Page: image of rounded scales watermark]

The Watermark Initiative

© 1997 Robert W. Allison. All rights reserved.

Commentary and Interpretation of the IPH Standard

with Proposals for a WWW Distributed Database System

The individual sections of this Commentary on the IPH Standard are linked to the list of IPH Data Fields in our Comparison of Data Fields in the IPH Standard and the WWW Distributed Database, which serves as a table of contents to this Commentary.


The objective of the Watermark Initiative is to devise a standard for a WWW-based distributed database system for the registration of watermarks and description of papers -- a standard which would be consistent with the IPH standard. This commentary documents our reading of that Standard.

The fundamental question faced by the Watermark Initiative is how to translate this Standard as effectively as possible from one medium to another, that is, from the medium of a free standing in-house database for consultation on the premises of the host institution, to that of an integrated WWW-based system of distributed databases in which a search from any one can be a search of all databases participating in the system. The objective of consistency with the IPH Standard requires interpretation of the Standard, which is not a cut-and-dried process. In many cases alternate translations and interpretations of the Standard can be imagined. This Commentary is intended as a basis for a public discussion and debate focussed on this process, in order to be assured that the final product will be true to the objectives of the IPH Standard and at the same time take full advantage of the open environment and interactive potential of this new medium, the WWW. As the discussion and design work progress, this Commentary will continue to grow.

An electronic discussion group named Watermarks Listserve has been established by the conveners of the 1996 Roanoke Conference as a forum for this discussion and debate. All interested persons are welcome to sign on and have their say. Contributors to the discussion are urged to reference their comments by citing IPH field numbers and/or (for expansions of the IPH Standard) our own www-based model presentation pages and field names:

  • Paper
  • Watermark
  • Facsimile (i.e., identification and type of facsimile; and orientation of the facsimile)
  • Mould (i.e. data on moulds which are still in existence)
  • Reference (i.e., editions, literary references)
  • or to one of the other data files accessed from the above pages:

  • Artistic (summary identification of the intellectual content of the paper-bearing object. E.g., literary or documentary texts, musical compositions, illuminations and works of graphic art, etc.)
  • Job (data about an employment or working residential relationship between a person and an organization)
  • Organization (identifications for organizations like publishers, printers, scriptoria, ateliers, binderies, paper mills, etc. All libraries and collections, institutional and private, are treated as "Organizations" in this WWW database system, so that private collections without formal or official names are entered under the name of the owner: The Thomas Gravell Collection.)
  • Person (basic prosopographical data for persons associated with papers (scribes, printers, artists, binders, papermakers, restorers, forgers, etc.)
  • Physical (codicological data for identification of the type of physical content of the paper: writing, printing, painting, glazes, etc.)
  • Seals (codicological data for description of stamps, seals and validations which, by their association with papers, are useful for dating, provenance and history of use of paper)
  • Source (official identification for paper bearing objects such as manuscripts, books, musical scores, drawings, broadsides, etc. Identification consists of country, city, name of institutional or private owner (all drawn from the Organization File); collection name, and shelf numbers or other unique and official access number)
  • The WWW as a Medium of WWW Publishing

    A WWW-based database has more and different potentials than a free-standing archive that might be provided and maintained by an individual organization. What are these potentials?

    To take full advantage of the WWW environment, the database design needs to be adaptable to different institutional pocketbooks and platforms, able to be adapted to specialized interests, and easily maintained by the institutions that will adopt it. As a research tool, It needs to be practical for data collection in field work as well as for searching. As a form of publication, it needs to provide for credit to the person or institution publishing the data, and a way of handling updates and revisions. It must also make it easy for private scholars and small institutional archives which lack the facilities to establish their own WWW servers to mount their databases through the services of a participating research institution or an affiliated consortium of such institutions, while still crediting the originating scholar or institution.

    Some General Observations About the IPH Standard

    We have proposed, in our working model, a number of changes from the principles and methods of the IPH Standard. Our proposals for changes are included in the commentary that follows. These changes represent an effort to think through the design implications of the WWW environment and in particular, to achieve the higher level of rigor and specificity in database design that are required in order to take advantage of that WWW environment. This last objective needs to be balanced against the danger of becoming overly restrictive.

    We begin, therefore, with some general observations and critiques of the IPH Standard in its present, provisional edition of 1992)

    The Commentary

    3.0.0. Entry Number of the database or card index

    Is this a numbering scheme intended perhaps as a cross-index to a paper-based system (card index), or is it a system for registering databases (some of which currently exist in the form of card files?

    If the latter is its purpose, then it is no longer needed in a WWW-based distributed database system. The WWW offers the potential for a seamless search of www-based archives all over the world. A single query can access all www-based archives which adopt the same standard and are registered in a central index of participating archives. This field is therefore not needed in the WWW-based database; the URLs for the individual archives replace it.

    3.0.1. Date of entry or revision

    The IPH Standard envisages a date of submission as part of every record (paper description) in the database. This field is not retained in the present proposal for the WWW archive. We currently envisage, rather, a WWW publication model in which any archive will make changes and release revisions of the archive as "editions." The identification of the particular archive and the edition/revision number and date will appear in the standard header or logo of that archive. This header will appear at the top of every view of data drawn from its database in response to user queries. (For example, see the Watermark Initiative header at the top of our model pages.) Citations in other publications drawing upon data from these WWW archives will, correspondingly, identify database, the edition cited and its date.

    It will be the responsibility of each institution maintaining an archive to keep copies of earlier editions and revisions. Editions and revisions of databases can also be issued on CD ROM for use by institutions or individuals in-house or by scholars engaged in field work. CD ROM releases would be especially useful to those institutions which lack access to the WWW, in which case these CD ROMs could also serve as a record of earlier editions of an Institution's database.

    Another possibility might be for the individual archives to maintain a "change history" file containing multiple dates and descriptions of changes. We have not yet attempted to work up a design for such a file in the present model.

    3.0.2. Whereabouts of Institution operating the watermark database

    This data field title indicates that it pertains to the host institution of the watermark database. Because this data field occurs in section 3.0, however -- "Data concerning the sheet," -- and because the following two fields (IPH 3.0.3-4) refer to the paper bearing object, one would naturally expect here information about the owner of the paper bearing object, not the institution operating the watermark archive. Did the authors of the IPH Standard presume that watermark archives would be established by the same institutions possessing the paper bearing objects (museums, libraries and other owners of books, manuscripts, documents, etc.)? Or was this field inadvertently merged with another for the identification of those Institutions?

    Of course, the institution operating the watermark database cannot be presumed to be identical with the institution owning the paper bearing object. So we take it that this field must have been intended to pertain to the holder of the paper bearing object. Cf. Maria Dolores Diaz de Miranda and Ana Maria Herrero Montero, Registro de Filigranas Anteriores a 1500 en los Archivos y Bibliotecas Asturianos, II Congresso Nacional de Historia del Papel en España, Cuenca, 1996: in this Spanish implementation of the IPH Standard, IPH 3.0.2 is likewise used for identification of the owner of the paper bearing object. (Reference courtesy of Tomás Stohr.)

    Identification of the Watermark Archive

    In the WWW-based distributed database system, the WWW pages of individual archives will bear their institutional identification in their individual headings and logos, just as the model pages of the Watermark Initiative Website do. Similarly, for purposes of searching these archives, this information is embedded in their URLs. In other words, identification of the watermark archive and its host institution is built into the working of the system. Nevertheless, the person inputting information needs to enter the key or ID for the institution which makes it possible for internal links among the various files making up a record to work. More on this below.

    Where, then, is the institution operating the watermark database identified, if not here? And what does the WWW system do about the IPH Standard's requirement of a "compulsory list of subcodes (abbreviations of all institutions operating a watermark database! e.g., FPBN = Paris, bibliotheque Nationale)?

    In the WWW system, all institutions -- holders of paper bearing objects as well as operators of watermark archives and others like binders or publishers or paper mills mentoned anywhere in the database -- are described in records located in the file called "Organization."

    NOTE: This holds true also for individuals who (like institutions) maintain watermark archives publically accessible through the WWW distributed database system.

    The Organization file input form asks for an ID -- an abbreviation or code -- for every organization. This ID serves as a key linking related files to the organization file. The IPH's existing list of compulsory codes will serve this purpose initially in the WWW archive system, but the list will have to be rapidly extended to accommodate the additional organizations. Persons inputting records of papers and watermarks will be able to select the correct IPH ID (code) from a menu, or create a new one. Thus, the compulsory aspect of this code disappears in the WWW system. Institutions will be able to use any abbreviation or code they choose -- the only requirement is that they use it consistently -- since the code serves primarily as an internal link among the related files making up a record within an institution's archive. This feature of the WWW system design eliminates a major administrative headache, and reinforces the basic principle of local administrative autonomy or independence in archive management.

    NOTE: It is conceivable, of course, that someone might want to search by the name of the host institution of a watermark archive, and a laissez-faire system of abbreviations could result in more than one way of identifying an institution so that a search by one of the abbreviations would not find those records in which a different abbreviation was used. But this shortcoming can be remedied in other ways -- for example, searching by the name of the city in which the institution is located. This minor inconvenience, we feel, is far outweighed by the advantages of reduced central administration of the system and elimination of the need for users of the system to work with codes (and therebyelimination of the risk of misreading them or making keyboard errors).

    3.0.3-4. inventory or accession no. of the paper-bearing object (e.g., MS gr 0219) and subunit (e.g., folio no. in a codex)

    Each sheet of paper, according to the IPH Standard, is identified by the official accession number or shelf number of the paper-bearing object or artifact (3.0.3; IPH example: "Ms Gr 0219") plus a subordinate number ("subunit") identifying the specific piece of paper within it, such as the folio number in a book or codex manuscript (IPH 3.0.4).
    Note: this approach differs from that of the Greek Mss Watermark Archive presented at the Roanoke Conference, in which "paper types" (identified by arbitrary serial numbers) take the place of the "subunit" (folio number). The IPH method is preferable because it eliminates the analysis required to identify a paper type, leaving that to the user, and allows for multiple examples of the same paper stock to be entered in the database as separate records.

    The shelf number and subunit Id alone are not a sufficient identification of the paper bearing source, however; nor is a single additional field sufficient for identifying the institution which owns the original paper bearing object. A full identification of the source requires, as indicated in our comments on IPH 3.0.2, four separate fields in addition to the shelf number: country name, city name, institution or owner name, and (to accomodate larger institutions with multiple collections) collection name or identification. The IPH Standard does not break down this information into separate fields, so we have added them in the WWW database design. These additional fields for identifying the institutional or private owner are located in the Organization File in our model.

    Larger archival institutions whose collections are subdivided into subcollections or series, or files and dossiers will have to create composite entries for their "collections" and "sources" (paper bearing objects). In the partially fictional example below, "collection" is a composite of collection and sub-collection, and "source" is a composite number combining file and volume numbers.

       country      Venizuela, 
       city         Caracas, 
       name (inst)  AGN (National General Archive)
       collection   Archivo General de Indias, 
                       Consul General's Office Records
       source       19,20 (= file 19, vol. 20, 
                       where file 19 = records for 1590-91)
       subunit      fol. 16 
    NOTE: A "composite" subcollection name or "composite" source number appears to violate the principle of only one type of data to a data field. In this case, however, we are in effect imposing a requirement on participating institutions that they create a unique collection title and a unique source number such that there will be, for purposes of the database, only one collection and only one source number. This can be thought of as analogous to our requirement that the institution hosting a watermark archive invent an institutional code and use it consistently. In this case, the alternative would be a database made extremely cumbersome by multiple levels of nested (hierarchical) collections and source numbers.

    In the above example, the institution might have files called "Consul General's Office Records" in several major collections representing different nations. The Institution could simply identify this one as "AGI Consul Genera's Office Records", as opposed, say, to "AGM (Mexico) Consul General's Office", etc.

    IPH 3.0.4 (the "subunit" specifying the exact piece of paper) is named "Paper_Unit_ID" in our model. this field identifies the piece of paper being described by folio or page numbers in manuscripts and books, or by sheet numbers in dossiers of documents, rolls or scrolls. (We suggest "Paper_Unit_ID" rather than the IPH term, "subunit" for the sake of more intuitive language to avoid confusion; "subunit" seems to relegate to a subordinate level what is actually the primary unit of identification of papers.)

    3.0.5-6. Bibliography

    The purpose and content of IPH 3.0.5, characterized as "References of a bibliography," and its relation to the following field (IPH 3.0.6, "Literature references") are not clear from the IPH description, which seems to raise a number of questions:

    Our model envisages for every participating archive a simple Bibliography Page as a resouce document separate from the data base and accessed by links from within the data base (a simple HTML page). Each participating archive may use whatever format for bibliographical entries it prefers. We do not feel that we should be in the business of establishing a world wide standard for bibliographical format.

    In our model, therefore, the concept of a data field for standardized abbreviations (or, more correctly, a set of data fields) is abandoned in favor of HTML links to a free-standing html page. Accordingly, IPH 3.0.5-6 disappear from the list of fields.

    3.0.7. Kind (i.e., kind of paper)

    IPH offers the following options:

    This IPH field presents several problems:
    1. Classification of papers by some of these categories may involve arbitrary judgments, for they do not appear to be mutually exclusive terms. Heavy paper that might be commonly used, for example, for drawing or painting, may also occur as folios in large codex manuscripts or used for charters, maps, or soft bindings.
    2. As presently conceived, this classification MAY overlap with 3.0.8, END USE OF PAPER, many of which are also original uses of the paper. Writing paper, for example, would appear in many cases to be an "end use of paper" rather than a distinct kind of paper. In both cases (type of paper, end use of paper), an unstated idea seems to be that the paper may have been used for something other than what it was "intended" for. How should the ambiguity of these categories be resolved?
    3. Marbled paper is listed as an example here and an option in IPH 3.0.13 Coloring? Is this a redundancy? Why is marbling included in a field where the fundamental idea is the paper's FUNCTION?
    4. The list of options appears, rather, to constitute a list of specialty papers, which does not allow for ordinary paper. What option would you select if you were describing, say, ordinary papers used in a a plain 16th century manuscript or printed book? Do we need to add an option called "ordinary stock" to meet this need. (see our model page for paper description.) Alternatively, would it be better to rename this field "Specialty-Kind" and provide an option, "Not-Applicable" for cases of ordinary paper found in books and manuscripts?
    In order to eliminate the need for arbitrary selection among any of the remaining options and/or other options that might be added to this field, and foreseeing that it may not be possible to achieve perfect mutual exclusivity of options, we propose, also, that the options be listed in the data input form as check boxes, so that more than one option is allowed per paper.

    3.0.8. End Use of Paper

    Options listed in the IPH Standard include (with some clarification here provided):

    This is a long list comprised of options which are not analogous kinds of terms, and are therefore of interest to researchers for different reasons. In our WWW-based Archive, this single field has therefore been divided into two separate fields which occur in two different files. One of these fields occurs in the description of the paper-bearing object, where it is named (and thereby redefined) as "Object_Type" (i.e., type of paper bearing object). The other, retaining the field name, "End_Use," occurs in the Paper file, where it identifies the end use of the particular piece of paper. This strategy solves several problems with the way that the field was conceived in the IPH Standard:

    1. Some of the IPH options for 3.0.8 are types of paper-bearing objects (e.g., document, manuscript, music book, and others), which may in fact include many different papers. Other options are (as the field name indicates) uses that a particular piece of paper may serve (e.g., endpaper, envelope) and which may or may not be found in various kinds of paper-bearing objects. This mixing of categories blurs the line between description of a piece of paper and description of the object bearing those papers. Division of the field name into two, distinct fields resolves this problem.

    2. Secondly, some of the end uses listed among the IPH options for this field are also original uses. In these cases, researchers may be interested in associating the paper with the type of the original source (i.e., paper bearing object), and secondarily perhaps with scribes, publishers, artists, etc. In other cases (e.g., end leaves), the paper may have been used originally (for example) as book folios then re-used (end use) as flyleaves in another book. In these cases, researchers may be interested in identifying restorers and binders by the papers they used. Both the type of paper bearing object of original use and the end use of the paper may provide the researcher with critical information (for example, to account for the limitations of the data provided regarding that piece of paper in the data base). It is clear, then, that two types of information are combined in this field, creating a problematic condition in the construction of a data base. This suggests that the field, "End_Use" should apply only to re-used papers, and that a field to indicate the "Original_Use" of the paper is needed, with a similar list of options.

    3. Some additional option categories for "End_Use" can perhaps already be foreseen, but can be added as need arises :
      • replacement folio in ms book
      • replacement page in printed book
      • restoration
      • inserted folio (illustration, text supplement, etc.)

    In our model therefore, we apportion the options for these newly defined fields as follows:

    Object Type End Use
  • printed text (book, broadside)
  • document
  • drawing
  • envelope
  • map
  • manuscript letter
  • bound ms
  • ms notebook
  • loose ms notes
  • ms score
  • printed music
  • painting
  • print (work of graphiic art)
  • other kind of object
  • backing paper (drawing or print)
  • endpaper folio
  • endpaper bifolio
  • replacement folio in ms book
  • replacement bifolio in ms book
  • replacement page in printed book
  • paper restoration
  • inserted folio in a ms book
  • inserted bifolio in a ms book
  • palimpsest ms folio
  • palimpsest ms bifolio
  • plain sheet (unused)
  • To these should probably be added a field, Original Use with a list of options similar to "End_Use", and maybe even a field "Intermediate_Use."

    3.0.9. State of Sheet
    The purpose of this field, as currently set up in the IPH Standard, is to indicate whether the paper being described is an uncut sheet, or an alternative to that condition (either a fragment or a trimmed sheet). It provides the researcher with information that helps with various kinds of analysis of data in the database (e.g., whether the dimensions reported in 3.0.10-11 are actual sheet dimensions or estimated ones; how to align the chain compartments in order to compare two pieces of paper and their watermark; whether the absence of a countermark may be accounted for by trimming; etc.).

    The term, "state." in this field name, however, prompts us to think of this item as part of a description of the state of preservation of the paper , referring to the substance of the sheet . We have renamed IPH 3.0.9 with the more concise label, "Sheet_Integrity" to avoid confusion with State of Conservation, which we have proposed as a new field to the Standard (see below)

    State of conservation of the paper is an omission in the IPH Standard that needs to be rectified. State of conservation should provide descriptors to account for such conditions as deterioration of the fibre network, acid-induced brittleness or holes, fading (cf. 3.0.14), etc. It thus pertains to the state of preservation or conservation of the substance of the paper. We have added this new field in the Paper Description file.

    3.0.10-11. Sheet Dimensions

    In most cases, sheet dimensions must be deduced from analysis of page size and book format (in printed books or codex mss), from location of the watermark(s), etc.. This involves a scholar's judgment and time consuming research to find other examples for confirmation of estimates. Unless the paper in question survives in an uncut sheet, the figures given here will remain only a hypothesis, requiring confirmation by others. In fact, figures provided here based on such analysis are likely to cause error and confusion, since they are very likely to be taken uncritically by inexperienced users of the archive. Inexperienced users of this database will not think to use this data only in conjunction with other information (Sheet_Integrity; Object Type).

    In our model, we propose to alert users to the danger here identified by adding a field, which can appear as "radio buttons" in the input form, to indicate whether the sheet dimensions are "measured," "calculated," or "estimated," and this information will appear as part of the presentation of the data.

    Explanation of the sheet width calculation in our model paper description:
    Sample Explanation:
    In this quarto codex, the folio height lies on the axis of the sheet width. In this example, the folio height in the ms measured 282 mm. The watermark was 30 mm off-center vertically in the gutter, so the vertical dimension (282 mm) was adjusted to 342 mm (doubling the 30 mm and adding it to the height dimension). Then doubling that figure (684 mm) because of the quarto format, allowing for 5-6 mm trimming on the horizontal edges, quadrupled for the quarto format (a loss of 20-24 mm) yields a calculation of 704-708 mm.

    Explanation of the sheet height calculation in our model paper description
    Sample Explanation:
    In this quarto codex, the folio width or the horizontal axis of the bifolio lies on the axis of the sheet height. In this case, the width of the folio measured 220 mm. (with no loss of exposure at the gutter). Then doubling that figure to obtain the bifolio width (440 mm), allowing for 5-6 mm to compensate for trimming on the outer edges and doubling the trimming to include both sides of the bifolio yields a calculation of 450-452 mm. for the sheet height.

    3.0.12. Local Standard

    Identification of a particular paper with a particular local standard is a case of analysis and involves someone's judgment informed by research.

    In the database, this is a write-in field; no standardized terminology is attempted here. Consequently, this information will not be a good search parameter, but rather will provide corroborative information for persons making comparisons of papers.

    3.0.13. Color of the Sheet

    The IPH standard allows for the following options:
    Why is "marbled" mentioned both here and in 3.0.7, Kind of Paper? The IPH Standard suggests that "marbled" is an appropriate category in 3.0.13, since all options describe methods of dying of paper; its mention in 3.0.7 is simply to indicate that if you have marbled paper, the appropriate descriptor in that field would be "decorated;" but see our comments on 3.0.7, Kind of Paper for problems in the definition and purpose of that data field.

    We have suggested a field name "Coloration" to avoid confusion with 3.0.15, where colors proper are entered.

    3.0.14. Intensity of colour

    This element of description is inevitably a matter of individual judgment. The IPH standard offers three options for standardized description: Of these categories, "faded" involves a judgment about the original color and the state of preservation of the paper. This descriptor ought to be part of a field for recording the present state of preservation of the paper (cf. our comments under IPH 3.0.9), since, unlike other kinds of data, this item indicates that other sheets of the same batch of paper may be of different color intensity.

    We propose that this field be defined for use with colored papers only, not natural or white papers The descriptors for colored papers would include:

    The field should be left blank for natural or white papers.

    3.0.15. Tone (Color)

    For natural papers we propose that natural and white papers be treated by a separate set of descriptors from colored papers (below) In addition, borderline cases could be described with ranges, which would be caught by any search looking for one of the basic terms (beige, cream, brown, white).

    White and Natural Papers:

    Dark color tones:
    dark beige (greyish-brown tones)
    dark cream (yellow tones)
    dark brown
    Medium (off-white) tones:
    medium beige
    medium cream
    medium brown
    Bright tones:
    bright cream
    bright white
    Borderline (partial list of examples):
    medium-bright cream
    dark-medium beige

    Colored Papers:
    For purposes of field work, a printed set of color samples and their descriptors should be developed, published by the IPH, such as are used by paper companies for marketing purposes. (The variation of representation of colors on computer monitors is so great as to assure lack of consistency!) Such a tool could become a standard gauge -- of both color and color intensity -- to be kept in all libraries and research institutions or other contexts in which papers would be described for addition to the WWW archive system. To allow for the subjectivity of color judgments, clusters of similar colors should be described with the same term, so that searches for that term would catch a wide-enough variety of closely related colors to make the search a viable one.

    To accomodate marbled and veined papers, and other special kinds of paper, entering more than one color should be permitted in this field.

    3.0.16. Pattern

    The IPH recommends establishment of a separate database for marbled papers by collections possessing large numbers of these papers. This need might best be met by use of a digital camera and on-line color images. Otherwise, the present field will inevitably depend on individually composed descriptions written in a wide variety of languages. As such, it will not be a searchable field, but rather, will provide information that could serve a corroborative function in the process of matching papers. This is a subject about which we know little, however, so we have made no attempt to interpret it or adapt it for the WWW-based archive.

    Searching for marbled papers might be accomplished by searching for a set of colors (in 3.0.15) and limiting the search to marbled papers (in 3.0.13).

    3.0.17. Side

    This field does not contain descriptive data about the paper, but rather, a piece of information about how the watermark facsimile was generated. It is essential, in attempting to match watermarks and papers, to be assured that two watermark prints or tracings were taken from the same side of the paper. Because this data field pertains to the generation of the watermark facsimile, it is located in the Facsimile File in our model.

    Submitters indicate whether the prints or tracings of watermarks were made with the impression side (that is, the wire or mould side, as opposed to the "felt side" or smooth side) facing down or up. A third option is "undetermined," which alerts users to the fact that they must try matching papers by viewing the watermark image from both sides.

    3.0.19. Fillers described in full (starch, china, clay, etc.)
    Data from analysis; for machine-made paper.

    3.0.20. Sizing type described in full (starch sizing, animal sizing, etc.)

    Data from analysis; for machine-made paper.

    3.0.21. Degree of Sizing. (full sized; half sized; unsized)
    Data from analysis; for machine-made paper.

    3.0.22. Type of Coating described in full (wheat starch, latex, casein, etc.)
    Data from analysis; for machine-made paper.

    3.0.23. Mode of coating (absent; double side; single side)
    Data from analysis; for machine-made paper.

    3.0.24. Dyestuffs described in full
    Data from analysis; for machine-made paper. IPH Standard specifies "if possible, in accordance with the Colour Index."

    What is the color index that is referred to here? Does the IPH envisage the development of a color guide like that described above (3.0.15)?

    3.0.25. Ash content (% according to DIN 54370
    Data from analysis; for machine-made paper.

    3.0.26. Smoothness (index according to Bekk (DIN 53 107)
    Data from analysis; for machine-made paper.

    3.0.27. Weight (weight of sheet converted to g/m2
    Data from analysis; for machine-made paper.

    3.0.28. Thickness of sheet in µm (according to DIN 53 105)
    Data from analysis; for machine-made paper.

    3.0.29. Paper grade described in full (according to DIN 6730)
    Data from analysis; for machine-made paper.

    3.0.30. Brand name in full in original spelling
    Data from analysis; for machine-made paper.

    3.0.31. Type of paper machine used for production
    IPH Standard offers the following options: As elsewhere in the IPH Standard, no way of specifying the "other" is suggested. It is the goal of our WWW model that the person entering the data should always have the option of writing in a descriptor if the descriptors offered as options are not appropriate (or in this case, to explain the choice of "other"). Idiosyncratic language will not make for efficient searching by users of the archive, but a search for "other" will still be useful to the researcher.

    3.0.32. Mechanical watermark application technique
    IPH Standard offers the following options:

    3.1. Watermark Data
    IPH Standard specifies that for each "kind" of watermark encountered in a sheet, a full description consisting of 12 descriptors (IPH fields 3.1.1 - 3.1.12) must be recorded. (We take it that 3.1.13, the data field for general information, was also meant to be included in this full description of every kind of watermark.) The "kinds" of watermarks are listed in 3.1.0 (see comments on 3.1.0 below).

    To accomplish this purpose, the IPH Standard calls for multiple values (descriptors) in each of these 12 (13) data fields, one for each of the kinds of watermark in the paper. These multiple values are to be numbered serially to coordinate with corresponding serial numbering of the watermark kinds listed in the first data field in this section: 3.1.0, KIND (of watermark). Thus, in the IPH example, KIND1 is the main watermark, KIND2 the countermark.

    The numbers are arbitrarily assigned by the person inputting the data. Thus, in a paper with no countermark, KIND2 might be, rather, a border watermark or a corner watermark. If 2 different kinds of corner watermarks were found, they might be numbered KIND2 and KIND3.
    In field 3.1.2 (Structure description) in the IPH example, the submitter will manually enter the code, "STR1" (= "Structure 1") into the field followed by a descriptor for the structure of the main watermark, and "STR2" and a descriptor for the structure of the countermark.

    This approach introduces two problems. First, it requires multiple values in a single field, contrary to the principles of entity-relationship database design. Second, this approach will not meet the needs of a WWW-based archive system, in which descriptions of different sheets of the same paper stock may be entered by different submitters in different archives, because these different submitters will use different arbitrary serial numbering for the different kinds of watermarks encountered in that paper, producing inconsistency among the various descriptions of the same paper and thus confusing the matching of papers by subsequent searchers. Even in the independent, free-standing databases envisaged by the IPH (leaving aside question of adapting the IPH system into a form suitable for the WWW), this method will significantly increase the liklihood of human error (errors in correlating these arbitrary numbers).

    Our model for a WWW-based archive system presumes creation of a separate record for each of the kinds of watermarks identified in the IPH Standard that might be found in any paper, namely:
    • No Watermark
    • Main Watermark
    • Countermark
    • Doublemark
    • Border Watermark
    • Corner Watermark
    • Dividing Lines

    Each such record consists of the same set of data fields for watermark descriptors, and each record is keyed to the particular piece of paper being described. When researchers viewing the Paper description page click on the "Watermark" hotlink at the bottom of that page, the system presents a list of the watermarks (i.e., kinds of watermarks) from that piece of paper for which descriptions are available. Researchers can then select any of the various kinds of watermarks in that paper for which they want to see descriptive data. Because the various kinds of watermarks are recorded as separate records, the problem of sequencing the descriptive data for several watermarks in a single field is eliminated.

    In the WWW model, then, the coordination of description fields with the various kinds of watermarks is automated and standardized. The result is that both the person inputting data and the searcher can concentrate on the description without worrying about coordinating fields by means of correct sequencing of watermark kinds.

    In both the IPH standard and the WWW standard, "No watermark" indicates the absence of any watermark at all; it does NOT mean that an existing watermark has not been described.

    3.1.0. Kind of watermark(s) in the sheet
    The IPH Standard offers the following options:
    At least one aditional option, the double mark (a pair of watermarks which occur in the centers of both sides of a single sheet), is needed here. .

    The IPH Standard indicates that subcodes are needed to describe papers in which several of these kinds of watermarks occur (KIND1, KIND2, etc.) and that the descriptor for each kind will be different. This use of subcodes, however, is a case of multiple values being entered into a single field (see our comments under 3.1, above). In the WWW model, a separate record is created for each type of watermark found in any piece of paper. This field, as part of that record, identifies which watermark is being described and serves as the key by which the database locates the record and lists it in any printout of search results..

    3.1.1. Structure of the watermark
    The IPH Standard offers three options for this field:
    • Line watermark
    • Shadow Watermark
    • Combined line and shadow watermark

    3.1.2. Position of watermark in the sheet
    The IPH Standard conceives of the uncut sheet of paper as divided into 9 zones in a 3x3 pattern, and thus offers 6 options (codes) for defining these zones: centre (C) lower part (D) left part (L) in the middle (M) right part (R) upper part (U)

    The IPH codes are not sufficiently defined. Which of the following interpretations is intended? (each table represents an uncut sheet of paper divided into 9 zones)
    SCHEME 1 left middle right
    upper ==> UL UM UR
    center==> CL CM CR
    lower==> LL LM LR

    SCHEME 2 left center right
    upper ==> UL UC UR
    middle==> ML MC MR
    lower==> LL LC LR

    The second scheme would seem to be what the IPH Standard has in mind, judging from use of the term, "center," in 3.2.3, and the term, "middle," in 3.2.4. Users of the IPH Standard should not have to go to this amount of trouble to figure out which is intended.

    In addition, the IPH Standard is unclear about how to define top versus bottom and left versus right of a sheet. Top and bottom cannot be determined by the alignment of the watermark, since in some cases the vertical axis of the watermark design may not parallel the chain lines (the vertical axis of the paper sheet) but rather, the wire lines (the horizontal axis of the paper sheet).

    Finally, when the paper is from (say) a book or codex manuscript in quarto, the fold between the left and right will have become the top or bottom edges of the codex and the codex will usually have been trimmed at the top and the bottom, making it unclear which bifolio is the left and which the right side of the full sheet. This descriptor is useless, then, unless the paper to be described is a full, uncut sheet.

    In all other cases, therefore, this field calls for analysis, and in most cases it will be left blank by persons conducting field work or feeling pressured for time.

    3.1.3. Watermark Motif
    The IPH Standard expects this field to be filled in with a code from its "Index of Classes and Sub-classes" of watermark motifs. The exact description reads as follows:
    Motif of watermark (abbreviated) coded according to appendix I

    In our model, we would like to use straightforward standardized language and eliminate the codes. In our model watermark description page, however, we have for the present retained the IPH code for discussion purposes.

    3.1.4. Class of Watermark
    As we read this section of the IPH Standard, it expects this field to be filled in with a full description of a watermark motif, expressed in the IPH alphanumeric code (IPH Standard, Addendum 1) and describing not only the main motif, but also its details, such as parts of the whole and their position in the watermark (cf. IPH Standard, section 5). The exact description reads as follows:
    Class of watermark with subclasses, coded according to section 5 and appendix I, with addendum I.
    The code is to be entered manually into the field by each individual submitter, interpreting the IPH alphanumeric code. One might compare, in a printed catalog, the coded descriptions provided in David Woodward's Catalogue of Watermarks in Italian Printed Maps ca 1540-1600, Chicago, University of Chicago Press, 1996.

    The IPH field name, "Class of Watermark," is therefore a misnomer, which does not describe what in fact the Standard expects to be entered in this data field. The IPH defines "Classes" (in Appendix I) as the names for the watermark motifs listed in that Appendix, and those classes and their codes constitute a classification system. Each item listed in the Appendix is either a Class or Sub-class. A better heading for this field would be Coded detailed watermark description

    Our objective is to use straightforward standardized language and eliminate the codes. For the present, however, we have retained the IPH code In the model watermark description for discussion purposes.

    3.1.5. Full (extensive) description of watermark
    The IPH Standard expects this field to be filled in with an extensive description of a watermark motif, written by the individual submitter in his or her own words. This will not be a searchable field, therefore, since the language is not standardized and there is, moreover, the problem of the use of different languages by submitters of different nationalities. Its purpose would seem to be to serve as a corrective for any possible misunderstanding of the coded description in field 3.1.4.

    3.1.6. Height of watermark in mm
    The IPH Standard calls for the watermark to be measured as a rectangle whose height parallels the vertical axis of the watermark, using a rectangular sliding rule. Thus, the IPH Standard defines the height dimension as parallel to the vertical axis of the watermark design, NOT (as per Briquet and others) as parallel to the chain lines. In some cases this makes for an arbitrary determination of which axis is the vertical one. What would be the vertical axis of a triple crescent moon watermark, for example? Are we to consider the crescents as open along the horizontal axis -- ( ( ( -- or along the vertical one (the open sides facing upward)? Since it is not always intuitively obvious which axis is the vertical one in a watermark design, and since such a determination is a form of analysis (and since not all users will make the same determination) it would seem safer to define the vertical axis of the watermark in purely mechanical terms: as paralleling the chain lines (as per Briquet), which eliminates analysis and variability. Thus, the vertical axis of the watermark is the same as the vertical axis of the sheet. That is the method adopted in the Greek Manuscripts Watermark Archive, and that is the method we propose for the WWW-based Watermark Archive.

    3.1.7. Width of watermark in mm
    To be measured, like the vertical dimension in IPH 3.1.6, with horizontal axis defined by the watermark design. The same criticism applies here as in the determination of the vertical dimension above. Accordingly, in our model, the width of the watermark is defined mechanically as the axis paralleling the wire lines, regardless of the watermark design. Thus, the horizontal axis of the watermark is the same as the horizontal axis of the sheet.

    3.1.8. POSL (Position from chain to left of watermark)
    The smallest horizontal distance between the watermark and the nearest chain line on the left, in mm. The rationale for 3.1.8, and likewise for 3.1.9-11 is to make it possible to distinguish sheets of paper from the same mould and to establish their relative chronologies by the movement of the filigrane in the mould.

    NOTE: in the Greek Manuscripts Watermark Archive, the position of the watermark is indicated by bracketing the compartments spanned by the watermark (cf. 3.1.12, below), with no attempt to provide this level of detail.

    Since the top and bottom of the watermark are not always intuitively obvious, and thus likewise, the left and right, there will be occasional cases of ambiguity in how this measure is to be determined, and thus instances of inconsistency from archive to archive. In any particular description, however, related data (the alignment of the watermark image and orientation relative to the mould or felt sides of the paper) should resolve the ambiguity.

    3.1.9. POSR (Position from chain to right of watermark
    The smallest horizontal distance between the watermark and the nearest chain line on the right, in mm. The rationale for 3.1.8-11 is to make it possible to distinguish sheets of paper from the same mould and to establish their relative chronologies by the movement of the filigrane in the mould.

    See additional comments in 3.1.8, above.

    3.1.10. POSB (Position of watermark from bottom of sheet
    The smallest vertical distance between the watermark and the bottom of the sheet, in mm. The rationale for 3.1.8-11 is to make it possible to distinguish sheets of paper from the same mould and to establish their relative chronologies by the movement of the filigrane in the mould.

    If this method is adopted, then this field should only be filled in when full, uncut sheets are being described, since in all other cases it is a hypothesis based on analysis or estimation (see 3.1.13).

    In our model, the measurement is always based on the size of the piece of paper in the paper-bearing source, which eliminates the need for analysis (calculation) at the input stage. The only place in our model where full sheet dimensions are given is in IPH 3.0.10-11, where the dimensions entered are also clearly defined as measured, calculated, or estimated.

    Since this data will be useless if the piece of paper is anything other than a full, uncut sheet, in our model this data is to be recorded only for that case.

    See additional comments in 3.1.8 above.

    3.1.11. POSH (Position of watermark from top of sheet
    The smallest vertical distance between the watermark and the top of the sheet, in mm. The rationale for 3.1.8-11 is to make it possible to distinguish sheets of paper from the same mould and to establish their relative chronologies by the movement of the filigrane in the mould.

    This field should only be filled in when full, uncut sheets are being described, since in all other cases it is a hypothesis based on analysis or estimation (see 3.1.13).

    In our model, the measurement is always based on the size of the piece of paper in the paper-bearing source, which eliminates the need for analysis (calculation) at the input stage. The only place in our model where full sheet dimensions are given is in IPH 3.0.10-11, where the dimensions entered are also clearly defined as measured, calculated, or estimated.

    Since this data will be useless if the piece of paper is anything other than a full, uncut sheet, in our model this data is to be recorded only for full uncut sheets.

    See additional comments in 3.1.8 above.

    3.1.12. Compartment number
    The number of the compartment (counted from left to right) which is touched by the furthest left point of the watermark.

    The term, "compartment," refers to the space between two chain lines, and this data field is based on I.P.H. 3.2.4, the list of compartment widths across the middle of the sheet.

    This part of the IPH Standard suffers from some of the same ambiguities described in the preceding sections. It is not clear which is the right side in watermarks with no obvious top and bottom. Also, the IPH Standard does not indicate whether it has in mind a full, untrimmed sheet, or only the (possibly cut-down) piece of paper occurring, for example, as a folio in a manuscript. If, as in 3.1.10-11, it means the number of compartments in the full, uncut sheet of paper , then the same qualification applies here as in those fields: this field should only be filled in when full, uncut sheets are being described, since in all other cases it is a hypothesis based on analysis and judgment (see 3.1.13).

    If, however, the IPH standard intentionally omits the word, "sheet," because it intends the end use of the paper which is usually trimmed or otherwise cut, this qualification would not apply. Here, as in other places, the wording of the IPH Standard in its present form is ambiguous.

    In our model, measurement is always based on the size of the piece of paper in the paper-bearing source, which eliminates the need for analysis (calculation) at the input stage. The only place where full sheet dimensions are given is in IPH 3.0.10-11, where the dimensions entered are also clearly defined as measured, calculated, or estimated.

    In the Greek Manuscripts Watermark Archive, the position of the watermark is indicated simply by bracketing the compartments in the compartment list (cf. I.P.H. 3.2.4) that are spanned by the watermark. This method eliminates the need to count compartments, or to calculate the number of compartments spanned by the watermark by comparing its width with the width of the adjacent compartments. It also facilitates the process of inputting data; the IPH Standard separates these data so that the inputter has to study the compartments twice for these two pieces of information.

    3.1.13. Additional information on the watermark
    The submitter may enter in this field any additional information regarding the watermark(s) that he or she considers useful. IPH examples:
    • fragment
    • can only be registered incompletely due to the use of the paper
    • registration difficult because stuck to a backing
    • assumptions as to the position of the watermark in the original sheet
    • identification of top in ambiguous watermark designs.

    3.2. Mould Data
    The IPH standard seems to conceive of this section of the data base as describing paper attributes that are a product of the mould (as opposed to the pulp, dying processes, etc.).But if that is the intent of the IPH Standard, this conception is not carried out consistently. Why, for example, are the dimensions of the full sheet (3.0.10-11) recorded in section 3.0 (Data Concerning the Sheet) instead of here as Mould Parameters? Why are the various measures for locations of watermarks, likewise, recorded in section 3.1 instead of here?

    The inclusion at the end (3.2.7, 3.2.8) of information pertaining strictly to moulds further compromises this conception. These last two data fields are for the identification of extant moulds (and 3.2.8 has a compound function which is inappropriate in entity-relationship database design; see our comments, below, 3.2.8). It is certainly appropriate to envisage descriptions of moulds by means of data describing sheets formed on the mould (cf. IPH wording to describe IPH field 3.2.8). In the case of the mould, of course, we would always have in mind the characteristics of the full, uncut sheet. However, papers and moulds each need additional descriptors which do not apply to the other. For papers, these additional descriptors include pulp attributes and results of dying or glazing; for moulds, they might include information about the state of conservation of the wires, evidence of use or ownership by particular paper mills or papermakers, and "special characteristics" of the mould (IPH 3.2.6).

    Our model for a WWW-based archive treats moulds and papers as independent sets of data, allowing for both overlapping and distinctive data fields, and structures the data base accordingly. (That is why, in our index of IPH database fields entitled IPH Categories: Comparison with WWW Database Fields, six of these IPH fields (IPH 3.2.1 - 3.2.6) are listed as occurring in two places: the Paper file and the Mould file.) See our Diagram of a Relational Database for a WWW Archive of Paper Types and Watermarks Thus, "mould parameters" are part of a paper description, and part of a mould description.

    3.2.1. Fabrication method
    The IPH standard allows 2 options:
    • Hand made
    • Machine made

    3.2.2. Paper type
    The IPH standard allows 2 options:
    • Laid (vergé)
    • Wove (velin)

    3.2.3. Laid lines in the sheet
    The IPH Standard gives the following explanation:
    Number of laid lines over a distance of 20 mm, measured along the chain line to the left of watermark or, for papers without watermark, along one of the center chain lines, first at the bottom of the sheet, second in the middle, third at the top (to be measured even on trimmed sheets)

    3.2.4. Compartmentation of the Sheet
    The IPH Standard gives the following explanation:
    Width of the chain compartments in mm in order from left to right, measured in the middle of the sheet; every number divided by a hyphen (to be measured even on trimmed sheets).
    Example: 15-20-22-20-22-23-20-14

    NOTE: in the Greek Manuscripts Watermark Archive, the location of the watermark on this grid is indicated by bracketing the compartments spanned either wholly or in part by the watermark.

    3.2.5. Shadow zones
    The IPH standard gives the following options:
    • zones below the chain lines (BS)
    • irregular zones (IS)
    • zones in the middle of the compartment (MS)
    • without shadow zones (NS)

    3.2.6. Wire
    Special characteristics of the mould, to be described in full by the submitter in his or her own words. May include (IPH examples):
    • faults in the wire
    • special traces

    3.2.7. Mould Access number
    The wording of the IPH Standard's explanation for this data field quoted below might suggest that the Standard envisages here only cases where both the mould and a paper produced from it are known and extant:

    Access number of the mould (in a separate mould database) on which the sheet was formed

    As indicated above (3.2), the WWW-based archive conceives of moulds and papers as separate entities. This design in fact carries out the intent of the IPH Standard when it proposes a separate mould database. In the WWW Standard, however, moulds may be preserved and described in this database even when it is not yet known what papers were produced from them, and vice versa. This design allows for the possibility that those papers or moulds might be discovered in the course of time.

    In keeping with sound database design, the elements of identification of moulds (as museum objects) are separated out into distinct fields (city, institution, collection, access no.), as elsewhere in our design. On our model data presentation page for Moulds we refer to this part of the IPH Standard as even though the Standard itself did not call for separated fields.

    3.2.8. Pair
    The IPH standard Description reads as follows:
    Information on the pair of moulds mentioning the access number(s) of object <and> data describing sheets formed on the second mould of the same pair
    This is another case of inclusion of multiple kinds of data in a single field. What is needed for handling this situation are separate complete records for each mould, which may be linked to their related papers and to their matched moulds (just as we need separate complete and interlinked records for each piece of paper in a pair of matched papers). Within each record, this data field (3.2.8) will function as the identification of the matched (other) mould. Each such identification of a mated mould will be in the form of a link or links to the independant record(s) containing the full description of that other mould.

    3.3. Bibliographical / Codicological Data
    Implementation of this section of the IPH standard is especially problematic because of the diversity of categories and types of information called for in it, the lack of analogy among them with respect to how they relate to paper, and the fact that some of these categories must relate to more than one of the others (one to many relationships). The IPH Standard defines ten data fields (3.3.1 - 3.3.10):
    1. authors
    2. artists
    3. writers
    4. titles of works of art or literature
    5. Country of Use
    6. Place (city) of use
    7. Publisher or editor
    8. Printer
    9. Earliest date of use
    10. Latest date of use

    The Problem

    How are we to use these fields when there are several persons, places and/or dates associated with the paper in different ways? For example, how will the user know with which person (author, artist, writer) the date or place of use are to be associated? How are we to indicate that the earliest date has to do with one figure, and the latest with another? What if two persons associated with the original use of the paper worked in different places?

    The complexity of the relationships among these disparate kinds of data requires more than ten simple data fields, as the IPH Standard seems to recognize in its repeated references, "If necessary, with reference to a special database." But this section (3.3) of the IPH standard does not specify what ought to be in these "special databases," or how they should function in relation to a registry of papers and watermarks. The clause, "if necessary," implies that the authors of the Standard themselves did not attempt to work out for the provisional publication of the Standard a clear concept of how these fields would function in a relational data base.

    The IPH heading for this section, "Bibliographical / Codicological," suggests that at least two kinds of information are included here:

    Does this distinction help us to reorganize these data in such a way that the relations among them are clear and unambiguous?

    Both of these groups of persons are recorded here as persons who put inks or colors to paper, but the distinction is that the creators of intellectual content (literary or poetic works or works of graphic art) may have put pen or brush to different paper than that now bearing their work, and reproducers are not necessarily also the creative originators of the works of literature, music or art or of the documents preserved on paper. But is this a useful or valid distinction for our purposes (for the registration of papers and watermarks)? How are we to draw the line between artists and illuminators who reproduce traditional designs or icons? Or between authors and scribes who authored substantive notes in manuscripts or may have played a role in the wording of documents, collaborating with the political or religious authorities whose authority stands behind the content of the document? Is it really necessary for persons describing paper to make this distinction (bibliographical and codicological)? How can we be sure that all persons submitting descriptions of papers would make these distinctions in the same way we would?

    This distinction suggests that we need a general file for all persons -- a prosopographical file -- which can be related to either the paper bearing object or to the paper itself.

    Another problem is that this bifurcation is not yet sufficient for our purposes. Where in this two-part scheme would we put the binder or restorer of a manuscript or printed book? Should we not also make room for owners or collectors (who may have commissioned the paper-bearing source) or even auctioneers and antiquities dealers (who may have commissioned otherwise anonymous restoration work?

    This observation suggests that we need a data field in which to specify the "role" of each person who actually utilized the paper. This field can not be in the person file, since the same person might have played different roles with different pieces of paper.

    Finally, IPH fields 5-6 (country & city of use) and 9-10 (early and late dates of use), as conceived in the provisional edition of the Standard, can only refer to the original use of the paper. But it doesn't work to make them part of the paper description file, because there can be different places and dates of use of the paper by different persons: from papermaker to scribe to illuminator to restorer to owner etc. Which one would the date refer to? Nor could they be part of the person file, because these persons may have worked with or manufactured different papers at different times and places, and this file must consequently relate to many different papers and paper bearing objects. These questions make it clear that the date and place of use of paper really pertain to the relation between these persons and papers. That is, they are not entities related to the paper in the same way as the persons of the other 6 IPH fields are. Rather, these data are relational, characterizing the use of the paper by these persons.

    These observations suggest that, as in the case of the "role" of persons who utilized the paper, so likewise place and date when they played these roles must vary for the same persons for their use of different pieces of paper.

    The Solution

    Our starting observation in our attempt to interpret and implement this section of the IPH Standard is that, for practical purposes, this section (IPH 3.3) constitutes a prosopographical index of persons who are somehow associated with the paper, either directly with the paper or with the source in which the paper is used. The names associated with a paper are important not only for dating or establishing provenance for the use of the paper, but also because their associations with particular papers provide medievalists (for example) with the data needed to date and place otherwise unknown scribes and centers of book production, binderies and centers of book restoration, artists' ateliers, and the like, or to reconstruct the work of any and all of those persons or workshops. Or, to offer a contemporary example, the association of a particular paper with a known forger of art works provides curators and police with information critical to their tasks of making decisions about authenticity, advising potential purchasers, and tracing the work of the forger in question. Names of artists and authors may also be relevant when particular centers of book production, for example, are known for publishing works of a particular composer or author.

    In our interpretation of the IPH Standard, therefore, we have created a prosopographical file called "Person" which holds basic, standard data about associated persons and can be accessed from a variety of other files. From source records (the records for identification of paper bearing objects, kept in the "Source" file), pursuing the "Artistic Content" link to descriptions of artictic/intellectual content generates a list of names of artists that are keyed to that source, each name a link to an "artistic/intellectual content" record for that particular artist. In that "Artistic Content" record, the artist's creative work preserved in the source (book, manuscript, document, etc.) is defined (his relation to that source) and his name is a link to the relevant prosopographical record. Likewise from the paper description records in the "Paper" File (the file for description of particular pieces of paper), pursuing the link to "Physical Content" generates a list of names of persons who worked directly with or upon the paper, names keyed to the piece of paper, and each name in the list is a link to a "Physical Content" record for that particular person. In that Physical Content record, the person's relation to the paper is defined (use of the paper, role and time and place of activity), and the name is again linked to the corresponding prosopographical record in the "Person" file.

    NOTE: Maintenance of these records in separate files means that there can be multiple records for a single person and a single piece of paper, thus providing for multiple types of content for a paper, but done by a single person.

    What becomes of the IPH data fields?

    The WWW Person File

    Adopting a minimalist interpretation of the IPH Standard for the prosopographical fields in IPH 3.3, we have constructed records for persons that contain only 9 fields plus a key by which it the records are linked to other files and searched:

    This is a minimalist interpretation in the sense that it does not call upon paper historians, manuscript catalogers, art historians or others who might use this database to go out and research each person whose name occurs in association with the paper. It does not, for example, ask for a breakdown of where the person lived at different times, identification of relatives and associates, or other prosopographical information to flesh out, in effect, a biographical dictionary article for that person. (Is this what the IPH Standard envisages for the "special databases" to which it refers?

    Relational Records

    Other pieces of information -- date and place of use of paper and basis for those judgments -- are included in the relational files to which they are relevant, that is, as indicated above, the "Artistic Content" and "Physical Content" files. The fields for "basis" will inform users of the rationale by which the submitter of the paper description drew conclusions about date and place of use. In the simplest case, the rationale will be

    1. that the writer, author or artist, printer or publisher wrote or printed the information in the paper-bearing source in question
    2. that it is clear that this information applies to the particular piece of paper (page, folio, endleaf, wrapper) in that source which is being described.
    More complex rationales are easily imagined. E.g., the submitter may be in a position to establish connections between related manuscripts, cite paleographical evidence, cite studies of the work of a particular printer or publisher to support conclusions about date and place of paper use, etc.

    Brief comments on each of the data fields in IPH section 3.3 follow, even though the IPH organization of data is not useful for the purposes of the WWW database.

    3.3.1-3 Author, Artist, Writer
    The purpose of these fields is simply to serve as an index, whereby researchers can search to find out what papers may have been associated with a particular creator of intellectual content in the source (the paper bearing object a document, codex, drawing, etc.). When the author is associated with a particular piece of paper -- as distinguished from a source (document) -- we are dealing with manuscripts that we would properly call "autographs."

    As elsewhere in the IPH Standard, these fields need to be broken down to accommodate multiple pieces of information. At least three fields are needed for a name:

    1. Last name
    2. First name
    3. Middle name

    3.3.9-10 Earliest & Latest Dates of paper use
    This and the following section are apparently located here in the IPH Standard because the life dates of authors, artists, composers, scribes, printers, publishers, and others who used the paper can contribute to the date of use of the paper. In fact, however, if this is truly a prosopographical file, the life dates or floruits of the persons in question should be recorded here, not dates of use (which in the IPH design are relational data relative to the paper). These persons may not have used the paper in question at all.. Moreover, the date of use of the paper can also be determined by dates of documents written on the paper, by analysis of restoration work which may result in dating of bindings and restorations which can then be applied to the paper, etc. Finally, the traditional method of comparison with other known examples of the same type of paper may contribute to a determination of the earliest and latest dates of use of the paper.

    Three conclusions from these observations:

    1. it is illogical to include earliest and latest use of the paper if we interpret this file as, essentially, a prosopographical file for recording information about persons -- creators whose works are preserved in paper bearing objects and reproducers who used paper directly.
    2. date of use of the paper ought, rather, to be designed as a data field relating the person file to the paper file.
    3. earliest and latest use of the paper ought to be thought of as a matter of analysis and judgment, and thus subject to debate and reassessment. This means that
      • provison must be made for the submitter to indicate the basis for any judgment about the date of use of the paper
      • provision must be made for recording subsequent judgments on the same piece of paper and for linking that new information to the original data for that paper.

    The WWW-based Archive provides for all of the above by taking advantage of principles of entity-relationship database design -- by treating IPH Standard section 3.3 as a prosopographical file, by creating relational files, and providing for rationale for data resulting from analysis:

    • date of use of the paper is a field in the relational file, "Physical" (pertaining to the relationship between paper and the person who uses it)
    • an added field is provided for entering the basis for date of use of paper
    • prosopographical files include fields for life dates
    • the model accomodates ongoing commentary and revision of scholarly judgments on dates and other matters of analysis. Any person who maintains an archive in the system can enter a differing judgment for the same piece of paper, and it will be found along with the original record by any researcher whose search turns up that particular record. Individual archives, likewise, will from time to time issue revisions and new editions of their databases

    3.4. Papermill Data

    The IPH Standard offers four data fields:

    The IPH Standard makes this comment:

    These data (in section 3.4.1-3.4.3) can be replaced by a reference to the relevant data set of a papermill sub-database.

    In a WWW-based relational database, data on papermills, as well as other organizations, are located in the file, Organization, which is linked to the Paper description by a link called "Papermill" (see our Diagram of a Relational Database for a WWW Archive of Paper Types and Watermarks). Among other records in this "Organization" file are represented binderies, libraries, monasteries, museums, papermills, scriptoria. These organizations are accessed from various records by appropriate links.

    3.5. Papermaker Data
    The IPH Standard offers three data fields
    1. Papermaker Surname
    2. Papermaker First name(s)
    3. Papermaker rank in the producing mill

    3.5.1-2 Papermaker Surname, First Name
    The WWW-based archive adds a field for middle initial(s), consistent with other identifications of persons in the archive, and reduces the content of the first name field (3.5.2) to a single name.

    3.5.3 Papermaker Rank in Producing Mill
    The IPH Standard offers four options for this field:
    • Fellow
    • Master/Foreman
    • Owner
    • Tenant

    4. Information about the Original Watermark (Sheet) and its Reproduction
    As the title indicates, this section is a composit of information about two different subjects. The IPH Standard offers the following, general theoretical comment on the purpose of the data called for in this section:
    As additional information about the object data described, it is very important to know if the data have been derived from the original or from a reproduction. Also, existing reproductions of an original have to be mentioned, indicating their type and their access numbers. Since a wrong interpretation may arise if a watermark is enlarged or reduced, even when the scale is given, the reproduction should be of original size.
    The various fields in IPH section 4.0 are distributed among the appropriate files that make up the WWW database design, namely, the Paper Description file (for information about the original sheet and watermark), and the Facsimile file (for information about reproductions). There is no corresponding composite division in the WWW database design.

    4.1. Original Sheet/watermark
    In the WWW database design, designation of the source from which the descriptive data are derived (from the original or from a secondary source) occur in both the paper description ("Paper") file and the watermark description ("Watermark") File.

    4.2-9. Type of Watermark Facsimile
    The eight fields for watermark types listed in the IPH Standard can be reduced to a single field in which the appropriate type of facsimile will be entered by selection from a pull-down menu on the data input form.

    In the WWW database design, this information is recorded in the Facsimile file.

    4.10. Facsimile ID Number

    In a world wide system of archives, facsimiles from many different institutions may be listed in the results of any search. Nor can it be presumed that the facsimile (e.g., the original Dylux print or beta radiograph print) is owned by the institution mounting the WWW Watermark Archive which publishes it in this electronic medium. Thus, a simple Facsimile ID number (e.g., an accession number), as called for by the IPH Standard, is not sufficient to identify the facsimile. Up to three pieces of information are required to identify the original facsimile: the Institution/owner, the name of the facsimile collection, and the access number. Thus, in keeping with the often cited principle that only one type of information may be recorded in any single field, three fields are required in each record. In our model, these records are located in the Facsimile File of our WWW Watermark Archive model, and the three fields are labelled:
    (4.10.1) "Organization"
    (4.10.2) "Facsim_Collection"
    (4.10.3) "Access_Number"
    The Organization is derived from a fuller set of data in the Organization file, to which it serves as a link, namely, country, city, and owner's name (i.e., name of the institution or private owner possessing the facsimile).
    NOTE: The collections and libraries of private individuals are treated as "institutions" in this database and thus entered in the "Organization" file. Thomas L. Gravell, before donating his wonderful collection of Dylux watermark prints to the University of Delaware (Thomas L. Gravell Watermark Archive) , would have been "institutionalized" as the Thomas L. Gravell Collection in Wilmington, Delaware.

    The Watermark Archive © was created by Robert W. Allison
    Dept. of Philosophy & Religion, Bates College

    Technical support and functionality by James Hart
    Information Services, Bates College Lewiston, Maine, 04240

    Last updated: March 3, 1998