3 ms·
What I don't understand is why John Deer with their incredibly expensive tractors doesn't just use Windows CE or any other proprietary OS. They can afford it an
by jeroenhd 4y ago
What I don't understand is why John Deer with their incredibly expensive tractors doesn't just use Windows CE or any other proprietary OS. They can afford it and it's not like Linux development is any easier than Windows development.
Offtopic: is it true what he says about Apple pairing their sleep sensor chip to their security signatures? That would be such a scumbag move.
- klysm 4y agoI find Linux development significantly easier due to experience and I think that feeling could be shared by a lot of folks. There’s also value in working with open tools because you can take that knowledge and use it anywhere. If you become knowledgeable in enterprisey stuff you can get stuck doing enterprisey things
- daemin 4y agoYou're saying the same thing but with two different outcomes, as if by virtue of using open source tools you're not going to get stuck working at companies that also only use open source tools. Become experienced in open source tools so that you can work in companies that use open source tools. Become experienced in closed source tools so that you can work in companies that use closed source tools. It's the same thing so really it comes down to a matter of preference and experience. I am pretty much a Windows developer as that's generally been what I've done, and have very little experience writing code for Linux and other operating systems. I could do it in a pinch, and with some time I could adapt and become a more open source based developer, but that is not the environment where I am most productive.
- klysm 4y agoCompanies that used closed source tools can also use open source tools and more importantly _any_ company can use open source tools while the same is not true of closed source tools. I don't think it's really symmetric like you say it is. I do agree it's up to preference and experience though at the end of the day. Having experience in closed source tools can be very valuable because the cost of that knowledge is higher. Not everybody can just tinker with it.
- 5e92cb50239222b 4y agoIf you pick hardware carefully (or write your own drivers, or pay someone else to do it, which is entirely possible in their case), FreeBSD would be an excellent option. I wonder why it isn't more popular among these kinds of manufacturers. (Everyone knows of the two or three major exceptions, which just proves the point.) IMHO it will get there if enough GPL enforcement cases pile up.
- calvinmorrison 4y agoYou'd think some embedded proprietary RTOS would be ideal
- mardifoufs 4y agoI don't think the linux part runs the machinery itself.
- throwaway894345 4y agoNo, it talks to the stuff that runs the machinery which was, and likely still is, JDOS (Deere’s home-grown RTOS). There was some interest in moving to a proprietary OS, but I don’t know if that was ever manifest. From what I remember, the only Linux system in ag equipment was the touch screen on the Command Arm, but maybe the guidance stuff also ran Linux?
- calvinmorrison 4y agoNot even proprietary, plenty of RTOS options that are MIT
- CamperBob2 4y agoThere's no longer any such thing as Windows CE. Closest thing would be LTSC IoT, which is still Microsoft Spyware Lite(tm). If they want a heavyweight modern OS that someone else maintains and that Deere, and not Microsoft, can control, the only real choice is Linux.
- throwaway894345 4y agoI was wondering the same. I might be misremembering, but I think early versions of the touch screen used Windows CE, and development on the new touchscreen (the only instance of Linux I’m aware of in John Deere’s ag lineup) started sometime around 2010 (depending on when you measure “started”), so Windows CE was still alive, but maybe there was some writing on the wall that it was being shuttered?
- cduzz 4y agoAs was stated earlier, FreeBSD (or I'll add netbsd as well) certainly is a full fledged OS that would allow free redistribution of all the work without having GPL encumbrance, while allowing any sort of "remote telemetry" desired. And it's free free free, if you ignore that developing for it is possibly more time consuming since you'll want to limit your ecosystem to non-GPL code. Nevertheless, if I were working through a product that involves "redistribution" of any of it, and I really didn't want to deal with any potential issues with compliance with the GNU requirements, I'd spec out that the whole product runs [free|net]BSD and runs only 3rd party apache/bsd/purchased licensed code/libraries. This isn't really even a difficult option.
- CamperBob2 4y agoI'd be worried that I'd have to spend half my time writing hardware drivers.
- cduzz 4y agoYou'd probably just spend 1/20th of your time looking at hardware BOMs. Freebsd has pretty good compatibility, and if you're buying in sufficient volumes or making / integrating boards from scratch you can just stick to the paved roads. It may not end up being quite as cheap from a BOM perspective, though.
- throwaway894345 4y agoI can’t remember clearly, but I think the prior version of the touch screen may have used Windows CE. I’m not sure why Deere pivoted to Linux, but it may have had better support for their custom boards or they might have seen the writing on the wall for Windows CE (according to Wikipedia, its final release was in 2013 while the new touch screen debuted in 2014 if memory serves). Deere is a big Windows shop and it was not fun trying to develop for Linux in a VMWare VM, and corporate IT seemed to have special prejudice for our unit for deviating from the standard Windows suite.
- bradfa 4y agoOr they could just use a permissively licensed OS, like one of the BSDs. There's no need to use a proprietary OS in order to avoid the GPL, pretty much all the BSDs shun the use of GPL code and they offer similar abilities to Linux. Generally, but have no direct knowledge of John Deere's situation, Linux is used because it's usually possible to have drivers which are already written for the various hardware which is desired in more embedded systems. Having to write or port something like a wi-fi driver to a different OS is very non-trivial but wi-fi is generally considered table-stakes now for lots of devices. So it saves significant cost and development time to choose Linux simply because you can get off the shelf drivers for your electrical design.
- larossmann 4y ago>Offtopic: is it true what he says about Apple pairing their sleep sensor chip to their security signatures? That would be such a scumbag move. Try replacing an angle sensor in a A2442. It won't work, even if it is from another Macbook. It has to have GSX run on it for it to work. What we do now when they're corroded is do our best to try and clean up the existing one. Maybe desolder the hall sensor, use a fine tipped iron like the Hakko 2032 w/ T30-KN tip to "file" the pins and pads as much as we can, and put it back on. Sometimes, it works. When it doesn't, we tell the customer "I'm sorry, no sleep for you. Only manual sleeping!" It sucks. The places with access to GSX, are the places not permitted to do component level board repair on machines with liquid damage. The places without access to GSX, are the places with the best ratings in the country for component level repair on machines with liquid damage. The world is a meme.
- nebula8804 4y agoOff topic but have you considered adding Chat-GPT added summaries to your videos as a TL;DW? There is a plugin called eightify.app that produces summaries. It just feel like sometimes you are limiting your audiences to the hard core nerds with your long winded rants.
- dmonitor 4y agoIncredibly off topic.
- larossmann 4y ago>Off topic but have you considered adding Chat-GPT added summaries to your videos as a TL;DW? There is a plugin called eightify.app that produces summaries. It just feel like sometimes you are limiting your audiences to the hard core nerds with your long winded rants. No. My channel was never started with the goal of having 100m subscribers. In my initial musings, I thought, MAX subscribership, BEST CASE scenario was 300 subscribers, since I estimated the people going down the LCD-cell-only/component level repair rabbithole with Apple products to be around 300 at the time 10 yrs ago. and that assumes 100% of them watch/subscribe to my channel, which is a lofty figure. getting to 1.7 million subscribers is ridiculous. This wasn't supposed to happen. A cursory look at my older videos, total lack of preparation, production quality, etc should make clear that the goal was never widespread viewership. but, if 1.7 million people ARE viewing, if I've blown out my initial best-case-scenario by 3 zeros and change.... why go out of my way to do it all differently? Why adapt my content to people who admit they don't want to watch anyway? It's not a movie, it's a 9 minute video... I totally get if someone doesn't want to watch a 9 minute video, but... I'm not making content for them. I wasn't trying to get big, I was just producing what I felt like producing.. if doing that got me to 1.7 mil, maybe being myself is the thing to keep doing.