Paper dating use case

This is part of the submitted UseCaseList.

Scenario

A description of the scenario that you have in mind.

Goal

The user wants to date an undated document based on the particularities of the dated documents in the Bernstein paper collections and other collections connected to the Bernstein workspace.

The dating can be done on a single or a combination of paper particularities as provided by the databases (watermark type only, watermark morphological proprieties, in conjunction with the laid lines density, or chain line distance - see list in PaperAuthenticationUseCase).

Procedure

1. User uploads image to date
2. User selects the parameters on which to base the dating from a list provided by the machine
3. The machine and the user take measurements of the questioned paper according to the selected parameters
4. The above steps can be skiped if the user provide the measurements (as numbers and text)
5. The machine searches in the databases if there are papers with such characetristics...
6. ...in which case it produces a text report on the quantities of found samples, their distribution, the position of the questioned individum in the found population as well as...
7. ...an opinion on the dating of the questioned paper expressed as a number representing the probability of a positive match

Additional dating methods

Following is a list of additional dating methods. Bernstein partners should decide if they should be used.

1. Quires hint - If two books feature an identical sequence of watermarks, the probability is high that they were produced in the same location, at the same moment in time.
2. Sieve structure alteration - WM get deformed during their lifespan under the physical activity exerced upon them during the process of handling the paper making sieve by the paper makers. Thus despite that two watermarks are not perfect morphological matches it can be that their are in fact imprints of the same physical objet. Also some parts of the WM can become loosened from the sieve wire structure and fall off, or some new parts sown to the WM (a "crown" over an "eagle"), or the WM taken from one sieve and sown to another (in which case the rest of the sieve imprint (laid lines, chain lines, paper paste) might reveal the move).
3. ...

Choice of reference data set

Not all paper files in the databases provide the same type of information. Say one might give you the laid lines density, one not, but that one would provide info about the chain lines distance which the first dosn't. Depending on which data set the dating is performed, the result might be different.

                                 1111111111222
file nr.                1234567890123456789012   qnty/type
WM sub-elements         --     ----               6
laid lines density          --------------       14
chain lines distance    ---------      -------   16
qnty/file               2211222332211112221111

The above graphics shows that the quantity and composition of the reference data set varies in function of the selected reference characteristics. For example a sieve can be dated by only relying on the chain line distance - in which case we have a large amount of data on which to rely (16 files out of 22); or it can be based on a more divers sample, say including information about the WM sub-emlements, the laid lines density and the chain lines distance - but in that case there are only two files to rely upon. The question is now to understand what difference it makes to do the dating this or the other way.

Statistics

The user has to be provided with statistical information about the dating (such as the number of record on which the dating is based in relation to the total number of record in the databases). The user needs also to know if different sets of data where used for different datings he has undertaken. And also what the error probability of the dating is. For these aspects see the DbSearchStatistics use case.

See the logical scheme of dating.

Importance

How important do you see this use case as?

• Must Support

Dependencies

What other use cases are affected by the implementation of this one?

• Databases usability / db integration / authentication

Input

Things that the user must/might supply to the system.

• Mandatory: Image to date or charcteristics thereof

Output

Things that the user will recieve in response to their request.

• number of matching samples in the databases and the statitical make-up of that population
• dating opinion

Difficulties

Areas in which you foresee problems/issues arising.

• communication with the databases
• amount of time necessary to search in the databases for items with specific characteristics
• absence of landmarking data for the WM stored in the DB

Example

An example supporting this use case.

• See above scenario description.

Other Information

Any other information that you think is important to include.

• ...

Comments

Comment from other partner regarding the use case.

• Please look at my comment in CartographyUseCase. Its will be similarly for this use case. (Where are which wire thickness, chain- and laidline distances in which period of time.) GeorgDietz

• "Quires hint - If two books feature an identical sequence of watermarks ..." -- within a ream of paper there are sheets produced with two moulds belonging to a pair of moulds, so one has to explain the meaning of "identical sequence of watermarks"; cf. DbSearchWatermarkTwins.- FriederSchmidt

-- VladAtanasiu - 19 Oct 2006