This is going to sound really nit-picky, but adding that many Strings together when a char array and a single snprintf would work hurts my eyes a little.
Getting off topic, back to the Mesh Networks, I have been out of the loop for about 6 weeks.
Has there been any update on the ability for developers to use multiple Argons in one network (HA = High Availability Networks)? Anyone want to comment on this? Are they working yet? Where are we at with these, has anyone made an HA as an example?
Has anyone been able to get a few Argonās working with OpenThread? Until the HA networks are perfected, I really think OpenThread is a viable option (Not an easy option, working with Openthread is a huge pain, but definitely something to use our hardware on.)
I would perhaps be good to split this off into a separate thread⦠Iāve been thinking about this issue as well 'cause I want more flexibility than Particle is interested in offering and I donāt care for their cloud service.
Aside from ranting my plan is to hack the firmware so I can set-up an argon or boron as if it were a xenon, while retaining some internet-facing (wifi/cell) code, so it can make outbound connections. Then a mesh would have a primary gateway which the Particle cloud would see and it would have some number of āsecondary gatewaysā that happen to have Mesh and internet connectivity. I would then do internet communication by running a gatewaying app on the primary and secondary gateways. Iām mostly interested in MQTT, so I would most likely gateway the Mesh pub/sub with MQTT pub/sub somehow.
Anyway, aside from ranting this is mostly just a plan at this stage, although I have started to hack the firmware, so itās not a black box anymore.
So this is fun. I am in my planned 2:00pm zoom.us meeting with @will having never used video conferencing before, in what looks correct [redacted]
The meeting room says [redacted], but no Will. Since I canāt find a 1-800 number, not much I can do except look at the really messy whiteboard behind me for another 10 minutes.
Anyone like the animation
So I asked my Google iHome Mini cell phone app what the 1-800 number for Particle in San Francisco was and somehow a note popped up that I had an email about it on my personal account.
Wow, Google and the Nest products are really impressive. The meeting was cancelled at noon today.
So Tuesday at 5:30 pm meeting hasnāt gone so well. Now itās 5:46 pm and no @will. Lucky I am on Spring Break and not wasting my Teaching Prep time. I will try again this week.
Think about it. If I purchased 31 Raspberry Piās that only worked if I was expected to pay a monthly fee, the news would go viral, but here at the Particle forum people seem reasonably OK with purchasing hardware that barely does what it was advertised to do.
.
Speaking of advertising, every time I login to my Instagram I get these ads, reminding me of the 31 Mesh Devices I purchased that are sitting in my cupboard at school.
Hey Jeremy,
I canāt say Iām frustrated. Iām patiently looking forward to the future and checking on how things unfold.
However, I can sympathize with your feelings. Itās not cool to expect something to happen or behave a way and then find out is not happening.
I had no expectations about the Gen3 hardware, and I do not mean that in a bad way. When I made the pre-order I was excited to get something from Particle (because itās their products that paved my way into IoT and connected devices) and I was excited to get something futuristic, whatever that means. Mesh, bluetooth and many other features captured my attention and I am still happy with what I got.
No bluetooth API for now? Well, not a big rock on my way since I would like to play with it but I can wait for it. Itās coming, itās in the roadmap and Iāll be happy when I get it. Iāll be happier of course if I would get it tomorrow vs getting it in 5 months, but hey, Iām ok with that. Of course, I will be sad if it takes 10 years but for now, it does not look like that.
Iām sad that you have experienced things this way, and that the kids you were about to teach this tech would not get it because of all the frustration you went through.
Anyway, Iām pretty sure of this:
if there is something I could do to help you de-escalate the frustration and let your students get to play with your Gen3 hardware, please let me know.
Not only that, write about it here in the forums and Iām one thousand percent sure that the community and Particle will try to help you.
I read once that optimistic brains get wired differently than pessimistic ones. Now you know a little bit more on how mine got wired some time ago⦠Still, I wish mine was wired like this.
which both took a ridiculous amount of time to make and also were very frustrating as nothing ever works the first time, or the tenth time for that matter.
Thanks for your post @gusgonnet. It is a bit of validation about what it must be like for most Mesh hardware owners who have a more flexible timeline than my perceived one.
Iām aware of those qualities from reading your posts around the community. I learn a lot from them, and Iām sure Iām not the only one. I guess not only your students learn a lot from you.
Thank you for mentioning your teaching sites, I want to learn a bit of tensorflow.js so Iāll check them out. Nice choice of yellow, BTW. Unforgettable
The almost daily updated github version of the same website is blue, I used to have them both the same color and would wonder why my code didnāt work, when I was checking the wrong website! Just checked and I have changed the repository 5,400 times (commits).
Some people might wonder why I mention Javascript or Machine Learning in a Mesh Devices blog.
**Imagine, almost unlimited sensors and actuators interacting using Particle Mesh devices also connected through fast Javascript internet websockets to Raspberry Piās, Cell Phones or Cloud Machine Learning software, continuously learning and reacting to the massive flow of data in real time from: your body, your car, your home, your farm, your business. Get the picture! Thatās like year 2025 concepts. **
For most of the above scenario I have a working proof of concept, I just need the Mesh Devices connecting.
If you want to learn Javascript or Machine Learning, last month was definitely the time to start. Tensoflowjs just reached version 1.0.0 which means breaking changes should be minimal for a few years (months ). Breaking changes mean everything done till then is basically useless for teaching. It has happened twice, since I started using the original Google Tensorflowjs called Deeplearnjs. (Yes, changing the entire name of the product is considered a breaking change.)
Here is a link to my, as yet unfinished, javascript online, fully interactive, (means you can change the code and see what it does on the page!) tutorial about javascript basics
Just checking in. Itās been a couple busy months for me, too. (Boy, when you miss a little here, you miss a LOT!
@Rocksetta, than you for your hints and pages and articles about your efforts along the way. Theyāve certainly helped me on my learning curve! Iām sure there are many ohers who have benefited by reading yoru posts as well.
I didnāt get the USB dongle, but rather a small board with 2x5 IO pins, but then read the fine print to see that my Pi Zero W wouldnāt work (that Iād need an ethernet conenction in order to plumb the wireless drivers. I did order a new B model and kit, but thatās when I ran out of time, and building in OpenThread gateway on a Pi as my home gateway is stalled until a few other projects are completed. (And work has been claiming more time than usuall in the past couple months.)
I think you still have my personal email. Iād be interested to hear more about your progress in the term ahead, if you arenāt going to be posting about it here.
When I get back to this, Iāll post more about my progress.
Current status: I have three Xenons, one on an Ether dock, reporting basic data, but I got stalled trying to get local publish/subscribe working. Thatās where I started trying to get a RPi to use an 802.15.4 dongle as a wireless sniffer, to look for the mesh networking packets (not the payload, but the underlaying network mapping).
Good luck, everyone, with all of your other projects. -Z-
Thanks @Zonker for the kind words. I have also stalled a bit while waiting for the Particle Cloud HA network to be active.
.
I think my Mesh Hello World program got local Mesh publish/subscribe working quite well. Even if the mesh devices got broken into pods each pod of devices communicated fine, but I havenāt used it for a while.
.
I recently had a long talk with @will, and I am back to being more of a team player. He reminded me that the pre-order people should be able to use 10 gateways and up to 100 devices for testing. Which takes a lot of pressure off my concerns about my teaching with 15 Argons and 31 devices having a monthly cost. (Only for pre-order people unfortunately).
I did mention my concerns about the Particle Cloud Border Router and @will mentioned perhaps a local Particle Border Router (Linux machine) might be something in the future. That sounds really interesting, with lots of positives, but that might not be for a while, if ever.
.
So that is my update, instead of stressing about the Mesh devices I think I will buy this puppy.
the kids are on board, just have to confirm things with my wife.
because my school locks down the Windows Device Manager and finding the correct COM port to use serial is a bit random, I just found out that the windows command shell either CMD or PowerShell command
mode
gives all the data I need to then use putty.org to communicate in serial mode
Status for device COM25:
------------------------
Baud: 9600
Parity: None
Data Bits: 0
Stop Bits: 1
Timeout: OFF
XON/XOFF: OFF
CTS handshaking: OFF
DSR handshaking: OFF
DSR sensitivity: OFF
DTR circuit: OFF
RTS circuit: ON
I still love using the console and it is fine for testing most things, but it needs a delay(1000); after each publish, so for much faster communication serial is needed, such as working with sound, acceleration etcā¦
@will and I have already talked through our differences and the minor mess up of the zoom.us meetings. Both @will and I were fine with moving forward. On a forum we are allowed to have respectful differences of opinion and occasionally make mistakes.
Have things changed? My opinion is that the product I purchsed should work. I ordered 31 mesh devices in Feb 2018, it is now May 2019, can we now connect 2 Argons together, it has been 15 months?
When I teach students, and someone does something out of line, I try to find something they have done really well. @mstanley, did you like that I have used Machine Learning with your product, connecting a cell phone through websockets to the Particle.io Photon to control a toy car based on a persons body posture, link here. Something few people on earth have got working or even thought of doing.
No worries. The only content changed was the [redacted] information. Since the zoom account is premium, they are staticāand I didnāt want to leave them on an open message board. Web scrapers are nasty things! There was no ill will in the editing of the message.
Happy to hear you and Will were able to work things out!
It is somewhat possible to do what youāre asking. We do not have high availability networks (2 simultaneous gateways talking to the cloud in one network) yet. However, it is possible add gateway devices as nodes. For instance, an Argon can be added to an existing mesh network, and it will act like a Xenon. Itāll talk mesh, just not cloud.
High availability will come in time. The focus is to improve reliability of existing services. Some recent efforts have gone into cloud side reliabilityāespecially in terms of sleepy devices. This has a significant positive impact on gateway connectivity for mesh.
Constant efforts are made to improve device reliability. At the time of this writing v1.2.1-rc.1 is our latest release. I encourage you to always update to the latest device OS and experiment.