ImpCentral for developers


#1

Does the new impCentral affect developer accounts and the existing IDE, or is it just for customers that have factory type imp products. Not sure if my few misc. single imp001’s and imp003’s are affected with the free development account.


#2

The new impCentral UI will eventually be the de-facto UI for all Electric Imp accounts. Developers using our free service will have the same opportunities to migrate models and devices to the new world as paid customers.

Though the particulars haven’t been announced yet, there will come a time –
not for months, yet – when any models and devices which have not been manually migrated will go through an automated migration process.


#3

Thank you, that makes it more clear for “free service” developers (like myself).

I’ve always been skeptical about the “developer free service”. Even a couple years ago, when I began using the imp001 for the first time, I questioned it. Free forever has never been a sure thing in my experiences. I would have to think eventually a person who calls himself (or herself) a developer, can only be a developer for some finite amount of time and have only a finite amount of active imps within that developer account. So eventually all customers will be paying customers?

I don’t mind paying for active imps, but what that cost is, will determine if I will stay with the imp platform. Again, I’m just speculating and thinking things in my head. The impCentral seems to be a step in the direction of “no more free service”. After all, it does cost money to have developers like me using the imp cloud every day, for several imps. The impact of that on the Imp company is unknown to me.

It would be interesting to get some insight into this and thoughts from others. Perhaps set the record straight on the topic of “developers using our free service”, and what that will entail in the future.

Thanks.


#4

We did go into this right back in 2012 when we launched - people asked then, too. No, we will not charge people for genuine developer use of imps.

TL;DR: we provide a really great industrial grade service that anyone can use to reliably and securely connect things to the internet. Many people who try this at home end up taking it into work because… well, it “just works” for years on end. Supporting developers for free proves how scalable our system actually is :smiley:

More detail:

Essentially, nobody becomes a commercial customer without using the developer service first, so it’s a marketing/customer acquisition cost.

Given how many commercial devices are running impOS (1M+), the developer support cost is not significant,

We do understand that there are many developers who will never become commercial customers, but who have many tens of imps being used for personal purposes (I’m one of those people, with imps in plugtops, garden lights, environmental sensors, solar inverter monitoring, halloween setups, etc). We reserve the right to limit bad behavior - like running huge amounts of traffic through imps - but those people are generally aware they’re being a PITA in those cases and take guidance well.

We have had “developers” who are obviously doing commercial deployments without paying us anything (hundreds of “developer” devices spread out around the world, or hundreds of vaguely identical accounts, all created by agreeing to the terms of service which explicitly deny that). We can easily find those people and talk to them about paying, though.


#5

IMHO one of the key reasons many IOT projects are failing or struggling to scale is because many projects side line or do not appreciate the value or necessity of extensive field testing. In my experience this is one the major cost drivers in developing a viable commercial product as field testing takes many months and involves significant redesigns and reliability/resilience fixes. It is very easy to develop prototypes in a lab but once it is out there it becomes timely and costly to fix problems and if it’s already a commercial product then this requires customer authorisation and the rest. This is where I have found, and still am finding, greatest value from the current Imp IDE as you can fix on the fly - as in, use a laptop out in the field to modify code and upload (ideally I would love to be able to do this via smartphone or ipad) or you can evaluate and do changes remotely from your desktop. I have yet to find other comparable systems with the same seamless workflow (not to mention auto versioning etc).

So while the current setup has limitations in terms of code development features, which the ImpCentral fixes, I would hope that you always maintain a cloud based IDE for the purpose of allowing fixes on the fly. I feel it would be a step in the wrong direction tying code development to a hardware device.

Then @hugo I may as well throw this comment / observation into the pot, regarding what it takes to move from precommercial prototype / field test to commercial product.

A key component that still seems to be shrouded in secrecy (as in there is no upfront transparency) is what are all the commercial / licensing costs. I am one of these people who always feels with one-on-one negotiations with someone in sales is that you could be paying a little more than your neighbour. Personally I find it a rather pointless exercise having to crystal ball and spin a convincing story as to future sales projections and timelines knowing full well all the risks involved and how many changes take place getting to a MVP from a proof of concept. Then if you haven’t “sold” your story well enough to sales they have, in the past, lost interest and are not very engaging / quite slow to engage (unlike this forum). So I would hope you could reconsider your approach as my investors and those with the hold on the money do not like surprises and want to know all cost upfront. Anyway fingers crossed my lengthy field trials are nearing completion with one potential system and we are now starting to evaluate and scrutinise the commercials again by doing various scenario analyses / cash flow modelling against various growth strategies. It is therefore really necessary to be able to have these figures upfront without having to divulge a potential product’s roadmap.


#6

We do have price lists - you just have to email for them; the problem with price lists is that applications tend to be somewhat unique and a price list isn’t a great way to communicate the flexibility we have in different dimensions. They are a good guideline, though, and if you need one then let me know.

It’s true that the very smallest customers will get less attention (from sales, anyway - some people might appreciate that even though all our salespeople are ex-engineers), and we are working on full “self service” which will be the zero-human-interaction Twilio model.


#7

@hugo ok great I will email. No I fully understand your position and most would follow the same approach. However, what tends to happen is that at very low volumes / smallest commercial customers things are more automated / transactional and do not need sales time + bunch of questions. Which is what I like to get things started and should suit both parties. As you grow and need more engagement / different pricing for services then I look to speak. Needless to say the Twilio model is also appealing too.


#8

To clear up any misconceptions:

impCentral is the IDE. Think of it as IDE 2.0. It does all the things you can do with the IDE — and a lot more besides, particularly in terms of wrangling devices, whether they’re for development or production. Like the IDE, impCentral is for any account holder — folk with free accounts simply won’t see the production functionality, just as they don’t at the moment.

@Gerrikoio mentioned iPad use. The IDE works reasonably on my iPad in landscape, but it could be improved. I haven’t tried impCentral on iPad, but I will.

The impCentral API is a ground-up rewrite of the Build API, adding all the stuff missing from Build: specifically production management and collaboration.

No one’s going to use the API on its own — it’s an enabling platform for other tools, such as build-cli, for which impCentral updates are in the pipeline. With the impCentral API, there’s no reason why someone with time and coding skills can’t build a native (for example) iOS IDE. I’m updating Squinter (macOS).