Initial code for the project

eclipse-thingweb / test-things

Collection of Things that can be used for testing different IoT protocols, security mechanisms and interaction styles

https://thingweb.io

Other

4 stars 2 forks source link

Initial code for the project #1

Closed hasanheroglu closed 9 months ago

hasanheroglu commented 9 months ago

This is the initial code for test-things. In development, I have worked with @egekorkan.

The development has started at https://github.com/thingweb/device-sandbox but that repository will be closed/archived. Already existing devices could be hosted at plugfest.thinweb.io and we want to extend the repository with more devices in the future.

For feedback on the overall structure, please provide a comment and not a GitHub review (it can be the final comment of a GitHub review).

It is a big initial contribution but we wanted to show the overall structure. The actual code is small, there is a wrapper and test code.

egekorkan commented 9 months ago

Just for further clarification:

My goal is to move every Thing we host to a dockerized setup
The current Calculator example can be of course changed. A follow up PR will contribute the node-wot version of those Things.
The utmost goal is to have multiple protocols that respect the protocol bindings where anyone can build a Consumer to test against these Things

egekorkan commented 9 months ago

I think it looks good. I added some comments (see below).

The main question I have is "why" we did not use node-wot for the exposed things. I see benefits of being neutral but somehow we partly implement it twice...

So the node-wot Things will also come but my idea is to not couple the protocol implementation of a Thing to the changes in node-wot. There are many changes happening to node-wot that do not need to be applied to these Things. At the same time, a longer goal is to have these as stable as possible so that node-wot consumer side can be tested against them.

danielpeintner commented 9 months ago

So the node-wot Things will also come but my idea is to not couple the protocol implementation of a Thing to the changes in node-wot. There are many changes happening to node-wot that do not need to be applied to these Things. At the same time, a longer goal is to have these as stable as possible so that node-wot consumer side can be tested against them.

Following you arguments I would say node-wot things should than never come. Having both does not make a lot of sense.. I think

egekorkan commented 9 months ago

Talked with @danielpeintner as well and we can merge it. I will create a discussion for the overall management of such Things