Launch is defined as when you start shipping the product to paying customers, so you can record revenue. Given this, Glowforge has not launched. The units that are out in the wild are not the final unit, and they do not belong to the people using them. Those folks can still cancel their orders and get their money back (and I’m sure Glowforge would also want their beta/pre-release unit back as well.)
The Glowforge launches the moment someone takes delivery of a box containing the unit that they own. That’s when Glowforge can recognize revenue, and that’s when I would expect to see a firmware release.
I definitely don’t expect to get the source for the “real” firmware, because I notice that the original message carefully didn’t promise that (while at the same time carefully not calling attention to that fact).
The exact quote is:
When we launch Glowforge, we’ll also release a copy of the firmware under GPL.
Glowforge firmware is user-flashable, so you’ve got both an escape hatch (if something happens to us) and a platform to experiment with. If you buy it, it’s yours – you should be able to do what you want with it.
They say they will release “the” firmware. The only way to interpret that is that it’s the firmware that’s on the device. If it’s user-flashable, it’s probably user-readable as well (though not definitely; nobody has reported on the innards yet and are probably prohibited from doing so by NDA.)
I expect the firmware to contain code that controls the servos, power supply, tube, lights, button, cameras, fans, sensors and switches, etc. at a low level. Beyond that I expect there to be enough code to initialize the configuration (get on the network), connect to the network after initialization, and register with the cloud service to receive commands. That should be all anyone needs to start working on a gcode-compatible layer. We’ll know more about how easy or hard this will be once people actually start receiving their units and we find out what microcontroller is being used and what the architecture looks like. But I would expect such a thing to be available to people around the time that the first warranties start expiring.
That’s also (give or take a few weeks) when I would want to see a firmware release. Any time before that, you would have a long list of interesting caveats about whether the firmware properly controlled the hardware as released, whether it worked flawlessly with the current version of the cloud software etc.
I think that when firmware is released many of us will be surprised at just how barebones it is, and how much of what we thought was essential to realtime control of the machine can actually be stuffed up in the cloud.
Dan,
Thanks for the cool video and update on the progress. I will use my glowforge to make custom engraving for the wooden sunglasses and watches i make, since our CNC machine can not do fine details.
Honestly I think it doesn’t need to be much more than a bootloader so it can do a bit of POST diagnostics and connect to the cloud. After that, the cloud directives could take over. Maybe need some small kernel of primitives support that all the complex movement & control commands can ultimately be expressed in and processed by the machine.
I don’t think it’s going to be nearly as useful as the off-ramp people are thinking
I’ve been saying that all along. I don’t know why people are so keen on the FW being released as if there’s any other back end for it to connect to. It’ll POST, it’ll control the machine. But without software driving it, it’s not going to do anything. So it’s going to take a team of devs a while to write software to make it useful. Then you can change the FW to point to your private server (or some 3rd party server, I suppose).
I think the solution set to the off-ramp question is going to involve a new controller board that plugs in and takes over. Then it’s open source territory.
And then… what… just use local software I guess? Sure, that’s a way. But since it’s designed to run in the cloud, would be easiest to start there. One could then just have a RasPi Glowforge server in their house.
Yep. The GF’s decision to go cloud doesn’t make it “easier” for the person who has a bricked GF and wants to keep using it. Plugging in a board and then loading some software is going to be easier for more people than setting up a server in their house.
Alternately, figuring out their cloud destination and then altering your home router to redirect that to another service based cloud server would make it really easy (except for people who can’t figure out how to touch their routers )
You missed my point. The decision to put it in the cloud doesn’t make it easier. The fact that that’s what they did makes it easier for somebody to point it to their own cloud (server) in the FW with minimal changes, as opposed to telling it to look nowhere and having software try to address the firmware directly.
I"ve always loved how Engineers, Craftsmen and Artists are the ones who build the world. That’s why I’m so excited to get a Glowforge I love tools that enable me to iterate quicker and easier so my train of thought isn’t lost after too many house of waiting or work or failed attempts. This is problem solving at it’s finest! Yes it may be a temporary solution, but it is a solution none the less and I’m exited!
Glowforge