Wikistamp

This is an initial proposal for "Wikistamp", a wiki-style online catalog of en:postage stamps.

Stamp cataloging

The stamp catalog is the traditional method of listing the types of postage stamps in a useful way. Catalogs began as literal dealers' catalogs, but have since evolved into important repositories of stamp data used heavily by all serious collectors. With some 300,000 type of stamps now in existence, catalogs have become rather large affairs: the en:Scott catalog is six large volumes in small print, with a new set each year.

Efforts to create a general online catalog have not been particularly successful so far [check the "Stanley-Gibbons" site - stanleygibbons.com/shop - and click "subscriptions" for a new attempt that's not free]. There was a "World Postal Issues Database" (WPID), and a couple of other efforts. User:Stan Shebs has an unpublished database "DWCS" that he's been developing for several years, and in the process of recording 145,000+ types, have developed some techniques to manage the mass of data more effectively. (More about that in a moment.) The makers of commercial catalogs derive considerable revenue from their print editions, and they've only recently even starting offering their material in electronic form.

What do collectors use catalogs for? Probably the main usage is identification; I have a stamp in my hand, I see it's an old Austra-Hungary, but it's one that has several minor variations in perforation and design, so I go to the catalog to make a positive ID. Another use is to build "want lists" of stamps still missing from one's collection. Often this entails scribbling in the catalog. Still another is at a dealer to check the dealer's price against the catalog's price, which is usually based on averages from surveys. Some collectors use those prices to judge the overall value of their own collections.

Stamp data

Data about stamps comes in several forms. The first kind of data is purely numeric, or has a limited number of possible values. Examples are year of first issue, face value, color(s), and so forth. Philatelists have developed terminology for describing all these fairly precisely (usually).

The second kind is textual, such as the name of a person being depicted, thematic elements (perhaps a sailboat or palm tree visible in the background - some collectors really like that sort of thing :-) ), the specific purpose of the stamp ("to pay overseas registered mail rate", etc), and miscellaneous notes ("only sold in Bavaria, but valid throughout Germany").

A third kind is graphical; in addition to basic images of the whole stamp, it's also useful to have closeups of specific features, postal markings, and usages on cover.

DWCS

DWCS ("Description-based World Catalog of Stamps", pronounced "dawks") is yet another attempt to do a philatelic database, but it has some useful properties. First, it is not based on the numbering of stamps; while this is traditional practice for catalogs, numbering is an editorial decision that necessarily collapses multiple varieties into one. DWCS creates records for all distinct varieties described. Second, the set of data fields and conventions is rich enough to cover stamps of all periods, such as complex multiple overprints, color shades, and inflation denominations. Third, the support software has deep knowledge of field interrelationships; for instance, an overprinted stamp must be on exactly one type already issued (or marked as "unissued"). Fourth, it does not require complete data in order to function.

The input syntax is designed for compactness and speed of data entry, with default values put on lines ending with ";", and the ability to use all the usual abbreviations and shortcuts. This excerpt shows the first few stamps of Thailand:

country=Thailand area=Asia;
money=128sol=64a=32sio=16sik=8f=4sa=1t;

8/4/1883 King_Chulalongkorn;
1sol bl o
1a
1sio
1sik
1sa
d=;

7/1/1885;
ovpt=1_TICAL 1t on=1sol
ovpt=1_Tical 1t on=1sol
ovpt=1_Tical_13.5mm 1t on=1sol

4/1/1887 issue=1887 King_Chulalongkorn_1887;
1a grn 2/0/1891 o
2a grn&car
3a grn&bl
...

The software can then crunch this into a standard database form, a palmtop download, etc, or output in the same form (which should match character-for-character if the input data is correct).

Internally, there are actually two kinds of records, one for stamp types and the other for individuals in a collection (created by the "o"s in the excerpt).

Although this has been a personal project of Stan Shebs for several years, he has designed and added to it in such a way that he is the sole owner of the copyright, and is interested in contributing both software and data to Wikimedia under GFDL or other appropriate license. This would put the online stamp catalog over the halfway mark to complete coverage.

Wikistamp vs Wikipedia

There are at least two ways to create a wiki for postage stamps.

One way is to make a dedicated wiki, a la wikibooks or wikitravel. This has the advantage of being able to be unabashedly technical and abbreviated. A downside is that cross-references to Wikipedia, where the concepts are explained, have to be interwiki references.

Another way is to build the stamp data inside Wikipedia. This doesn't necessarily mean 300,000 new articles; there's not much to say about each modern stamp. If one were to have per-year per-country articles like "1984 stamps of Denmark", complete coverage would entail articles of 500-2,000 words, and there would be about 20,000 of them. 300,000 images would be a lot, but a composite image of a "year set" (all of a country's stamps for a given year) is still useful, and fewer are needed.

A useful approach to the nontextual data might be to store it in the Wikicommons, and fold into per-language articles in some tricky fashion.

Another further-out idea is to add some sort of ability for collectors to maintain their personal collection data in conjunction with the Wikimedia, so for instance each collector sees checkboxes by each type, and can check off stamps in his/her own collection.

Related projects

Wikibooks features a World Stamp Catalogue, but it hardly has any content yet.