hardware Rosetta Stone?

Royce Williams royce.williams at gmail.com
Mon Sep 19 05:02:28 UTC 2011


The tl;dr version: Let's start documenting our implicit knowledge
about hardware.  It will be significant force multiplier.  If we start
a stub on the wiki, it could become useful relatively quickly without
a lot of effort.

The long version:

I've followed a few list threads that end with "Go talk to Vendor X
about that."  Developers, vendors and hardware owners are all busy, so
this can be hard.  Most hardware vendors don't have someone like
Intel's Jack Vogel - competent, constructive, and paying close
attention to the relevant freebsd-* lists.  Even if Jack misses a
message, everyone else on the lists knows to say, "Hey, send that to
Jack."

I've also spent a lot of time painstakingly reconstructing knowledge
from multiple forums in order to arrive at The Whole Truth about a
specific piece of hardware.  Someone knows the answer off of the top
of their head, but that person is busy frying bigger fish -- and they
should be.  Caching the results of that work would be a high-leverage
activity.

I propose creating a hardware Rosetta Stone of sorts.  It would be
sort of a cross between the Hardware Compatibility list, parts of
Jeremy Chadwick's list of known issues, some FAQs, and a list of
FreeBSD folks to coordinate between vendors and the project.

As a quick start, I propose a wiki page that would contain something
like these tuples:

* Hardware family
* Name of vendor
* Brief advice on how to work with that vendor.
* Identifying device info (PCI/USB IDs, etc.)
* Links to significant PRs.
* Link to a FAQ page/section for that hardware/drivers.
* Name of coordinating volunteer(s)

I'd bet that there are already some de facto vendor "ambassadors" of
sorts who could seed parts of the initial list pretty quickly.


Benefits:

* Developers (who already get peppered with questions about given
hardware) only have to answer a question once, and simply link to the
Rosetta Stone thereafter.

* Non-developers can take their hard-won research and put it where
others can reliably find it.

* Newbies spend less time chasing their tails and FAQing the lists.

* Vendors get a consistent voice from, and more deliberately connect
with, the FreeBSD community.

* Developers who want to tackle a hardware family can easily survey
the state thereof.

* Vendors who are hard to connect with might be swayed by a large list
of "this vendor won't work with us, but hey, their competitor is
totally helpful" data points.

* People could rotate through ambassadorship, to spread knowledge and
prevent burnout.

* Everyone gets to spend more time doing what matters.


If I'm reinventing a wheel, any breadcrumbs appreciated.

(I got this idea from a club I'm in that has a semi-official
ambassador to eBay -- which has really benefited both eBay and the
club -- and also thinking about Jack, and Jeremy's list).

Royce



More information about the freebsd-doc mailing list