Comment Important is a hard problem because it allows users to
attribute arbitrary words to other people.  This is the reason no
"migrate" tool uploads comments -- no LJplatform allows a user to
upload comments attributed to someone else.

I'm quite certain that Comment Import is the do-or-die adoption
feature for DWS.  If you don't have it, the only sufficient motivation
to move to DWS for 99% of users will be the complete eradication of
their journals, i.e. they'll only move if LJ tanks.

Possible solutions/approaches:

1) Allow uploads, but MARK them as uploaded.  Add a column or two to
the the comments table, "uploaded BOOL" and maybe an appropriate
datestamp.  Display them with appropriate disclaimer "UPLOADED BY
JOURNAL OWNER ON $DATE".

2) Provide means of original comment maker finding and authorizing the
comments.  (See azurelunatic's recent post, re doing this with OpenID.)
Could be combined with the above to strike the "UPLOADED" flag.  Note
this will confront many users with thousands upon thousands of entries
to approve.  Batch approval would be both necessary and a gruesome
security hole.

3) Provide original comment maker a means to proof some/all uploaded
comments against an authoritative server, i.e. upload from text files
then compare against livejournal.com to see if the comments match.
Have *no* idea if this is feasible at all, but even if it could allow,
e.g., just all the public comments to be programmatically authorized,
that would reduce how many the user needed to approve manually.  This
would involve some serious coding.

4) Allow uploads as screened, and unscreenable and frozen, comments,
so that the journal owner has them to read, but they're published to
no-one else.

5) Only allow uploads directly from authorized servers, such as
livejournal.com.  Involves divulging one's user/pass to the importing
server (so it can log in as you directly to LJ), which is another sort
of security breach.

That's what I got so far.

Finally, a non-solution anyone thinking seriously about this problem
needs to consider very hard:

A) HTML snapshot of the comments on the original journal, dropped
right into the body of a post on the new journal.

This is a grody, grody kludge and SHOULD NOT be promoted.  HOWEVER, it
makes a point: there's NOTHING stopping the end user from doing this
already, so any solution DWS comes up should be compared to that as a
reality check!  There is a tendency to make things more righteous than
they need to be, and reminding yourself "Or, of course, the user could
dump a scrape of the comments into the body of the new post" helps
cool that fever.

This is why I think importing with labeling (option 1) is quite
reasonable: it's better than the kludge (option A) without sacrificing
security over it.

-- Siderea
_______________________________________________
dw-discuss mailing list
[email protected]
http://lists.dwscoalition.org/cgi-bin/mailman/listinfo/dw-discuss

Reply via email to