On Thu, Oct 04, 2001 at 08:34:46AM -0700, Ian Clarke wrote:
> It would certainly be great to get these in if they can be implemented
> in the next two weeks, but I don't think the release should be delayed
> if they are not.
Rudimentary implementations of at least the first two would not
necessarily be that difficult, though I guess it would take someone who
knows his way around the code decently - which means me or Tavin, and I
do not have time until at least after the conference.
> Correct me if I am wrong, but all of these can be implemented in a
> backwards compatable fashion, and I think that under the "release early,
> release often" mantra, it is more valuable to get the current codebase
> the wider testing it would receive with a .5 release in about two weeks,
> than to wait longer and get them in.
It's doubtful whether load balancing could be made backwards compatible
- it depends on whether unknown fields are preserved in the StoreData
message at the moment, which they should be, but well...
> The most pressing issue for me is resolving the seednodes issue, and
> making it easier for people to set up non-centralized seednodes. The
> best thing I can think of for this is the addition of a FCP command
> which extracts a list of nodes (with a good CP) from the datastore and
> allows a client to create a seed.txt file which could be uploaded to a
> website (perhaps automatically).
My experience with the installation topology of software is that it
consists of a landscape of pyramids, where more knowledgeable users show
software to those below them. Those at the top of pyramids are not
always hackers, and often not even liberated software users (ie still
Windows slaves) but they usually know enough to communicate with others
and read multiple step instructions. I still believe that if importing
and exporting node references was just made painless enough that they
could pass painlessly down these pyramids, that would be enough. (Not
all the pyramids are very steep, but it still provides much better
entropy then a single top).
>
> Ian.
>
> On Thu, Oct 04, 2001 at 12:08:44PM +0200, Oskar Sandberg wrote:
> > Probablistic caching :-
> > Load balancing :-
> > Address resolution keys :-
--
Though here at journey's end I lie
In darkness buried deep, above all shadows rides the Sun
beyond all towers strong and high, and the Stars forever dwell:
beyond all mountains steep, I will not say the Day is done,
nor bid the Stars farewell.
(JRRT)
Oskar Sandberg
oskar at freenetproject.org
_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl
