Revisions to stepper-2.0

[dtex]

Drop member number from multiStepper commands

Comment updates

Change bias for exponent values to -17

[dtex]

I still need to add details for the float composition

[soundanalogous]

For just one of the 3 sections that describes how the float is composed, can you provide an example for a specific number or two (maybe 500.0 and 0.000234 or something like that) how this would be broken out into the 3 parts specified in the table?

[soundanalogous]

Even better if the example(s) align to a specific duration (such as 1 step per hour for "slow" and 10 steps per second for "fast"). You probably have a better sense for practical examples than I do.

[dtex]

I thought I might mention the custom float format up front, describe it in detail at the end of the document (with examples) and for each method where it's used just refer users to that detailed description.

Sound good?

[soundanalogous]

sgtm

[dtex]

Working on those examples and testing as I go, I've realized that I don't have a clear understanding of how AVR-GCC fakes floats. The data format probably should be transmitted in a format that closely matches what we actually need. I think the root of the problem is I've been using decimal32 representation instead of binary32.

[soundanalogous]

Arduino is not just avr-gcc anymore. Various architectures each have their own compiler and may handle the float to int conversion in different ways. As long as we use a normalized representation of a float we should be okay. On the firmware side that unpacked value will be assigned to an Arduino float so it should then be set properly in memory at that point.

[dtex]

I added those examples, and I have one last thought. Both maxSpeed and minPulseWidth require separate calls to change their default values in accelStepper. Since a reasonable default exists for both should we remove those values from the configuration command and give them their own command?

[soundanalogous]

We can pull them out to their own commands and if they don't already have default values within the AccelStepper implementation, we can give them default values upon receiving the config command. That will allow the user to not have to worry about those values in many cases.

[dtex]

Done.

[soundanalogous]

minPulseWidth

[dtex]

Thanks, fixed.

[soundanalogous]

We have to decide here if we want to specify in steps/second, in which case we probably need to use the same 4 byte float we're using for speed and accel. Or we can allow specifying steps/sec as an integer here since it will always be > 0 and < 1000 and we shouldn't lose too much in rounding. However we should probably call out the need to round if that is the route we want to take. I prefer the former for consistency, but the later for simplicity.

[dtex]

I think consistency should rule here. Besides, accelStepper's setMaxSpeed receives a float so it's not like we're making a performance improvement by using an integer.

[soundanalogous]

sg, lets change setMaxSpeed to use the 4 byte float then

[dtex]

That's done.

[soundanalogous]

Cool, this LGTM. I'll merge now unless you wanted to make any other changes.

[dtex]

I feel pretty good about it so I'd say merge away. Starting in on configurableFirmata now.