Clarifications about the FIT Protocol specification

manlio_perillo over 3 years ago

Hi.

I'm implementing a FIT file decoder, and I think I found some issues with the specification:

1. The endianess used in the File Header (table 1) is not explicitly documented.

2. In the "Normal Header" section (4.1 in the old PDF document), the HTML formatting does not

make it clear that "Definition" and "Data" are subsection.

The same problem is with the "Local Message Type" subsections: "Definition Message" and

"Data Message" (this problem is also present in the old PDF document.

There is also an empty "Data Message Header".

3. When I tried to print the HTML document with chrome, I found that both the document

header (with the FIT logo) and the footer are visible.

Thanks

Manlio Perillo

Ben FIT over 3 years ago

In the file header any value with a size greater than one byte is labeled with LSB and MSB, which indicates the byte order.

I agree that some of the CSS makes parts of the page hard to read. I think that H4 and H5 have the same style. We will take a look at improving that.

Thank you for the feedback!

manlio_perillo over 3 years ago

There is another small issue.

In Table 2 "Normal Header Bit Field Description", the table use the LSB 0 numbering, where the Normal Header bit is 7.

However this may confuse the reader, since it is possible to do:

header & 0b10000000

Strictly following the text, one may instead do:

header & 0b00000001

Thanks

manlio_perillo over 3 years ago

Sorry, I forgot to report some other issues.

1. It seems that old FIT SDK are no more available, on both https://www.thisisant.com/developer/resources/downloads/ and

Are they still available somewhere?

2. In table 7 "FIT Base Types and Invalid Values", the text specifies that integers use the two’s complement format.

However nothing is said about the float32 and float64 format.

Thanks.