Hmm. This usually happens when the feed doesn't provide a stable entry identifier that lets LJ or DW tell an edited entry from anew one with same or similar content. In the absence of such a field, the synsucking task assumes it's a ew entry, hence the result you see. So, yes, it's ugly, but I'm reluctant to blame it on the LJ/DW side. More to the point, LJ and DW *do* provide a stable entry identifier, so what you see for that feed shouldn't happen (provided the feed aggregator on the other side is sane). It would see them as changed existing entries, not as new entries that just happen to have (nearly) the same content.
Also, crossposting isn't the same as importing feed content: if you crosspost an entry from DW to LJ, it's DW pushing entry content to LJ, whereas if you syndicate your DW journal to LJ (create a LJ feed of it), it's LJ that pulls (public) entries from DW, and the formats used to pass entries are different (syndication uses standard formats - RSS or Atom, whereas crossposting uses something ad-hoc, I believe). That is to say, it could well be that there is no such standard, but the automagic image hack doesn't prove it (it's the only way that applies to crossposting, but not or not necessarily for syndication).
no subject
Also, crossposting isn't the same as importing feed content: if you crosspost an entry from DW to LJ, it's DW pushing entry content to LJ, whereas if you syndicate your DW journal to LJ (create a LJ feed of it), it's LJ that pulls (public) entries from DW, and the formats used to pass entries are different (syndication uses standard formats - RSS or Atom, whereas crossposting uses something ad-hoc, I believe). That is to say, it could well be that there is no such standard, but the automagic image hack doesn't prove it (it's the only way that applies to crossposting, but not or not necessarily for syndication).