[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]
Re: Splitting the pool (Re: [zzdev] Re: Poolin pilkkominen --)
- To: Tuukka Hastrup <Tuukka.Hastrup@xxxxxx>
- Subject: Re: Splitting the pool (Re: [zzdev] Re: Poolin pilkkominen --)
- From: Tuomas Lukka <lukka@xxxxxxxxxx>
- Date: Fri, 8 Feb 2002 17:15:08 +0200
- Cc: zzdev@xxxxxxxxxx
- In-reply-to: <Pine.LNX.4.44.0202081101230.22432-100000@xxxxxxxxxxxxxxx>
- Mail-followup-to: Tuomas Lukka <lukka@xxxxxxxxxxxxxxxx>, Tuukka Hastrup <Tuukka.Hastrup@xxxxxx>, zzdev@xxxxxxxxxx
- References: <20020208074154.GC1095@xxxxxxxxxxxxxxxx> <Pine.LNX.4.44.0202081101230.22432-100000@xxxxxxxxxxxxxxx>
On Fri, Feb 08, 2002 at 11:13:48AM +0200, Tuukka Hastrup wrote:
> On Fri, 8 Feb 2002, Tuomas Lukka wrote:
> > Tuukka, in the future we should discuss these kinds of things
> > on the open lists.
>
> Oh well, just thought this was our mess and we should clean it behind
> closed doors.
It's our mess but it's still useful to discuss things publicly.
> > Well, currently I don't think there are any parts we shouldn't
> > distribute except the .ps / .pdf files we don't have copyright for.
>
> It's more like who takes them out? And how to ensure we all know about it
> then? Otherwise first sync with old MS blocks will bring these back. Maybe
> we could have an id kill list in the Makefile for now.
Hmm... maybe we should put a name for the main pool that's different from
the empty one now...
The kill list idea is important in itself - we need to be able to put
in metadata: "Person X doesn't want block Y to be distributed" for potential
future problems.
> > This is actually a non-issue, if you think about the structure of the
> > mediaserver P2P space: it doesn't matter where you sync your pool!
>
> I see that the propagation path is not important right know, and it has
> little to do with MS servers' naming. However, I thought we would have
> some official central repository _available_ still for a while, as it
> makes things easier for now; of course we can begin syncing with each
> other, but it's practical to have some URL for Makefiles and such, some
> URL where the data's guaranteed to be available.
Sure, it's a good idea. I just wanted to emphasize that the permanency
doesn't need to be emphasized.
> > Remember: this is not the WWW where it's important *WHERE* something is.
>
> But you do realize we don't have working P2P with discovery and such
> running at the very moment. Basicly I was suggesting us having pointers
> http://ms.gzigzag.org/* for the pools we manage.
Due to Ted's trademark stuff we should probably not put it at ms.gzigzag.org...
How about asking the department for an alias machine (mediaserver.it.jyu.fi)?
Tuomas