[Nut-upsdev] New naming patterns for USB hid2nut entries in drivers

[Jim Klimov]

 Hello fellow devs,

  The discussion of https://github.com/networkupstools/nut/pull/431 slowly
led the new driver submission from EVER Power Systems to being almost
acceptable, with one raised issue being how they would "NUT-name" certain
HID entries they could not find equivalents for in the existing NUT
documentation.

  One proposal in the PR was to define a loosely-structured `vendor.*` tree
for such values, which can get the proposal moving forward but may
undermine the interoperability in the long run - especially when we have
more devices supporting some concept that is new today, so it gets a name
and at best old drivers have to get their names remapped (legacy alias that
someone could have scripted against, new official name), and at worst the
new features would be slightly different and so can not be a 1:1 mapping.
Given that new players do come into the market, having established names
and features expected of those can drive their engineering decisions into
making something quickly compatible with little traction, to everyone's
benefit.

  What would the collective conscience and historic practice suggest? Feel
free to chime in on that PR discussion thread right away :)

Jim Klimov
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/nut-upsdev/attachments/20201113/749467bc/attachment.html>