would be very expensive (about $4 million each in 1987), even if shared among five OCRs
($800,000 per OCR). If the memory size is reduced to 3 billion bits, the cost would be about $300,000 per memory, or $60,000 per OCR if used on a shared basis (as above).
In contrast, a local directory would require much smaller memory size (e.g., about
72 million bits for the Philadelphia metropolitan area). OTA estimated that the cost of a
typical billion OCRs
local directory would be times a multiplier of 3x). may not be necessary. But
about $20,000 (roughly $300,000 times 72 million/3 At this low cost, sharing a directory among several if shared among five OCRs, the cost per OCR would be
further reduced to about $4,000. In all likelihood, the capital would be small compared to either the total cost of multi-line directory data conversion and maintenance.
cost of local
OCRs or to
the cost of
Feasibility of Single-line to Multi-line Conversion
OTA reviewed the technical feasibility and cost of converting single-line OCRs to multi-line. OTA concluded that conversion would be technically feasible and that the USPS estimate of conversion cost -- $200,000 per machine -- is as good as can be developed from available information.
Conversion is relatively simple because a large part of the single-line OCR could be retained almost as is, as discussed below.
Letter sorter -- no change.
Letter transport -- no change.
The current mechanical transport
move faced and bottom function is common to
justified letters at a constant speed past an OCR single- and multi-line OCRs and would not change.
is designed to window. This