Backwards Compatibility BUG

[utu]

.
Here's an interesting comparison, looking at Knopppix 6.4.4 relative to
luther's Posts #16 and #20 for Knoppix 6.4.3:

knoppix@Microknoppix:~$ ls /lib/modules/*/kernel/drivers/net/wireless
adm8211.ko       atmel_cs.ko   iwmc3200wifi       p54            wl.ko
airo.ko          atmel_pci.ko  libertas           prism54        wl1251
airo_cs.ko       b43legacy     libertas_tf        ray_cs.ko      wl12xx
at76c50x-usb.ko  hostap        mac80211_hwsim.ko  rndis_wlan.ko  wl3501_cs.ko
ath              ipw2x00       mwl8k.ko           rt2x00         zd1201.ko
atmel.ko         iwlwifi       orinoco            rtl818x        zd1211rw

I think this means 6.4.4 has a kernel with wl built-in, lacking b43;
whereas 6.4.3 has a kernel with b43 built-in, lacking wl.

Hint:
wl.ko shows up in 6.4.4 which I use, and b43.ko shows up in luther's 6.4.3

So, I think this means we have to consider not only offering (at least)
two drivers, but two kernels-to-match as well, for Broadcom users.
Interestingly so, this actually occurred with Knoppix 6.4.3 and 6.4.4
taken together as one distribution. One size does not fit all here.

In trying this on Klaus K, I think there's a better, more plausible explanation:
The kernel in 6.7.1 may not have been built to accept b43 modules,
whereas, it appears 6.4.3/6.4.4 if they contain the same kernel DO accept both wl & b43 modules.

If that's the case, 6.7.1 surely favors only wl at the expense of ALL potential b43 users.
This turns out to be easily remedied by making sure future kernels don't preclude using
all likely useful modules. It does NOT imply a proliferation of kernels.