On Aug 15, 2008, at 2:05 AM, Afuna wrote:
(I swear one of these days I will Get It Right and send it to the list
correctly the first time)
For duplicate handling, looking at either the user-set time or the
server-set time, then comparing entries made within $x minutes of each
other can narrow down the list of possible duplicates, and then a
compare the full text of the entry -- or hash relevant attributes and
compare that instead.
But that leaves the problem of merging comments (merge all the
comments from one source first, before working on all the comments
from another?), possibly of merging tags, and what if one entry was
manually edited afterwards, it wouldn't be a duplicate/wouldn't be
marked as such, but could be similar enough that it would seem
inconsistent (at first glance) that they weren't merged.
It may be much simpler to do without merging altogether, but given how
some people are crossposting to two/three/four sites these days, I can
see how that would feel redundant.
I'm also thinking about how we're going to handle differing security
levels from different sites -- for instance, if you posted the same
entry to both LJ and IJ with different security levels, or different
people on your friends list/trusted list, etc. We're going to have to
do some really careful architecting work, and more than that, some
*really* careful documenting work, so people know exactly what the
import is going to do and can make intelligent decisions about how to
handle importing their stuff.
But we have a little bit of time left before we do that yet -- the
watch/trust split is where most of our attention is going right now.
(It's always more complicated than you think it's going to be.)
--D
--
Denise Paolucci
[EMAIL PROTECTED]
Dreamwidth Studios: Open Source, open expression, open operations.
Coming Summer 2008!
_______________________________________________
dw-discuss mailing list
[email protected]
http://lists.dwscoalition.org/cgi-bin/mailman/listinfo/dw-discuss
