Forums: Index > Watercooler > Charlemagne's descendants

Excellent project. Has anyone else done it apart from Fred's contact (now suffering ill health)?

Best discussed here rather than spread around user talk pages, I suggest, because a growing number of us will be interested.

I see there's a need for renumbering because someone has discovered a shorter line. Solution to that would have been (and minimising future renumberings would be) to get all of his children sorted, then all of his grandchildren, and so on, and anyone coming off a longer line would be obvious because they would already have the smaller gen number.

Remember the short line from Geo III to Geo VI. I think the error-prone tends to use the royal line instead.

Robin Patterson (Talk) 15:57, 9 April 2009 (UTC)

(Somebody who read this didn't even think to ask about the short line from Geo III to Geo VI.) — Robin Patterson (Talk) 15:03, 27 July 2009 (UTC)

It is indeed best to work from the early generations onwards, but the current ones are what attracts attention! I'm afraid we'll need a bot to occasionally renumber generations as more lines appear.

Note that there are two templates.

  1. {{OrderCharlemagne | N}} puts people in [[Category:Descendants of Charlemagne (Generation N)]]. We could have {{OrderCM | N}} which is shorter but less clear.
  2. {{SovCM | CYear}} puts people in [[Category:Ruling descendants of Charlemagne]] where C is the country code and year is the year of accession. Rtol 18:03, 9 April 2009 (UTC)

Generation 2[]

CM's kids now all have pages and info pages -- the same cannot be said of Wikipedia. This also includes a count of the grandchildren: 26, 19 of whom have a page already. Rtol 10:14, 12 April 2009 (UTC)

Well done, gentlemen! Some Wikipedia administrator should ensure that WP has a link to FP. — Robin Patterson (Talk) 11:02, 12 April 2009 (UTC)
Link is there already: {{Familypedia|Charlemagne (747-814)}} Rtol 11:27, 12 April 2009 (UTC)
Good work. I've not checked whether all his kids that are on WP have the template. I did study the template and give it a /doc page, and I'll be writing about that elsewhere on this wiki if I don't fall asleep on the job. — Robin Patterson (Talk) 15:58, 15 April 2009 (UTC)
The kids on WP are now linked back to here. Rtol 17:46, 15 April 2009 (UTC)

Disputed descendants[]

Among the 26 grandchildren, 2 are disputed. They are singled out by the * in [[Category:Descendants of Charlemagne (Generation 3)|*]]. Rtol 10:14, 12 April 2009 (UTC)

I note that for the "Descendants of Charlemagne" categories to make any sense, none of the descendants of the disputed people should be categorized as descendants, otherwise we'll need categories like Category:Descendants of Charlemagne (2 disputed links) etc. I notice that the Wikipedia page for Clémence of Aquitaine (1060-1142) implies that it is only a guess that she is the mother of Gerard II van Gelre (c1098-c1131), so there goes a whole branch into the "disputed" basket. Thurstan 22:05, 12 April 2009 (UTC)

That may be better indeed. And I would think that we would just put the first disputed into a category "Descendants of Charlemagne (disputed)" and not trace their children etc through the generations. Rtol 22:29, 12 April 2009 (UTC)
We surely need a bot. Removing the disputed line from Petronilla (825-?), William I, King of England (1027-1087) shifts a generation and he is the main anchor point for all later generations. Rtol 07:11, 13 April 2009 (UTC)
My opinion: Just categorize undisputed lines. Also note that William the Conqueror has multiple lines of descent from William the Conqueror. Also, it would be much easier to remove the categories manually than with a bot. -AMK152(talkcontribs) 02:16, 14 April 2009 (UTC)

Generation 3[]

Now complete, including a count of the 4th generation: Charlemagne had 45 known great-grandchildren, 4 of whom have a page on Wikia Genealogy. What happened is that people focussed on the lines that lead to England and Holland, and forgot about France, Germany, and Italy. So, putting up generation 4 will restore the geographical balance somewhat. Rtol 07:57, 15 April 2009 (UTC)

Ruling descendants[]

I changed the categorization and the templates so that we allow for personal unions. Rtol 18:09, 22 April 2009 (UTC)

Generation 4[]

Now complete, including a count of generation 5: 65 (incl. the first double descendant), 15 of which are on Wikia Genealogy. rtol 08:18, 24 April 2009 (UTC)

Chapeau ! Fred Bergman 21:05, 3 May 2009 (UTC)

Info pages[]

I added OrderCharlemage to the info pages and to the info categies template. rtol 21:01, 3 May 2009 (UTC)


There is now a semantic test (thanks, Phlox) that automatically categorizes descendants. rtol 10:09, 17 May 2009 (UTC)

{{Descent Charlemagne test}} — Robin Patterson (Talk) 12:31, 17 May 2009 (UTC)
It uses two parser function calls per generation, which means over 80 for recent people - restricting other things we can put on the page. Now that we have a couple more of the earliest generations finalized, can we cut a couple more early tests off? {{OrderCMTest3}} and {{OrderCMTest4}}? — Robin Patterson (Talk) 23:20, 23 May 2009 (UTC). See — Robin Patterson (Talk) 02:50, 24 May 2009 (UTC)
But there are still bugs, among others the sisters Katja and Nina Bleijenberg are wrong displayed. Fred Bergman 10:12, 17 May 2009 (UTC)
Katja and Nina were manually categorized. Errors are introduced that way. To automatically categorize them, the page needs to be re-opened and saved (I already did that for the girls). So it will be while before automation is completed, and then every time a link is added we start anew. rtol 10:24, 17 May 2009 (UTC)

To give most people their lowest possible number as soon as possible, we can concentrate on family lines from the youngest people in each generation. Have we an easy way to see who they are? — Robin Patterson (Talk) 23:20, 23 May 2009 (UTC)

Let's first finish the semantic descendants' lists. Then we can do "Descendant of Anybody (Generation N)" without expensive autocats. This would also remove the need for many of the current categories, including the "Ruling descendants ...". rtol 05:26, 24 May 2009 (UTC)

Generation 5 complete[]

including a first count of generation 6 (71, 28 of whom are already there)

however, with auto-categorization (and re-categorization) it is not that important to do this generation by generation and I'll pursue family-lines for a while rtol 20:49, 19 May 2009 (UTC)

SMW is here too[]

SMW allows us to show the descendants of anyone without having to create categories. The autocategorization of Charlemagne's descendants has caused a general slowdown of Familypedia.

The plan is therefore as follows:

  1. Remove the "Descendant of Charlemagne" categories;
  2. Create a new category VIA; and
  3. Create a template to show at any page a list of Very Important Ancestors.

This plan messes up the bookkeeping of the project to put all Charlemagne's descendants on Familypedia. I therefore created a page Charlemagne (747-814)/project. rtol 05:30, 29 May 2009 (UTC)

Rate of propagation[]

Descendants of Charlemagne are now marked by the signature {{Project Charlemagne}} and placed in Category:Descendants of Charlemagne. This autocats. The rate of propagation is good. I started with Charlemagne and Amaudru on June 5. Autocat has found some 160 descendants in five days, and the rate increases with the number of descendants found. rtol 07:44, 10 June 2009 (UTC)

Generation numbers[]

Generation numbers are now part of {{Project Charlemagne}}. This is done with a new Property:Order of Charlemagne that will also allow easier queries per generation. Propagation requires a double flush from old to young. rtol 18:30, 19 June 2009 (UTC)


While {{set order of Charlemagne}} can be done manually, it should not. Humans introduce error, and as information percolates so do the errors. Restoring the damage would be a lot of work. rtol 06:39, 27 July 2009 (UTC)

Visible categories or easy equivalent[]

I know that this has been raised before, but I was not satisfied with what I understood of the answer. If I want to see which generation one of my relatives is in, I don't want to have to poke around in a list of facts or whatever they are that involve a lot of scrolling and may be different tomorrow. I want to see it in the customary place (a category) or something like the "project" logo with a generation specified. On Adolphus, Duke of Cambridge (1774-1850), who has been edited several times since I created his page, there is no logo, a fact that tells me that that system isn't reliable yet. He should have one, because he is a son of George III of the UK, and his descendant George VI of the United Kingdom should have a number based on it, not on George V's number. Maybe I will be sure to see the generation number in the signature panel (or get used to looking for it in a list of facts) when that system settles down and we have all become accustomed to it, but it's not in "alpha" mode yet.

Is there any reason why those of us who feel inclined should not add by hand a numbered category to the page of any known descendant?

Robin Patterson (Talk) 15:03, 27 July 2009 (UTC)

As I have now argued many times, manual categorisation is error prone and too much work. There are hundreds of pages already, and there will be tens of thousands.
Adolphus went wrong because the page did not call {{t|Descent Charlemagne test)), which is because Adolphus is an experimental SMW page. As the SMW pages will be overhauled again, there is no point in fixing them now. rtol 20:27, 27 July 2009 (UTC)
Would he have been better with an ordinary info page? — Robin Patterson (Talk) 13:55, 29 July 2009 (UTC)
The issue with SMW pages is now solved.

The going ons with the Windsors over the last week clearly illustrates the advantages of automation of generation numbers. Last week, Prince William was generation 41. We add the Tecks, and everyone from George VI shifts a generation. We connect the Duke of Edinburgh to the House of Hesse, and Prince Charles shifts two generations. We connect the House of Hannover to the House of Anhalt, and everyone from Prince Frederick Lewis shifts two generations. Manual adjustment would have been a drag.

And there is no reason to believe that Prince William really is generation 38. There is only a handful of lines between generations 6 and 30, and there may well be shorter lines through the House of Saxe or the House of Hesse-Kassel. Other studies are unreliable guides, because they did things manually as well. rtol 00:18, 31 July 2009 (UTC)


Is there an easy explanation for the occasional appearance of the logo without any generation number? — Robin Patterson (Talk) 13:55, 29 July 2009 (UTC)

You need a double flush as generation number needs to be determined and saved before it can be displayed.

When there is a generation number, I would like it to be a link to the relevant page. Can that be managed easily? — Robin Patterson (Talk) 13:55, 29 July 2009 (UTC)

Can be done. Requires string manipulation. You're welcome to try. rtol 14:07, 29 July 2009 (UTC)

Generation 6[]

Complete. rtol 11:05, September 5, 2009 (UTC)

April 2010 update[]

Now that the SMW is settling down nicely, I would like to be able to publicise this project in places such as the soc.gen.medieval mailing list.

But it needs better documentation. Even I don't understand who does the double flushing that has been talked about. I presume that a bot is involved, but I've no idea which bot (perhaps because at least a couple of them don't follow the simple procedures requested) or when it might operate.

The Project Charlemagne page recently proudly stated that generation 8 had been completed and that that all of its children had Familypedia pages. Great! But that same section had a direct link to the /count subpage, which has a table saying that that level and the two immediately preceding it are incomplete. Visitors would get a poor impression if they noticed that; and we want them to read most of the documentation if they are to become useful contributors.

There is still an unanswered question about who says a particular level is complete. The second part of Rule 2 - "(preferably including several "source" links each)" - is more honoured in the breach than the observance. Where is the documentation showing that two or more authorities agree that Person Xyz had only seven children and noting that we have pages for all seven? As an example, the first listed page for Generation 8, Baldwin IV, Count of Flanders (980-1036), had six double flushes last year, has not been edited in the last 12 weeks, but still has no "citations" or external links from which a diligent enquirer might do a spot check to see whether there was any evidence that Familypedia has listed all of Baldwin's children.

Respectful but puzzled --— Robin Patterson (Talk) 14:21, April 4, 2010 (UTC)

Generations 1-5 are complete in the sense that all known descendants have a page.
Generations 6-8 are complete in the sense that all descendants that could be found have a page. However, we also know that there are a bunch of 5th generations descendants who left no trace, but lived long enough to have children.
Generation 9 is complete in the sense that all descendants that could be found have a page, but their children do not. rtol 14:37, April 4, 2010 (UTC)
double flushes are for inbreeding, not for descent
I cross-check data with the sources that I have access to. If someone else wants to cross-check with other data, they're welcome. Given that we are talking about thousands of pages, that person may want to introduce a category:cross-checked. rtol 14:46, April 4, 2010 (UTC)

Semi-automatic procedure[]

In July 2009 rtol said (above) "You need a double flush as generation number needs to be determined and saved before it can be displayed". Now he says you don't need double flushes for descent. Something has changed for the better and I don't need to know what.

As I understand from explanations, from examining the code, and from observing what happens in practice:

Whenever any user saves a sensor page, whether or not it already has an Order of Charlemagne in its "facts" list, if it is properly linked to the page of a parent whose sensor page has an Order, the page being saved will thereafter have its own Order, either a new one or an update, based on the current Order of the parent (the one with the lower number if the parents have two different numbers).

So there is a bit of work to be done by users. Those users could include a bot that routinely saved a null edit of every sensor page in Familypedia every now and then.

A bot could be dedicated to this particular project and instructed to go through Descendants of Charlemagne (Generation 10) doing all of those sensor pages then gen 11 then gen 12 and so on. That should get every descendant with the right Order (as far as we know) "in short order".

Robin Patterson (Talk) 13:17, April 18, 2010 (UTC)

Double flush was needed for the transition from info- to sensor-pages. Not needed for newly created sensor pages. rtol 13:23, April 18, 2010 (UTC)

Citation and documentation[]

If one's evidence comes from the Web, it takes only a few seconds to paste each source URL into the "Sources" box on the person form. If it's from a book, it takes only a few seconds to type the book's shortcut (e.g. {{BPB}}) and page number. — Robin Patterson (Talk) 13:17, April 18, 2010 (UTC)

By 2011 we still have imperfections[]

Too few listings[]

As I've just commented on the talk page for one of the generations, something has recently gone seriously wrong with the machinery that should pick up and list everyone of a particular generation. Only one person in generation 10, only a couple in each of the next few generations (including one that has one person listed once and one listed three times). Then gen 17 has nobody, which inspired me to change the default message on an earlier generation. Robin Patterson 01:32, February 16, 2011 (UTC)

Bot use[]

I've not noticed any response to my suggestion (a few paragraphs above this) that we have "a bot that routinely saved a null edit of every sensor page in Familypedia every now and then". Nobody has admitted to operating such a creature, but we still get told that such things happen. What's going on? Robin Patterson 01:32, February 16, 2011 (UTC)

Two and a half years later, I still have not noticed any sort of reply to my plea. George Neville, 1st Baron Latymer of Snape (c1414-1469) has had order 21 for some time, possibly months, but several of his direct documented-on-FP descendants have been on FP for months if not years and have no order numbers. Example: William Bentinck, 2nd Duke of Portland (1709-1762)/sensor was created 4 years ago and has no order of Charlemagne. If the bot is still running, why is it not spreading the numbers down? If it is not, would somebody please give the rest of us enough information to get it running again? Is there any relevance in the fact that Special:Contributions/Rtolbot shows no action in the last 3 years? -- Robin Patterson (Talk) 01:29, November 1, 2013 (UTC)

Display of icon[]

Discontinued. I'd like to see it restored, though I appreciate that it should not be part of the standard page display making more server work for every article load. How about a simple template that users can add to a person's page when they know that the person is a descendant of KtG? The determination of the number would still be programmed, not manual. In rare cases, a disputed link in the chain might mean some person pages continued to show the icon without being entitled to, but that would be a very small number. Robin Patterson 01:32, February 16, 2011 (UTC)