite monitor works in one COG very nicely due to the lower refresh rates common to TV and older CGA / EGA / VGA display systems... I've an older amber screen monitor that is just killer for text and basic graphics on Propeller, easily rendering 80 - 100 columns of text, sharp, easy to read, etc... P2 video will be a very significant step up, turning an ordinary VGA capable display into something as easy and flexible as that monochrome NTSC composite display is today. Give it a little time.
Re: Interrupts
Most of that discussion was scale blind. Again, I saw folks not really willing to discuss anything new with any real effort. If scale were properly incorporated into that discussion, half the "but it can't... or won't..." comments were simply irrelevant. Tough crowd.
Getting them to do that work is paramount, otherwise all I see is slinging around of things, grunts and suppositions that are entertainining, but little else.
A strong alternative would be a few, "can you do this?" challenges, again with code and dev notes so that the value can be clearly seen. One disadvantage of this would be experienced users of technology will respond with their own challenges and that could get into the weeds on scale / applicability axis, but there will again be some investigation on, "how did that happen again?", which is what is needed to get people past the obvious and well trodden misconceptions / exceptions.
These are very common new technology problems and the primary barrier is always to get people to do the work to actually understand where the value comes from. Once they do that, then exploring use cases gets them to picture themselves employing the technology, which then gets us new users to talk to! In general, this is a very strong reason why there are so many "legacy" or "compatable with" kinds of investments that often detract from the core new technology! This is done to get people to do the work thinking they are simply expanding on some work already done, which is arguably the same work anyway, but it will be seen as minor as opposed to "new" which is significant as any new technology adoption process is.
I for one would like to see more responses from experienced Propeller users. The thread appears to have gone off on too many tangents, leaving several issues hanging.