Check if the transducer has bounded lag, i.e. that the difference of length between the input and output words is bounded, for every word accepted.
It is a pre-condition for transducer synchronization.
import vcsn ctx = vcsn.context("lat<lan_char(ab), lan_char(xy)>, b") ctx
a = ctx.expression(r"'a,x''b,y'*'a,\e'").automaton() a
This automaton has a bounded lag: there is at most a difference of 1 between the length of the input and the length of the output (e.g., $abba \rightarrow xyy$).
b = ctx.expression(r"(\e|x)(a|\e)*(b|y)").automaton() b
This transducer, however, doesn't have a bounded lag: there can be an arbitrary large difference between the input and output. For example, $ab \rightarrow xy$, but $aaaaaaaaab \rightarrow xy$.
ctx_3 = vcsn.context("lat<lan_char(ab), lan_char(jk), lan_char(xy)>, b") c = ctx_3.expression(r"(a|j|x)(b|k|\e)*").automaton() c
In the case of more than 2 tapes,
has_bounded_lag checks that every tape has a bounded lag compared to the first one (incidentally, if that is the case, it will insure that every tapes has a bounded lag in respect to every other). This transducer has a bounded lag if you only consider the first 2 tapes, but the third tape doesn't.