Kinetis SDK poor documentation

[matheuspinto]

Hello,

I'm doing this post to get it off from my chest the frustration about the lack of documentation on the Kinetis SDK.

I'm really trying to migrate from the "old" processor expert components to new KSDK components, but I think this is a useless task.

There is no documentation. The help in processor expert doesn't exist. The "reference manual" is too incomplete.

I don't know which is the source for some ISRs to be called, what arguments put in functions, etc. Is really hard (if not impossible!) try understand too many configurations!

That is it

[petec20]

I've arrived on this forum after spending several months writing assembler on TI's TM4C1294, which is somewhat comparable to the K64/K65. I'm about 1/2 way through reading the reference manual for the K65. It was my first venture into TI ARM MCUs, and other than some involvement into TI's MSP430, my first ever venture into TI MCUs. However I'm a long time user of (what was then Motorola) devices - 6800, 6809, HC05, HC11, 68K. What I found was that the TM4C1294 datasheet (what Freescale term a reference manual) was visually very good, and somebody has made a very good start on it, but completely lacking in finer detail, the detail you'd need to actually set some of the device up. Okay, so TI have 'Tivaware' which is their version of the SDK but I found this to be sometimes poorly coded, and sometimes in direct conflict with the datasheet. What is apparent is that the K64/5 reference manual does at least go through some of the important basics of much of the module setups for things like the SDHC controller and DMA.

This was one of the key things lacking in the TM4C1294 documentation for the USB controller; even if you appreciate the basics of USB setup & enumeration the information needed to bridge that knowledge in to writing a driver for their USB controller isn't there. The flowcharts often seen in good documentation are not present, nor can TI supply them. At least in the K64/65 reference manual they are there, if only in text form, and I'm about to find out how good they actually are.

Thiago mentions Doxygen and it does indeed seem to be the downfall of modern documentation. In an age when companies claim to be so driven by user feedback and the need to perform, why do they not use user feedback, freely available, usually their own forum, to refine what they do? It's not rocket science. So in the drive to get people to adopt their products there seems to be an emphasis on providing quick to knock up projects (usually revolving round an accelerometer!) rather than actually teaching people how to truly use their devices. Answer people's questions well, once, in the reference manual, and save a hundred or a thousand people's time who are trying to adopt your product.

One of the main things often missing from SDK documentation is an overview of what's going into a module and what's coming out. ie there's comments on individual segments but the overview of larger pieces is missing.

So I'm about to start to pick up Kinetis K64/65 and see how far we get.

[ivadorazinova]

Hi Peter.

Thank you for your feedback.I passed it to the responsible team.

Best Regards,

Iva

[ivadorazinova]

Hi Matheus Pinto,

thank you for your valuable feedback.

Currently we don´t have the document you need, but we are already working on the improvement which will fully support your requirements.

At the moment please refer to documentation what is new for PE in KSDK, see bellow.

Best Regards!

Iva

[thiagopalmieri]

I have to agree with you.

However, as I see it, it is more a development team misuse of Doxygen than a tool fault per se.

I have seen lots of DOxygen generated files nowadays that do nothing more than describe the function inputs and outputs. This kind of documentation, in my opinion, is very poor, nobody seems to care about examples, flow graphs, state machines, block diagrams, etc....

Luckly, the Freescale SDK has example projects, but the documentation must indeed be improved.