User Guide

Organizing Automated Forms Processing
Data capture basics
Batch processing
Batches are collections of forms. Each batch has a unique iden
tifier. An important advantage of this approach is that it structures
information streams and facilitates administration, routing and
storage of data.
Operator specialization
Production capture solutions operate like assembly lines where
each person is responsible for a specific operation. Specialization
increases productivity and makes the system highly scalable. You
can easily add, say, more Scanning Stations and scanning operators
without interfering with the job of the recognition or verification
operators.
Scalability
As has been said above, the entire system consists of highly spe
cialized modules whose number can be easily increased or reduced
to meet particular processing requirements. Suppose the initial
configuration included one recognition module (a powerful two
processor server) and eight verification modules. If verification
becomes a "bottleneck", you can easily add any number of verifica
tion modules to increase verification throughput. This makes the
system very flexible and manageable and saves customers a lot of
money.
Processing queues
Batch routing is an important concept in data capture. Batch move
ment cannot be arbitrary but should be optimised to reflect the
logic of forms processing.
Depending on the stage of processing of a particular batch, the sys
tem a particular status to the batch. In a complex data capture sys
tem several batches can have "verify" status at a given moment.
They will be placed into the verification queue, and as soon as one
of the Verification Stations is freed up one of the queued batches
will be sent to this station. This allows the system to evenly distrib
ute the workload between the stations and operators, so that they
do not stand idle or become overwhelmed.
Data flows
To optimise batch processing, batches are routed and placed into
processing queues. If a problem occurs with any document (e.g. it
was poorly scanned), the problem batch will be immediately taken
out of the queue so as not to interfere with the processing of the
other batches. As a rule, problem batches are set aside to be
processed manually  the operator will have to identify the cause of
the problem and select the right solution. In this particular exam
ple the document will be sent to the Scanning Station to be re
scanned. It should be noted that the processing of the other
batches will continue at the same speed.