I have worked in this space for one of the big players.
It’s bad, okay? Terribly terribly bad. Mountains of protocols designed-by-commitee, 4 major generations of everything and every non-toy device is supposed to implement all of them. The list of documents is incredibly long too.
And that’s still the better part, additional software that you do not need but really really do want or you’d need 10x more employees. The one used to monitor, manage, make reports, forward packets, upsell. It’s all unfathomably terrible.
That there are working new modem implementations could be considered some divine miracle the complexity is so ridiculous.
Reading this article felt like being beaten to death with a large brick which had been meticulously engraved with every acronym ever written. I can only imagine the pain the author must have felt reading the actual full specifications. Kudos to them.
I have worked in this space for one of the big players.
It’s bad, okay? Terribly terribly bad. Mountains of protocols designed-by-commitee, 4 major generations of everything and every non-toy device is supposed to implement all of them. The list of documents is incredibly long too.
And that’s still the better part, additional software that you do not need but really really do want or you’d need 10x more employees. The one used to monitor, manage, make reports, forward packets, upsell. It’s all unfathomably terrible.
That there are working new modem implementations could be considered some divine miracle the complexity is so ridiculous.
Reading this article felt like being beaten to death with a large brick which had been meticulously engraved with every acronym ever written. I can only imagine the pain the author must have felt reading the actual full specifications. Kudos to them.
That line from Mickens’ The Night Watch popped into my head unbidden
It’s interesting that the firmware of the same modem (nRF9160) was the source of the extensive debugging story in this post from a month ago:
https://lobste.rs/s/7jpnyh/hunt_for_error_22