A New Registry of Deeds Computer System

The Registry of Deeds will get a new computer system in 2019. I wrote the following article about the features it should contain for the Real Estate Bar Association of Massachusetts.

Thoughts on a New Registry Computer System

By Richard P. Howe Jr.

Just before Memorial Day 2002, a team from a computer company called ACS arrived in Lowell to begin the first-ever Massachusetts installation of the 20/20 land management computer application. Just a few weeks earlier, a group of registers of deeds assisted by personnel from the Secretary of State’s office had selected 20/20 as the preferred replacement system for registries of deeds in the Commonwealth.

The Middlesex North installation went live on July 1, 2002, and remains in use today. With periodic upgrades, 20/20 is still a reliable system but much has changed in the world of computers over the past 16 years. The time has come to replace 20/20 with a new system built with the latest features and capabilities. The formal selection process has yet to begin but our plan is to make the choice and begin deployment in 2019.

The new system will perform all existing computer tasks better and faster and will add even more capabilities. Here are some of the features I expect to be included.

Full Text Search. Optical Character Recognition (OCR) is an old technology that keeps getting better. OCR converts the print in a digital image into searchable text. Google has scanned millions of books, run the resulting images through OCR, and now makes the content of those books fully searchable. The same can be done with recorded documents. By running record book images through an OCR program, we can make the full text of all documents – the typewritten ones, at least – fully searchable. This would not replace our index but would supplement it, adding a new and powerful search tool for those who use land records.

Index Verification. Having an accurate index is essential to the registry of deeds. The 20/20 system uses a process called “blind rekey verification” which works as follows: At the point of recording, one employee types information from the document into the various fields of the index. Later, a second employee “verifies” those entries by viewing the document and retyping all the index entries without ever seeing what the first employee typed. If both entries match, that document is verified and the process is repeated for the next document. If there is a discrepancy, the verifier can view the original entries and then decide whether to restore them or keep the new values. Although rekey verification is more time consuming than visual verification (something we all call proofreading), it is also thought to be more accurate.

I believe technology is changing that calculation. With electronic recording, which accounts for a majority of all documents recorded at this registry, the customer makes all index entries and the registry clerk processing the document verifies those entries when the document first arrives. To have another registry employee then retype those same entries to verify them a second time seems unnecessary or at least inefficient.

Beyond that, OCR also has a place in index verification. The OCR application can detect names and numbers from the recorded document image, convert them to text and then compare that text to the values entered into the index by the registry employee (with walk-in recording) or the customer (with electronic recording). This verification-by-computer will ensure a high degree of accuracy in the spelling of names. But there is a qualitative aspect to verification – determining the proper document type, for instance – that will still require human judgment. And in those qualitative decisions, it is better that the employee doing the verification actually see what the first employee entered. For these reasons, I believe that the new system should move away from blind rekey verification to a more basic visual verification method operating in tandem with OCR verification.

Standard Searches Across Platforms. The 20/20 system employs two different search applications: one available at the registry of deeds and the other online. Although they pull from the same data and yield the same results, they look and respond differently. As more and more people use the website for their research rather than doing it at the registry, this has become less of a problem. Nevertheless, there should be a single search system and that system should work well on tablets and cell phones.

Hyperlinks. Long before anyone had heard the term hyperlink, the registry of deeds was using them. They were called marginal references. The 20/20 system has an automated marginal reference feature that works well but a new system could do better. For instance, a deed should have a link to the recorded plan that depicts the lot conveyed by the deed which would allow the user to quickly toggle between the deed and the plan or show both simultaneously. The new system will have this capability; 20/20 does not.

This list is far from exhaustive but it offers some ideas on how a new system might change the way registries of deeds operate. Still, the most important characteristic for a new registry computer system to have would be the flexibility to adapt to technological change that never stops.