Lifestyle
 
[Phlox]

Phlox 22,994 edits since September 6, 2007

22,994

User:Phlox/TBD

From Familypedia

Contents

[edit] Current

2009-04-29

  • Info page optimization- was that an explode thing, or wasn't it some esoteric hack that minimized expansion?
  • Are events separate subpages, or are they possible with above optimization
  • evidentiary/ conclusions data model?
  • How to handle alternate theories- separate afns for separate theory of parentage, or in the same record (GED6.0 did this I think)
  • How to reconcile GED5.5 data model with more powerful data models? Do we lead by picking a winner or wait and see. Database inertia will tend to propagate whatever model is chosen, so that argues in favor of picking a model that will gracefully degrade to accurate GED5.5 output, but can represent information in the highest quality form preferred by serious genealogy scholars. Quality and collaboration becomes our differentiators over ancestry.com et al.
  • pdf output script like wp
  • Data entry hand holding- javascript form.
  • Companion open source software to the web site? Like how familytree maker is produced by genealogy.com?
  • This whole notion of believed reality is a recipe for massive edit wars as the genealogy wikia population scales up. Either that or it is massive redundancy- where everyone has their separate realities. Do we do a me-too quantity thing like the other sites? Sure we can we pile on Gedcom after Gedcom until we have hundreds of thousands of nodes, but is that a good thing? Don't we want to go for quality? Gedcom Lineage based models focus on believed realities. Links are made with certainty, and but the reality is that it is all much murky than that. Eg. oftentimes it is not clear whether one or two people are being described in a set of documents.
    • Would an evidence model allow for greater collaboration?
    • On WP- evidence focuses the community on shared reality. That's why citations are so crucial to get out of POV viscous cycles. Certainly, there will be differences of opinion, but this shift means the difference between- everyone with a gedcom gets a new article versus- versus one or two versions of an article (if not one common article with alternate scenarios on parentage/ other questions).
    • Typical Problem: say some contributor with deep psychological need to establish ties to Scandinavians does some good research work, but a few of his conclusions are viewed as highly improbable. The GEDCOM way is to totally fork the tree. Mitigation: AFN's can allow a separate AFN identity for the person with the alternate parentage based on this conclusion. So see what this does for duplication? It's nuts. Instead you have something like:
  1. father hypothesis 1: Ward Cleaver / evidence_A / probability_A / evidence_B / probability_B ...
  2. father hypothesis 2: Hans the Milkman / evidence_M /probability_M ...
    • The idea is that each person could override the probability numbers assigned- their home software could display the view the tree with the probabilities they assign. A Bayes algorithm could determine most likely trees. What does that imply for familypedia linking structure? How do you private user preferences on a public site?
    • Maybe confidence probability is by votes? Hmmm no, that's a formula for propagating sock puppets.
    • Immense weight of the chain of evidence? If you encode the data formally, then wouldn't the mountain of evidenciary material expressed in a shared way bring the wikimedia engine to its knees? I mean, storing the whole breadcrum trails.. for instance of Aunt Bee was recollecting at the old folks home that Andy was in El Paso in 1944, and then you have his enlistment papers and the unit he trained with used the base near El Paso, but went to another base in 1943, so it must have been not 1944 but before 1943 etc etc. If you formalize all that junk, jeez well you wind up with maybe Andy was in town on the right year and there is a 10% probability he could be the father. These discussions [1] [2] [3] touched on the issues.
  • statistics on current usage of info pages- which fields, how many edits, dead pages, how many editors

[edit] Research TBD


[edit] Periodic maintenance

[edit] Infrastructure

  • use simple forms extension with javascript for info page article generation Simple forms used with Blog. Much easier would be to make a custom php extension, but this would allow forms to be created by other contributors. ~ Phlox 05:22, 6 November 2007 (UTC)
  • try reloading template:Convert after sysop status. Maybe spam filter will allow. admin upload.
  • Internationalization: Omega ultimate wiktionary: [4]


[edit] Microformats

  • New article at [5]. Point him to our stuff- all info pages can generate his stuff, and he may profit from examining our data fields.

[edit] New ID

  • World connect has an reference scheme we can use for autolinks to their database.
    • Database:Individual ID eg:db=:1099716&id=I23425986 is Benedict (1667)
  • figure out how we want to do potentially huge lists of such source material
  • also a potential problem with alternates and long descriptions maxing the preexpand limit- I think we are going to a high use/low use split in info pages.
    • High frequency calls- small size. EG. fields Father, Mother, Birth death, children.
    • Low frequency use- huge size- descriptions, short bio, alternate names, residence lists. Sources, other language encyclopedias.

[edit] YSTR and DPL

  • Idea- instead of creating a useless category for each DYS value, use a raft of dummy templates. Example Template:DYS393~14 is for a count of 14 of the marker DYS393]]. Then you do an and query for articles that include all templates with your 12 digit code.

[edit] Microformat encode events

  • represent events in some sort of info pages
  • showinfo them as calendar Vevents
  • So we can do: [6]
  • Install Exhibit Timeline action by Rosie Morales to grab events from a page and display them this way. [7]

[edit] piddly stuff

  • vde doesn't belong on counties
  • form for county create new article- use preload on the http line and editintro for help above edit form.
    • do the same for blank info pages.
  • help on captcha -notice on confirmed email
  • info page display page sorely needs updating.
  • weird comma missing problem in dates parser function time? What about old dates?
  • Fixup Info pages
    • International (French templates) rationalized
    • Update info fields
    • Add preload so fields can be auto enterred with help.
    • Figure out microformats and communicate with that guy about them.
    • either rename Template:Info table person (interferes with dpl), or embed.

[edit] Forms

[edit] Multilingual tips

  • Postpend the lang code to the Mediawiki message
  1. Every Wikipedia that goes well has enough basic information about Wikipedia/WikiWiki/editing pages/how to contribute/HTML-codes/how to entitle pages etc. translated so that newcomers easily learn how to contribute.
  2. Every Wikipedia running well has at least one person doing the pushing, who continues to put in new entries so the RecentChanges page looks alive! Once it looks dead, it will stay dead.
  3. The HomePage looks up-dated, and the links in the titlebars have been translated, too.

[edit] Mediawiki

[edit] Extensions we haven't been using

Retrieved from "http://familypedia.wikia.com/wiki/User:Phlox/TBD"