HEADSUP: ibcs2 and svr4 compat headed for history

Vladimir Dyuzhev vladimir-dozen at sympatico.ca
Tue Jun 29 19:38:00 PDT 2004


> I disagree with the first part and agree with the second.

> The reason I disagree with the first part is that it would be interpreted
> as a lottery under most jurisdictions since sending $20 towards ibcs2 only
> gives you a certain chance of ibcs2 improving, but no certainty (until
> you buy all the tickets in the lottery).

   Well, I agree that the idea lacks some important details... still,
   both FreeBSD project and FreeBSD users would benefit from an
   established negotiation way...

   For example, I now see an old Unibasic app under SCO box; my
   customer under any circumstances wouldn't pay the complete price
   for fixing ibcs2 under 5.x; but ten or twenty similar customers
   may make the trick...

   I don't know how it should be and if it possible at all.
   What I thought about is something like:

   1. Functionality A is broken. Nobody has time to fix it
      right now. Too bad.
   2. Someone who knows the code gives a time estimation for
      the A. Like "to fix it we need about 500 mh".
   3. This estimation (adjusted up ~30%) is
      published on FreeBSD "need users help" page.
   4. Users that need A register and say things like "I'd pay
      $N for that (not later then Nov 1)".
   5. When (and if) the critical amount is registered the task
      is assigned to one of the core(?) developers and users are
      asked to actually do what they preached -- transfer money.
      (Some would refuse; 30% correction above would hedge it).
   6. Developers have smth for living and users have their
      beloved functionality. Everybody's happy.

   I believe it's much better than just "well, guys, we won't support
   ibcs2 in 5.x". First at all, users become warned long before
   HEADSUP; second, they have a place to complain and to influence the

> What this means is that if some people can agree to donate $N towards
> ibcs2 development and find a developer who will do that for $N, then
> fine, they're go.

   True. The problem is that the users do not know about each
   other and all of them believe that the sum is too high to be
   paid from their and only their pocket; but your financial experiment
   showed that, given enough users, you may finance even very general
   tasks ;)

> If some entity, possibly a person in the project wants to hunt for
> donors based "I want to improve ibcs2", then that is fine too.

   Exactly. But each time that one person have to
   make all "hunting" preparation himself; and prove himself not
   a scam; and somehow deal with pay systems; and ... well, it's
   quite ineffective and not what a developer good in.

P.S. Trusted pay system would only help the project at whole.
   The reason L*x makes such huges steps forward is that it
   is financed well; there is nothing wrong in targeted fundrising
   for FreeBSD; after all, if someone wants to do the job for free,
   it's completely okay; but some tasks need a lot of time, and that
   means -- money. I know that FreeBSD was developing successfully
   long time without this kind of actions, but -- time's changed,
   why shouldn't BSD?

P.P.S. We may launch an experiment again: we may setup the fund for
      financing ibcs2 porting. That's a good candidate for proving
      (or dropping) the concept.


More information about the freebsd-current mailing list