[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: The fate of libc5

>>>>> "Ben" == Ben Armstrong <synrg@sanctuary.nslug.ns.ca> writes:

    Ben> On 11 Jul 2000, Turbo Fredriksson wrote:
    Ben> because it makes for more things that can break.

Very good point...

    Ben> I think you greatly underestimate how much effort is
    Ben> involved.

Maybe so, but maybe not... I have tried to make the transition from
Roxen1 to Roxen2 as smooth as possible, without breaking things. But
since Idonex (author/owner of Roxen) don't support it 100%, neither
can I (unless I do some SERIOUS coding in pike, which is a language
I don't fully master).

So I took a different route, I made them installable side by side...
That's also a way of having 'backward compability' (Al bight by cheating :)

    Ben> What does this have to do with Debian?  Well, in a
    Ben> round-about way I am saying that users will always drag their
    Ben> heels about upgrading, and it really is not in their best
    Ben> interests to do so.

True, and I have been there myself. But I have the VERY firm belief that
you NEVER, EVER mess with a working (production) system! NEVER!

It _WILL_ give you problem. But yes, one day you absolutely have to, because
the feature you have been waiting for/begging for FINNALY arrived...

    Ben> But  in the  end,  users are  better  served with  reasonably
    Ben> up-to-date software

Couldn't agree with you more on this...

    Ben> At some point you must draw the line, for your
    Ben> developers' sake, and for your users' sake.

This was my main key point  about Commodore and the Amiga... I _LOVED_
(strangely enough,  considering my  previous statement perhaps  :) that
they  basicly said  'Do it  our way,  or it  will (!!)  break  when we
release new  OS versions'... I think  that showed guts,  and quite big
balls (pardon  the expression).  But it's one  thing to  say 'backward
compability is not our main concern' from the start and another to say
(and show!) that 'backward compability is a big concern for us'... The
last is  what we  have said  (maybe not officially,  but we  have sure
SHOWN that  this is the case) from  the beginning.

 This is  one of the  big strength in  Debian, and I  understand fully
that maybe we must draw the line somewhere, I just don't want it to be
decided out of the wrong reason. I'm just not convinced that we really
have  that overhead  that's  discussed.   Maybe if  we  decided to  be
backward compatible NOW (without all that code base), then I'd just say
'forget it',  but now... With  all that existing (working?)  code, I'd
say 'be cool, think twice'...

Nitpicking, yes, but that's me :)

But maybe  you got  the solution there...  Upgrade in  steps... That's
backward compability to... The cheat solution, but it works :)

I keep quiet about that, I'm convinced (i think :) that 'upgradability
in steps' _might_ be the correct way.
jihad CIA Marxist Kennedy NORAD cryptographic FBI attack colonel
Mossad ammunition [Hello to all my fans in domestic surveillance] Uzi
fissionable SEAL Team 6

Reply to: