This doesn't change the huge red flag that should give prospective customers pause.
The device will still operate only as long as Glowforge's servers stay up. If there are internet outages, or the company goes out of business, the device will no longer function as a laser cutter.
The firmware is a very small piece of the puzzle. The cloud service is being portrayed as doing all of the heavy lifting. That's image processing, CAM, toolpath creation, and motion optimization/lookahead. Once the motion is optimized so that the motion system will move as fast as possible within its acceleration limits, the result is a list of explicit instructions for the motion system and laser. Accelerate at rate A for T seconds. After X steps, pulse the laser at a power level of P.
Thus, the firmware simply processes these instructions, and actuates the motors and lasers. That's not to say the firmware is trivial. But in comparison to the overall codebase, it is a very small chunk of the complete CNC system.
I'm not arguing that Glowforge is under any obligation to open source anything. But this is a fairly small concession that does not address the main concern that most people have voiced, the inability to run the complete system on your local machine.
These folks didn't invent the CNC ecosystem. They didn't even invent the realatively inexpensive laser cutter. If the firmware is GPL, I'm fairly certain plenty of existing toolkits will get drivers for the it in short order.
But then you still have a totally different product.
If I buy hardware and software, and can execute the software on my own systems, I can run that combination for as long as I like exactly like on day 1, and as the product was advertised.
If I buy hardware and only borrow the software, I have a big problem the moment the access to the software goes away. The open source release of the firmware is a good step to ensure I might get out of the story with more than a paperweight, but won't let me keep using the product I once bought and wanted to use.
The maker world already has lots and lots of alternatives for existing laser printers, homebrew things etc. The toolchains are getting more robust by the day - heck the CUPS just added compatibility layers for gcode based "printers" (laser cutters, 3d printers, etc). This simplifies even further the common use case for low-end laser cutters - turning an illustrator or pdf file into a cut device.
Our competitors in the 3D printing space did the exact same thing, they open sourced their "client". But all of the real processing is on the server, so it's really inconsequential. In my opinion it's just a way to market your product as "open source" even though 90% of it is proprietary.
The hackers/makers that they target with this branding should see through it pretty easily. I'm not sure if there is a segment of users who would be fooled by it.
Completely agree. But devices should work without a connection to a cloud, even if that cloud was open source. In this case, I see no reason for performing logic in the cloud, except that it makes DRM easier: You can't copy code you don't have.
You can continue to use proprietary hardware that is no longer supported by the manufacturer, as long as it is self contained. I think that SaaS is an unfortunate misstep, and making hardware depend on SaaS is even worse.
Given the firmware source, it generally isn't too hard to port an existing open-source motion control software (which there are plenty of) to run on this machine. Electronics schematics would help but the firmware source is enough.
With source code escrow of their services, you can presumably use dns redirection of their site and the open source version of their software to run your own local instance of the stack.
Truly Open Source doesn't mean runs without any other components, it just means you can replace any of the components yourself.
This doesn't change the huge red flag that should give prospective customers pause.
No it does not. And among the people who care, they're just digging the hole deeper. It felt evasive to ignore the question about the cloud over and over while conspicuously answering other questions, and responding with a disingenuous "thank you" while throwing out a half-measure (that really doesn't help that much) makes it worse.
There is one response to this that will work: Oh we get it. It was our business model but its feeling like more rotten DRM to our customers. We have to change it.
Might I suggest: "We have this cloud service that we think adds tremendous value to the product and is offered for this amazing low price, but if not, here is exactly how to use the product as a standalone ordinary laser machine if you so choose."
It would be so much more friendly to have a functional UI hosted locally on the device and to serve updates and optionally populate additional functionality from the web. If your web offering is good enough, most people would have no problem with this setup; it satisfies everyone's needs. If they don't want it to connect out, they just block external connections from this device.
Comments
This doesn't change the huge red flag that should give prospective customers pause.
The device will still operate only as long as Glowforge's servers stay up. If there are internet outages, or the company goes out of business, the device will no longer function as a laser cutter.
The firmware is a very small piece of the puzzle. The cloud service is being portrayed as doing all of the heavy lifting. That's image processing, CAM, toolpath creation, and motion optimization/lookahead. Once the motion is optimized so that the motion system will move as fast as possible within its acceleration limits, the result is a list of explicit instructions for the motion system and laser. Accelerate at rate A for T seconds. After X steps, pulse the laser at a power level of P.
Thus, the firmware simply processes these instructions, and actuates the motors and lasers. That's not to say the firmware is trivial. But in comparison to the overall codebase, it is a very small chunk of the complete CNC system.
I'm not arguing that Glowforge is under any obligation to open source anything. But this is a fairly small concession that does not address the main concern that most people have voiced, the inability to run the complete system on your local machine.
These folks didn't invent the CNC ecosystem. They didn't even invent the realatively inexpensive laser cutter. If the firmware is GPL, I'm fairly certain plenty of existing toolkits will get drivers for the it in short order.
But then you still have a totally different product.
If I buy hardware and software, and can execute the software on my own systems, I can run that combination for as long as I like exactly like on day 1, and as the product was advertised.
If I buy hardware and only borrow the software, I have a big problem the moment the access to the software goes away. The open source release of the firmware is a good step to ensure I might get out of the story with more than a paperweight, but won't let me keep using the product I once bought and wanted to use.
CAD/CAM software is not cheap either. When their cloud suite goes dark, you'll have to cobble together your own toolchain.
The maker world already has lots and lots of alternatives for existing laser printers, homebrew things etc. The toolchains are getting more robust by the day - heck the CUPS just added compatibility layers for gcode based "printers" (laser cutters, 3d printers, etc). This simplifies even further the common use case for low-end laser cutters - turning an illustrator or pdf file into a cut device.
Yes!
Our competitors in the 3D printing space did the exact same thing, they open sourced their "client". But all of the real processing is on the server, so it's really inconsequential. In my opinion it's just a way to market your product as "open source" even though 90% of it is proprietary.
The hackers/makers that they target with this branding should see through it pretty easily. I'm not sure if there is a segment of users who would be fooled by it.
Completely agree. But devices should work without a connection to a cloud, even if that cloud was open source. In this case, I see no reason for performing logic in the cloud, except that it makes DRM easier: You can't copy code you don't have.
You can continue to use proprietary hardware that is no longer supported by the manufacturer, as long as it is self contained. I think that SaaS is an unfortunate misstep, and making hardware depend on SaaS is even worse.
Given the firmware source, it generally isn't too hard to port an existing open-source motion control software (which there are plenty of) to run on this machine. Electronics schematics would help but the firmware source is enough.
"The firmware is a very small piece of the puzzle."
Probably right. What's needed is source code escrow for the cloud service, so if Glowforge tanks, the code goes open source.
This won't help you if your internet connection is down.
With source code escrow of their services, you can presumably use dns redirection of their site and the open source version of their software to run your own local instance of the stack.
Truly Open Source doesn't mean runs without any other components, it just means you can replace any of the components yourself.
Sure it will. GCE just runs containers.
That's a good idea in theory, but what about maintenance and related devops? It's not enough to assume the code would be reliable on its own.
What kind of organization would provide this service? And who pays for the infrastructure?
No it does not. And among the people who care, they're just digging the hole deeper. It felt evasive to ignore the question about the cloud over and over while conspicuously answering other questions, and responding with a disingenuous "thank you" while throwing out a half-measure (that really doesn't help that much) makes it worse.
There is one response to this that will work: Oh we get it. It was our business model but its feeling like more rotten DRM to our customers. We have to change it.
Might I suggest: "We have this cloud service that we think adds tremendous value to the product and is offered for this amazing low price, but if not, here is exactly how to use the product as a standalone ordinary laser machine if you so choose."
Its not too late.
It would be so much more friendly to have a functional UI hosted locally on the device and to serve updates and optionally populate additional functionality from the web. If your web offering is good enough, most people would have no problem with this setup; it satisfies everyone's needs. If they don't want it to connect out, they just block external connections from this device.
way more expensive to put all that processing power inside each cutter though
The founder has said it would be on the order of a few hundred dollars, which is less than the cost of any of the add on packages.