:: [DNG] Drive-by critique
Forside
Slet denne besked
Besvar denne besked
Skribent: Rick Moen
Dato:  
Til: dng
Nye-emner: [DNG] A concrete proposal [was: Re: Drive-by critique]
Emne: [DNG] Drive-by critique
A very odd thing happened on #devuan, and then an incrementally odder
thing on dev1galaxy, on the day before Pearl Harbour Day.    Both 
involved a whirlwind visit by Eric Raymond -- who came and went with
some quite ranty opinions.   Being sleepy at the time, I good-naturedly
promised Eric I'd raise his points on Devuan's mailing list.  Mulling
over same, though, I've had to have second thoughts -- and to my regret
can't keep that promise, but will post some observations.


What Eric left for us:
https://dev1galaxy.org/viewtopic.php?pid=13144#p13144

The big problem is that Eric wants to hector Devuan into being things it
isn't, that its careful four years of planning have not aimed towards.
I doubt tat he took the trouble to properly understand Devuan's chosen
mission and strategy. (I think I basically understand it, tough I'm a
newcomer.)

A key part of the basis of Eric's argument is irritatingly and obviously
untrue:

If you want to have more devs, you need to attract a larger userbase.

I'm surprised to see Eric advance this non-sequitur. Open source
projects attract developers because their needs and objectives (or the
needs and objectives of their employers, which amounts to the same
thing) are met and served. Moreover, that is a truth that has long
been advanced by Open Source Initiative, which Eric co-founded, and I'm
pretty sure it's also an insight in Eric's own writings.

Anyway, in effect, Eric wishes a systemd-free distro existed, one he
didn't realise can't be Devuan, that has only _one_ installer image (per
supported CPU architecture), not several to choose among, that merges in
all possible proprietary firmware BLOBs, and that consistently uses a
cutting-edge installer kernel and installed kernel (hence, maximum
possible new-hardware support).

I don't think Eric has any clue about the developer cost of maintaining
bespoke installer images, or about the weird bugs and instability that
can come with bleeding-edge kernel instead of (as Devuan and Debian use)
stable kernel versions with backported fixes. Also, he may not know
about the developer cost of too many avoidable differences from Debian,
our sister distro whose work Devuan is smart enough not to duplicate if
possible, and where possible works with.

Someone advocating views like Eric's might be able to make a case for
them if he/she were willing to stick around and pitch in to make them
happen, but in that area one finds the other problem: His advice
appears to have been on a drive-by basis, near as I can tell.

I'm put in mind of one of the traditional Debian answers when a visitor
gets demanding and wants to know when something will be fixed or
released: 'Sooner if you help.'