Phil's swipe (taken in good stride ) aside, there is the case for it right there.
Look at the meta task discussion. Meta task being discussion about making the things needed to do the thing needed, work.
There will be some merit in a system that will do things consistently, sans meta task issues. About the only real trouble will be getting code from a discussion onto it.
One can type it in, write to a file, copy paste serial...
Also note Peter has Tachyon there now. All he really needs to do is serial comms.
Just putting this out there for the, "what good is a self hosted development environment?" ongoing discussion.
The way I see it, there are three basic paths here:
One is a nice set of PC tools. Various OSES, open code, etc. Mainstream and by far the preferred path. This one is the most potent, few limits.
Blockly is an extreme in abstraction. Keep the number of details needed down and present tasks in a non text, graphical, visual way. The weight of visuals will limit project size and or complexity.
A self host would go the other way, also keeping details down, but doing so by presenting text with a minimum of environmental encumbrances and dependencies. Project size is a limit, but complexity may not be.
A variation on these could be done with a RasPi. Lately, on a startup I'm involved with, using the Pi and a preset OS image has been a nice bonus. The idea is there are any number of environments and networks out there. It's actually darn tough to deliver broad based support and focus on the core tech, whatever it is.
We could set up a Pi nicely, intended to be dev station, and build on that as a default option. Looks like blockly is on it already. Good.
Package up Blockly, C, SPIN, serial tools, loader, IDE, libraries and it's a one stop shop. Very capable too.
A Pi is more than enough for even serious and sizable projects.
On chip tools could be written to work via serial, and could be redirected to a local console, should one be available.
That's it, just wanted to share some thoughts and experiences.
Having RasPi devices in the field has already saved me a ton. Users can get going with those. Some stick with it, and it's fine. Cheap, fast enough, no BS.
Others have gone on to incorporate their PC and more complex environment too. When troubleshooting, I call up the Pi and run a baseline. Gets right at where trouble may be, and provides a consistent path to make progress while other setups are being brought online too.
Couple these things with the language and presentation discussion here and the product could be broadly useful, appropriate for kids as well as various adults, and pretty lean in terms of potential hazards along the way.
Comments
http://forums.parallax.com/discussion/165996/sd-card-speed#latest
Phil's swipe (taken in good stride ) aside, there is the case for it right there.
Look at the meta task discussion. Meta task being discussion about making the things needed to do the thing needed, work.
There will be some merit in a system that will do things consistently, sans meta task issues. About the only real trouble will be getting code from a discussion onto it.
One can type it in, write to a file, copy paste serial...
Also note Peter has Tachyon there now. All he really needs to do is serial comms.
Just putting this out there for the, "what good is a self hosted development environment?" ongoing discussion.
The way I see it, there are three basic paths here:
One is a nice set of PC tools. Various OSES, open code, etc. Mainstream and by far the preferred path. This one is the most potent, few limits.
Blockly is an extreme in abstraction. Keep the number of details needed down and present tasks in a non text, graphical, visual way. The weight of visuals will limit project size and or complexity.
A self host would go the other way, also keeping details down, but doing so by presenting text with a minimum of environmental encumbrances and dependencies. Project size is a limit, but complexity may not be.
A variation on these could be done with a RasPi. Lately, on a startup I'm involved with, using the Pi and a preset OS image has been a nice bonus. The idea is there are any number of environments and networks out there. It's actually darn tough to deliver broad based support and focus on the core tech, whatever it is.
We could set up a Pi nicely, intended to be dev station, and build on that as a default option. Looks like blockly is on it already. Good.
Package up Blockly, C, SPIN, serial tools, loader, IDE, libraries and it's a one stop shop. Very capable too.
A Pi is more than enough for even serious and sizable projects.
On chip tools could be written to work via serial, and could be redirected to a local console, should one be available.
That's it, just wanted to share some thoughts and experiences.
Having RasPi devices in the field has already saved me a ton. Users can get going with those. Some stick with it, and it's fine. Cheap, fast enough, no BS.
Others have gone on to incorporate their PC and more complex environment too. When troubleshooting, I call up the Pi and run a baseline. Gets right at where trouble may be, and provides a consistent path to make progress while other setups are being brought online too.
Couple these things with the language and presentation discussion here and the product could be broadly useful, appropriate for kids as well as various adults, and pretty lean in terms of potential hazards along the way.