[meteorite-list] MADML RFC (was "Data in auctions")
From: Jamie Stephens <j.stephens_at_meteoritecentral.com>
Date: Thu Apr 22 10:28:37 2004 Message-ID: <3FA1E81A.1050504_at_morphism.com> For posterity, here's the first (and almost certainly the last) Request For Comments for the Meteorite Auction Data Markup Language (MADML) version 1.0 ** If you're a big seller and you think you'll never provide this ** ** structured data, please let me know. I don't want to waste my ** ** time or others'. No hard feelings (from me) at all. ** ** If you want to flame me, please be creative. Privacy-invader, ** ** user-agreement-abuser, never-seen-the-light-of-day geek, and ** ** irresponsible loser are old. I'm a white male. Got to be ** ** something there. ** Please send suggestions and comments to me directly. (Apparently the List can't take much more Jamie-inspired traffic. I'm sympathetic to the List in this regard.) I'll revise and republish if I have the guts. The goal of MADML is to provide a *simple* way to provide pertinent auction data in a standard format that's easily parsed. With MADML, it'll be easy to track fine-grained stats over time: $/g ranges for Howardites $/g ranges for NWA 998 $/g ranges for unclassified NWA Number of bidders on Gibeon slices Number of winners of Eucrites Etc. (Note that $/g will be ranges, not single numbers.) All could be easily graphed over time. N.B. No EBay ID's! Since EBay auction titles must be short and since sellers will likely want to maintain complete control over those titles, MADML goes into the item description, not the title. Approach 1: An HTML table: <!-- MADML 1.0 --> <table border=1> <tr><td align=right>Name:</td><td>Dar al Gani 607</td></tr> <tr><td align=right>Mass:</td><td>2.75g</td></tr> <tr><td align=right>Cut:</td><td>Partslice</td></tr> <tr><td align=right>Classification:</td><td>EUC</td></tr> <tr><td align=right>Shock stage:</td><td>1</td></tr> <tr><td align=right>Weathering:</td><td>1</td></tr> </table> All bidders could easily see the data. Sellers could wrap whatever HTML they wanted around the table (including *hiding* the entire table in an HTML comment). Approach 2: An HTML comment: <!-- MADML v="1.0" name="Dal al Gani 609" classification="EUC" shock="1" weathering="1" mass="2.75g" cut="partslice" --> The data would be invisible to bidders. (You can see raw HTML by using your browser's "View Source" function (available in my browser in several places: e.g., from the menu you get by right-clicking on a blank spot on a page).) The rationale for this HTML comment approach is to avoid messing with any visual aspects of an item description. Approach 3: A text line: MADML [ name="Dar al Gani" mass="2.75g" ... ] A simple, consise, inelegant visible presentation. Not pretty. Approach 4: Approach 1 wrapped in an HTML comment The benefit is easy switching from visible to invisible data. Just remove or add the HTML comment syntax: <!-- Nothing will show up, but the data is still there. <!-- MADML 1.0 --> <table border=1> <tr><td align=right>Name:</td><td>Dar al Gani 607</td></tr> ... <tr><td align=right>Weathering:</td><td>1</td></tr> </table> --> Which approach do you prefer? No matter what format, we (!) want several fields: name: Required. Either 1) the unabbreviated name from the Met Bulletin or 2) unabbreviated, unofficial, pre-Bulletin name. Examples: "Northwest Africa 1694" "Bensour" mass: Required. The mass of the specimen offered in grams. Examples: "2.75g" "2345g" "0.003g" cut: Required. One of "Individual" or "Partslice" or "Fullslice" or "Fragement" or "Endcut" or "Specks" or "Speck" or "Dust" or "Vapor". classification: Optional (suggested). The standard (?) abbreviated classification or "unclassified". See 'http://www.geocities.com/CapeCanaveral/9278/systemat.htm'. For achondrites, use the first three letters. All caps. Include petrologic type. Examples: "H5" "L3.8" "EUC" "HOW" "CV3" "H3-6" "H/L3" "CV3R" "IAB" "H IMB" "C-UNGROUPED" "MES" "PAL" shock: Optional (suggested). 1-6 weathering: Optional (suggested). 0-6 Comments: Readers can use online sources to look up classification data from names, so it's not essential to replicate that data here. However, it'd be much easier to have it here. Other standard data (e.g., fall/find, year, location, etc.) is probably best left to external sources. Feel free to provide it: find="yes" year="1917" location="Los Angeles, California, USA" Comments, please. Flames expected (from people with lives, presumably and ironically). BTW, I'm looking into getting a life. First I'm going to read about it on the Internet. --Jamie Received on Thu 30 Oct 2003 11:42:02 PM PST |
StumbleUpon del.icio.us Yahoo MyWeb |