Data on metadata: on the question of indexing technical drawings

From the translator . We continue a series of articles devoted to the problem of creating electronic archives of project and design documentation. Today's article is devoted to the problem of indexing scanned drawings. Indexing (attribution) in this case means the analysis and subsequent input into the system of metadata, on the basis of which the classification, structuring and search of documents in the archive is carried out. We invite our readers to discuss and exchange practical experience on this issue.


There are no easy ways to index large documents. Often you hear about organizations that outsource projects to scan and index large volumes of images. Unfortunately, these organizations often do not pay due attention to the definition of specifications, standardization, verification of indexed data. As a result, numerous errors are detected after digitization. A few recommendations on how to avoid these errors will be given below.

The first point to consider is the following: those who carry out scanning and indexing, in most cases, do not understand the meaning of the scanned documents and are unlikely to understand it. They can ensure the quality of scanning, but the work on systematization and finding of documents in many respects falls on the shoulders of the customer.
')
The work on indexing large format drawings cannot be automated. There are no programs for such automation and may never appear.
The name of the drawing can be located either in the lower part or on the right or left. The text may be located in different parts of the drawing; in addition, numerous dates can be indicated on the drawing. Recognition of drawing fonts or handwritten inscriptions often does not bring the expected results. A drawing may have several developers: which one should I choose? The only possible solution in this case is to carefully review all the drawings, analyze the information and enter it into the system manually.

You need to understand data entry standards and abbreviations before you sit down at your computer. All queries in the database are based on parameters, not on the translation of information. Make a list of all abbreviations used and bring it to the attention of the employees participating in the project. Remember that data entry standards are advisory, not mandatory.

Having defined data entry standards and abbreviations, attention should also be paid to the punctuation used. It is recommended to abandon the use of punctuation marks altogether - unless, of course, they carry important semantic meaning (for example, hyphens and dots divide the numbers of drawings into semantic parts, and therefore they cannot be abandoned).

Finally, you need to constantly monitor the indexing process. Be prepared for constant questions like: “What do we enter here?”, Etc. Indexing is a team work. And with joint efforts you can competently arrange the drawing metadata and make the process of working with a document really convenient. Otherwise, why digitize documents?