Familypedia
Forums: Index > Watercooler > Future of Familypedia


Hi everyone,

About a month or so ago, I started thinking about the future of Familypedia. For a whole bunch of reasons, I like the idea of open collaborative online genealogy, and Familypedia is the best site I have found on which to do this. To make the "collaborative" part work, though, I feel that we need a lot more regular users (so that there is at least one person on the site who shares ancestors with me :-) - say, a factor of 10 or even a hundred. I then started thinking that wikia might not want to keep supporting us if we got that big.... Today I came back after a hiatus, discovered that form-based pages (and the forms themselves) were acting strange, and found the discussion at Forum:SMW_on_Familypedia - it seems we are already starting to outgrow wikia, and any dedicated server would have to be supercharged.

So the question that I hate to ask (because so much work has gone into it): Is SMW sufficiently "scalable" (i.e., able to accommodate a 10-100 fold increase (or more) in people pages) as an engine for the kind of data processing we need? If not, are there reasonable alternatives out there, either as a back end to the mediawiki-driven interface we have now, or as a separate platform? (I don't know the wiki/web 2.0 world very well, so don't have any real ideas here, unfortunately.)

I don't want to be a wet dishrag, but I also don't want to invest too much of my limited genealogical time into adding content to a site that may be running into a ceiling.....

Bruce Kendall 01:02, November 27, 2010 (UTC)


Good points, Bruce. Until Wikia tells us exactly what was going wrong and whether the imperfect patch can be fixed, we won't know the full answer. User:Phlox, who designed the SMW, did tests to satisfy himself that the project was scalable, but maybe he overlooked something. There is probably a solution of staying with Wikia and growing exponentially if we simplify the pages so that each page is not much more burden than any other wiki's pages (e.g. the million-plus articles on the Lyrics wiki); that could mean abandoning some of the clever links between generations (e.g. cutting our ancestor trees back to the same size as Wikipedia's) and/or abandoning the hope of full multilingualism. — Robin Patterson (Talk) 11:02, November 27, 2010 (UTC)
I'm just confused why phlox made every single variable into a property. Why can't you just set the parents and be done with it for basics. I made this for narutopedias family trees
{{#switch:{{lc:{{{2|}}}}}
|children={{#ask:[[Concept:Characters]][[Parent::{{{1}}}]]|?Name|format=list|headers=hide|mainlabel=-|link=none|sep=,}}
|parents={{#show:{{{1}}}|?Parent|link=none}}
|siblings={{#ask:[[Concept:Characters]][[Name::!{{{1}}}]][[Parent::{{#arraymap:{{#show:{{{1}}}|?Parent|default=NOPAGENAME}}|,|$|$|{{!}}{{!}}|,}}]]|link=none|sep=,}}
}}

It can be used such as {{FamilyTree|{{FamilyTree|{{PAGENAME}}|Parents}}|Siblings}} to get grandaunts and granduncles {{FamilyTree|{{FamilyTree|{{PAGENAME}}|Children}}|Children}} would print out grandchildren.

It should be easy to change the OR logic in siblings to AND logic so that half siblings and step children can be tested. Maybe I'm missing the big picture/issue... SimAnt 20:50, November 27, 2010 (UTC)

Well, the first basic thing we want to do (inside {{showfacts children}}) is to query each of the children for birth and death dates and places, because that is standard info you want to see on the parents page. We also have an old "standard" (not always followed now) of storing the children on the father's page, so {{showfacts children}} on the mother's page queries the father's page for the children. Thurstan 21:15, November 27, 2010 (UTC)

Just noticed something verrrry important, the category system is a total mess for smw. I think that it needs to be my first priority to start fixing things. SMW has Special:BrowseData, which is suppose to replace a lot(over 95% probably) of these categories, and each page should only have a single category which are then sortable by properties on the pages. Considering the number of categories, and pages with categories manually added someone is going to have to run a bot to trim it down. SimAnt 22:48, November 27, 2010 (UTC)

This seems to me to be a bit of a "chicken and egg" problem: I see you boggling about the number of SMW properties that we have already. But to replace categories like Category:British MPs 1707-1708 and Category:Australian military personnel killed in World War I we need a lot more properties that we haven't defined yet (they were put in the "to do" basket). It seems to me that we want to get the basic stuff working again, as it was before, and then see what we need to add. Thurstan 01:16, November 28, 2010 (UTC)

Okay that's fine, as long as duplicate results don't bother you for now in the special:Browsedata. SimAnt 18:38, November 28, 2010 (UTC)

On the subject of making every variable a property: could you have queries such as Semantic MediaWiki/demo query-subquery without doing that? — Robin Patterson (Talk) 02:58, November 28, 2010 (UTC)

That query, yes. I would think of that not many queries use all of those properties. If you needed a property for a query you would just have to add it to the template, but there is no real good reason to add it by default. SimAnt 18:38, November 28, 2010 (UTC)