My Brief Thoughts on Roadtesting

Here's a few random thoughts on the subject:

Some companies have always tried hard to lock you in to their own tools although most seem to be trying to reduce the cost of development. In some product areas (like FPGA) they don't really have any options other than provide part specific tools.

I know several companies where the boss isn't that keen on buying software (or hardware) tools.

I don't mind (too much) paying for general purpose tools but I don't like paying for vendor specific tools - so I do all my embedded C development with Keil/ARM MDK - not cheap but supports multiple vendors.

It's all part of the big picture you need to look at when choosing parts - unless your volumes are big then the cost of development is usually more significant than the cost of parts.

I do find it very puzzling that some vendors offer Roadtest stuff with short time limited licenses - not just dev boards but instruments as well. Some Roadtests go on giving product exposure long, long after the official test is done.

On a very specific point, the Bridgetek Eve family of graphics controller chips embedd a lot of graphics library stuff in the hardware  - you can buy displays with these chips built in. There are free tools.

MK

Thanks for your valuable words of wisdom. Sounds like your approach of Keil/ARM MDK works well for you, balancing your investment with ability to use something that works well.

I know several companies where the boss isn't that keen on buying software (or hardware) tools.

That resistance is a boss thing . Once they see some costs showing extra income they tend to be more happy to hand over the company credit card !

I'll be watching out for Bridgetek Eve products now as I hadn't heard of them before, sounds like they might be easy to use if much of the GUI code is embedded already.

michaelkellett  wrote:

 

... I do find it very puzzling that some vendors offer Roadtest stuff with short time limited licenses - not just dev boards but instruments as well. Some Roadtests go on giving product exposure long, long after the official test is done.

...

Indeed. It also hampers with the ability to answer questions.

When there's someone trying to replicate or use things you've described in the test, and they are stuck, you can't help.

The vendor could have profited from the support of an already convinced believer, for no extra investment.

I'm a fan of ever-lasting evaluation licenses.

I think it is fair to say: the license can only be used for evaluation and test purposes.

But by giving the license an end-date, it results in the kit that will never be used again (as already told above) and that the new fanboy is rendered powerless .

I agree that license for compiler should be included with roadtest or provided for free for everyone developing non-commercial projects. For example Segger Emebeded Studio (which I am using now as part of Arduino Nano 33 BLE Roadtest) is free for non-commerical projects and MCU manufacturer negotiated that when used with their (but only their) devices it is free even for commerical projects.

Also note that GCC for RX72N is avalaible. It is part of installer but it is less supported than their proprietary compiler and provided examples contains some compiler specific stuff which makes hard to compile some examples using GCC (you must do appropriate modifications. Then it also should be compilable using GCC).

I do not think that board become useless when license expire, but it will be more hard to use it.

I'd like to see more open source support for such embedded boards with manufacturers perhaps issuing licenses to their software to allow the 'hobbyist' end of the roadtesters to continue using the product. Should they create something appealing then the manufacturer would likely benefit from the increased orders for the hardware itself? However being able to identify useful open source solutions that roadtesters can use and build skills in should be encouraged.

I'd love people who have the capability to roadtest products to see it as an opportunity to support the hardware with open source in mind as well, however what ends up happening, even with hardware that is designed with open source in mind (beagleboard?) is that people expect some level of "I don't want to have to work this out myself" but that fundamentally ends up being the situation when something is open source - because, no-one wants to do it for free, or give something away for free, and it only tends to come to fruition when someone has a devout interest or dedication in bringing it to everyone for the altruism for it.

I believe we also have a side difficulty when we're given products to roadtest in that we look at it and go... "you realise you can't only give us the microcontroller/hardware board for this right? you can't expect roadtesters to have everything necessary to test this properly?" and it's likely software support is overlooked in some cases, or the supplier realises having an engineer test the kit is a bit more involved and needs the infrastructure.

What was the term coined these days? Software as a service or similar?

I personally find it frustrating either way, and maybe I went off on a tangent

There are some good points above.

The bottom line is that there appears to be enough members who are willing to go through the learning curve and the road test blogs are sufficient to make it worthwhile for manufacturers to continue this method of exposure.

Manufacturers used to give out free dev kits to prospective clients, probably still do, but they didn't get much of a wider marketing effect beyond the specific user. A blog on this forum provides much wider exposure than that.

Some members will participate just to learn about the products and their tool chains, even though they can't afford to maintain them. Such experience can even be useful when applying for a job.

I am not generally in that camp, especially when it involves a huge investment in time, because anything that complex will fade from memory if it isn't used on a continuing basis. But there is still enough interest by others to keep the marketing model viable.

No tangent...all good points. Maybe also a bit hypocritical of me spouting on about the need for open source when I've never participated in directly contributing to an open source project myself (mainly because I don't think I have the required skill level). The word 'devout' sums these people up perfectly, as I imagine they spend hundreds of hours developing, modifying and supporting their creation so that others can use it for free. We need a global day of thanks to recognise their work.

Hopefully some manufacturer's will read these kind of comments and get a better understanding of what engineers desire in their toolsets.

Lastly a note to sponsors: please carry on kindly providing the great boards etc for roadtesting regardless of your licensing terms of your software.....we'll work something out in the tests

because anything that complex will fade from memory if it isn't used on a continuing basis

Likewise, and that's where writing an E14 blog helps me. I can recap using my own online material should I forget the process

14rhb  wrote:

 

... spouting on about the need for open source when I've never participated in directly contributing to an open source project myself ...

My claim to fame: https://apache.googlesource.com/ant/+/refs/heads/master/CONTRIBUTORS

I actually removed lines of code instead of writing something . I flagged a change that caused a regression and what not-needed lines of a newly submitted funtionality caused it.

I had to stop roadtesting because my ADD issues made it too difficult for me to address the challenges of reading through all of the data to figure out how to use some devices.

I do like a challenge, but when you cannot grasp the essentials, it is time to stop.

DAB