$B9=B$2=J8=q$,9-HO0O$J%"%W%j%1!<%7%g%s$GMxMQ$5$l$k$K=>$C$F!"(B $B$=$l$i$r4IM}$7!"A`:n$9$kBN7O$NI,MW@-$,9b$^$C$F$$$k!#(B $B9=B$2=J8=q$O!"F1$8L\E*$NJ8=q$G$"$C$F$b!"3FJ8=qKh$K:YIt$N(B $B9=B$$,0[$J$C$F$$$k$3$H$,B?$$!#=>$C$F!"J#?t$NJ8=q$r!"(B $B$=$l$i$N9=B$E*B&LL$r4^$a$FE}0lE*$K4IM}!"A`:n$9$k$3$H$OFq$7$$!#(B $B$=$3$G!"K\9F$G$O!"9=B$2=J8=q$N9=B$$N0lIt$rCj>]2=$9$k$3$H$K$h$j!"(B $B:YIt$N9=B$$,0[$J$kJ#?t$N9=B$2=J8=q$r!"E}9g$7$F(B $B07$&$?$a$N%b%G%k$rDs0F$9$k!#(B
Structured documents can partly differ in their structures from one another though they have been made for the same purpose. This fact makes difficult the uniform management and querying of them with effective use of their structural properties. In this paper we propose a data model for treating such documents uniformly without ignoring their structural properties. The model is based on the nested relational model, and into which an abstraction facility is incorporated to view the structures of documents as the same ones at a certain level.