Lazy Reading for 2018/04/22

Accidental theme this week… video games, though not strongly.

 

7 Replies to “Lazy Reading for 2018/04/22”

  1. »On Linux, where reinventing wheels in the system is a constant.«

    *cough* HAMMER2 *cough* ipfw3

  2. Gnome wheel reinventions isn’t fault of Linux developers. You are crying on the wrong grave here. And at the time JWZ wrote that, there weren’t a thight integration with Linux(wayland, systemd, dbus…) so, not Linux fault AT ALL.

    We have to be honest here: On FreeBSD you have a lot of repeated tools to do the same frigging thing, like, 3 tools to manage disks:

    https://www.freebsd.org/cgi/man.cgi?fdisk
    https://www.freebsd.org/cgi/man.cgi?geom(4)
    https://www.freebsd.org/cgi/man.cgi?query=camcontrol&sektion=8

    Guess what?

    List disk. Linux version: fdisk -l
    List disk. FreeBSD version: geom disk list or camcontrol devlist

    What more? 3 options of firewalls:
    https://www.freebsd.org/cgi/man.cgi?ipfw
    https://www.freebsd.org/cgi/man.cgi?query=ipfilter
    https://www.freebsd.org/cgi/man.cgi?query=pf

    And you have features scattered through aaaaall of them.

    I Use BSDs(mostly OpenBSD and FreeBSD) and Linux. But this game of blaming isn’t going anywhere .

  3. I know. But worse than that, they have imported an old version of OpenBSD pf, and didn’t gave really much attention on maturing the tool or keeping it sync with OpenBSD -HEAD cause all the propaganda that was being advised at the time was “FreeBSD pf was more SMP-Friendly than OpenBSD, we rock!, and ipfw can be thrown asside, since we have a new shiny tool”. Pf now seems dead in the water on FreeBSD cause, those smp patches made it hard to keep sync with OpenBSD pf that is constantly implementing new features and maturing(and also being SMP aware).

    And the same way the new pf kid in the block made ipfw development slow down…there is an IPFW-ng idea being developed, and guess what? – https://wiki.freebsd.org/IpfwNg – There are some of the main efforts being inspired by Linux iproute2, like a /etc/ipfw directory with the “tables” file, chain dependencies inspired by “iptables”, an ipfwsync to mymic pfsync(cause now, pf ain’t cool anymore), object oriented stuff and of course lots of “modernizing” since ipfw is also an old asset the same way pf came to hit a dead end at the road(it’s at the Handbook, a big red box that says you have to be careful since its an OpenBSD 4.5 for of pf).

    See? Sometimes you HAVE to reinvent wheels, but people keep blaming Linux while they are doing the same, quietly and with less media attention and failing.

  4. Yeah, I think this criticism was a bit uncalled for. There is about as much wheel reinventing in BSDs as in Linux, and it’s not a bad thing.

    Besides, some of the guys working on this BPF stuff (the B stands for Berkeley by the way :D) are as smart as Matt (e.g. Brendan Gregg, who wrote the book on modern Unix performance analysis).

  5. I think the message is being missed; my complaint is that wheels get reinvented, and then reinvented again before the previous reinvention gets to its full potential. It happens more on Linux.

    New versions of software are natural and happen with Linux or BSD or whatever, but I think there’s a potential for software to be refined rather than replaced. I do get a certain enjoyment out of knowing a particular chunk of software has made it through multiple decades of ‘doing what it says on the tin’.

Comments are closed.