>>>> (b) how we make a very large
>>>> quantity of data available.
>>>
>>> Bittorrent with all mirrors.
>>
>> Problem with that is updates, you'd only want to update it on a
>> bi-weekly or monthly basis to avoid the constantly changing dataset.
>
> Let the trackers just update the torrent file contents :) That will
> allow to have different torrent file distribution, while the contents is
> shared over the different updates.
Are completed seeds rescraped? I'm not sure it will start to download a
torrent that is already completed if the scrape fails... On the other hand, how
about gittorrent?
http://code.google.com/p/gittorrent/
_______________________________________________
dev mailing list
[email protected]
http://lists.openstreetmap.org/listinfo/dev
