OK, good; by good luck I am already using the M5 which is the future one.

@felmue yes, the help() function helps but its quite a slog to figure things out. For example I've been able to use the 'Lcd' class to print some text to the Atom S3's display; but when I wanted to change the font, it took me quite a while to figure it out.
(In case anyone else is looking for this, the command seems to be

Lcd.setFont(Lcd.FONTS.DejaVu18)

when I incorrectly tried

Lcd.setFont("DejaVu18")

it crashes the device!)

I have an ATOM S3 and I have burned UIFlow 2.0 on to it using the M5Burner tool.
I am using the micropython REPL to program it. (May explore UIFlow later).
The module "M5" is present in this build of UIFlow.
In other example code, I have seen a module "M5Stack", but this is not present for my device.
Can anyone explain what the difference between these modules is?
Also, I'd like to find a link to documentation on the M5 module, since that seems to be all that is available to me.

Adding to my own comment here: I believe my problems were caused by setting the 'Start Mode' parameter incorrectly. When I reverted it to "App Mode" the ATOM LITE seems to respond normally to the terminal (including Thonny) and I get the >>> prompt.
I'm not clear what the other App Mode values are for, but for now it seems to be doing what I need.

I am burning with M5Burner, and using any of Thonny, PuTTY or python -m serial to connect afterwards. All of them fail. Sometimes I get a stream of "flash read error" messages; sometimes I just an ascii-art logo printed but then no response to the keyboard.

I don't know about inter-erasure: can you point me to docs about that? At one point I did a "python -m esptool --port COM7 --chip ESP32 erase_flash" before re-burning, but without success. As noted, the burn process always ends successfully but after connection is when the trouble starts. I'm beginning to wonder if my Atom Lite is bad, but it was working before.
(I said above that I could revert to the Atom S3 version, but in fact it is not working now either)

@ajb2k3 On more careful inspection I see that my module is just ATOM LITE, not ATOM S3 LITE. So I have selected "ATOM" in M5Burner and then re-burned with "UIFlow_Lite" v1.11.8. However this does not appear to boot successfully, I cannot get to the >>> python prompt.

Can you comment on what is going on? Am I using the wrong firmware for an ATOM LITE? (I can revert to the ATOM S3 version, which was working fine, but I am trying to learn how to do this properly.)

Actually I mis-identified my module, it's ATOM LITE not ATOM S3 LITE on closer inspection (the names of these things can be confusing). I got the pin numbers scl 32 and sda 26 from the legend on the device itself, and I think they must be correct since it is able to reply with the [35] from the scan() function.

I tried looking at the Arduino code from GitHub to see if it does some kind of initialization of the DLIGHT before reading, but I can't figure it out.

Anyway thanks for replying Terry.

I have an ATOM S3 Lite with a DLIGHT connected to the Grove interface. DLIGHT is the "Ambient Light Unit" with a light sensor of some kind. I have the UIFlow v2 on the ATOM, so microPython is working. I want to read the values from the DLIGHT sensor.
I have code like the following:

 import machine
 i2c = machine.SoftI2C(scl=machine.Pin(32), sda=machine.Pin(26))
 i2c.scan()
 i2c.readfrom(35, 2)

the scan() function returns [35] so I presume it is connecting successfully.
But the readfrom() always returns b'\x00\x00' no matter what brightness the light is on the sensor.
Is there something else I need to do, to initialize the DLIGHT?
Or do I need to pass 'freq' or other parameters when I create the i2c object?

@ajb2k3 OK, I will try using M5Burner. I have an Atom Lite, so should I select "ATOM" in the list of devices on the left? And then should I select "Micropython" from the list on the right? I only want micropython, not UIFlow (at this stage).

I was able to flash my S3 Atom Lite using the file M5STACK_ATOM-20230426-v1.20.0.bin from this location https://micropython.org/download/GENERIC_S3/
I have not tried to use thonny with it but I can connect via the serial port and execute Python statements.

Good info; thanks everyone