Arduino Create (online)

[GremlinWrangler]

That's a question to ask over at code bender, since getting it working is on their side of the house. Would suspect it's well down any list since codebender works best with network connected devices you can download to wirelessly while Teensy is very USB centric (hence midi keyboard and joystick options). Among other things one of the clever parts of codebender is the curated libraries, which Teensduino gets via Paul so a lot of duplicated effort.

 

[Pensive]

A second problem is what happens if the company providing the service goes away. I've seen companies that provide remote backup services go away, leaving their users stranded.

Exactly.

Look what happened when google bought revolV for the talent, not the product. 18 months later they shut the cloud service down which rendered thousands of peoples home automation system in-operable.

If you use a proprietary cloud software of any kind and invest significant time into the environment without hope of transferable work, your risk is off the charts. It's a worrying trend.

 

[Epyon]

Definitely, but it will take a couple of months to a year until an OS/Java update makes it unusable. Even if not, you're stuck with a dead, legacy system. In the case of JetBrains, and other offline solutions, companies (i.e my company) can't ask anyone for support anymore. And with every passing day the world moves on and we're stuck on a dead platform. So maybe we won't update that same day, but we would in a couple of months maximum.

My point is, it's really how you handle it. For example, codebender can stop developing stuff and stay alive for months with minimum cost. Even better, I can notify my users in advance, disable registrations, and keep the site running for a year with my personal budget.

Doesn't that disrupt you less than, say, IntelliJ shutting down one day and in a few months their IDE is so stale you can't use it? (I'm using IntelliJ here because we're using PHPStorm to develop codebender so I have personal experience with it)

And honestly, that goes the same for everything. If Microchip goes down, you can still use the 10 processors you have lying around, but they're essentially worthless to you because you have to switch to a new processor, one that you can base your product on. Same for your computer vendor, you can still use your computer but you might as well throw it away because there's noone supporting it, and it's going to have open, known security issues soon, so the faster you throw it away the better. Yes, this issue is more enhanced with subscription-based models, but that depends a lot on their handling of their situation, and -my point here is- it's more of a theoretical & psychological threat than a real one.

Of course, you don't have to agree with me, and of course, I'm biased, for obvious reasons

There's one (obvious) flaw in your argument. If Arduino.cc shuts down, you can keep developing with the existing IDE or switch over to other IDEs with (almost) perfect code compatibility (Eclipse e.g., or your Codebender ). If Microchip shuts down, all your existing products keep working and you can switch over to other MCUs for your new products, maybe even with some form of code compatibility depending on the toolchain. But if stuff like Particle goes down, you've lost all your code and the only way to access all your products - and worst of all maybe your products stop working too! That's creating a single point of failure. As a professional developer I cannot live with such a dependency on one single party. Working like we do, with a local Github repo that syncs to the online repo and a shared Dropbox for sketches looks to me like having all the benefits of online collaboration without the potential risks (not one SPOF, everything has a local copy).

If your Codebender goes down, that fortunately won't have the same impact as Particle shutting down, but I still don't see any benefits compared to the Dropbox/Github method that justify the added risk of an online-only IDE. Unless for educational purposes or short term collaborations, like stated before in this thread.

 

[PaulStoffregen]

Any progess towards supporting Teensy LC in codebender?

No, not really. Truth is, we've talked several times about someday supporting Codebender, but so far no actual work has been done towards that goal.

Among the technical challenges is Codebender (and Arduino Create) have browser plugins built around the concept that everything is a serial port. Teensy uses HID protocol for uploading, not serial protocol.

From Arduino, this non-serial is done by launching the Teensy Loader and then establishing communication between Arduino and the Teensy Loader to ask it to do the work of reprogramming Teensy. That works really well for the Arduino IDE. However, from a browser, launching other native programs and establishing communications with anything other than the server which sent the web page is a huge obstacle. Browsers pretty much don't allow such things, for security reasons.

I'm sure some amount of work could overcome these technical challenges. But it's not a simple matter for these reasons. So far, most user feedback I've heard has echoed much of the conversation in this thread (keep everything local), which has tended to push this daunting task to the bottom of my priority list.

 

[tzikis]

Actually, things have changed since the last time we spoke, and we do also support native USB devices as well, only it's a lot of work for us to implement each protocol, as trivial as it may be. But it's doable and Teensy is the second most requested board on codebender (surpassed only a while ago by the ESP8266)

One of the issues that still remain, AFAIC, is that Teensy still works as a plugin instead of just an Arduino core. Am I right on this one PJRC? The way we're moving forward is by making sure we are 100% compatible with the Arduino IDE, and couple that with our insanely intensive testing every time we make a change, and it becomes unmanageable for us to keep maintaining Teensy, unless PJRC also invests significant time

And lastly, I'm afraid it's also a business decision. We're still a small, struggling startup. Given the market size of the Teensy (significant, but 2% of the total Arduino market that we already support), it only make sense for us to invest that much time if the we get a significant number of Teensy users using codebender. And that probably won't happen, because Paul wouldn't want to promote codebender as the only/main solution, and quite frankly, the % of Teensy users who'd use codebender is smaller than the % of the AVG Arduino user who would find value in codebender. For one, Teensy users are more reluctant to use a Cloud service (as stated in the comments), and they would also prefer a more sophisticated tool/UI (although to be fair we're currently working on something like that)

So why spend significant time and effort on something that might send us 500 users, when we could be spending that effort on improving the life of our 80K users, (i.e Collaborative Editing), or adding features that would attract a larger number of users (i.e.GitHub integration)? As much as I'd love to support Teensy personally, I have a company to look after

 

[PaulStoffregen]

Is upgrading to the paid version your main business model? What do people get for the paid version?

I'd imagine your business justification would involve the number of people who do convert to paid version, or have other revenue generating activity if you make money other ways?

One of the issues that still remain, AFAIC, is that Teensy still works as a plugin instead of just an Arduino core. Am I right on this one PJRC?

Yes, that's correct. Teensyduino is an installer which does more than can be accomplished using only Arduino's boards manager install. Most of its special customization involves patching the serial monitor.

Long-term, supporting Teensy probably will involve more maintenance than other boards. Long term this only makes sense if both PJRC and Codebender benefit. There may be an opportunity, but it's easy to imagine this involving more investment than returns.

 

[tzikis]

Is upgrading to the paid version your main business model? What do people get for the paid version?

I'd imagine your business justification would involve the number of people who do convert to paid version, or have other revenue generating activity if you make money other ways?

Great question. For the most part, yes. It's the standard Freemium business model, and users pay for the ability to keep their projects private. We're also thinking of adding more advanced paid projects, i.e. adding a debugger for paid accounts

We're also doing deals with manufacturers & other partners, whenever that makes sense. So in other words, the manufacturer takes care of the bill for us supporting our users, because we expand their market (their products can be used in Chromebooks), we make their products easier to use, and we take care of customer support (vs the manufacturer supporting & patching the Arduino IDE)

 

[Epyon]

Not want to make this a 'told you so' but: https://codebender.cc/next-chapter

Looks like the investors pulled the plug quite hard, everything is closing in less than one month time.

Makers aren’t interested in paying for Software.

No. Makers are interested in paying for software, but not for software that is multiple steps back in functionality and flexibility. The fact that the IDE is online does not make up for the lack of features a lot of makers really care about, like being able to tweak libraries or compiling for different targets. I'm still amazed they started such a venture around a concept that offers no real benefit compared to the traditional way of a local IDE combined with Github for collaboration. Online IDEs are great for education, but there never was a market for anything beyond that seemed clear to me.