Talk:Descendants of Charlemagne (couples)

Here's the other bulldozer - the one whose driver knows what he can see and is very pleased in principle but knows it looks wrong and may not know whether it should come right. Boots on, chaps!

Numbers
I counted 59 marriages. Twice. Columns 1 and 2 cannot therefore contain more than 118 people (and actually contain fewer because of people listed for two qualifying marriages). The clever counting mechanism says 132, so we are not listing everyone who is counted. We should explain why to the readers, some of whom may repeat the check while it's still humanly possible. We don't want an apparent error to send someone away for ever.

I presume that the first figure counts each individual found by our database crawler to have been at any stage married to any other descendant (not counting anyone reported to have married himself). The introductory text would be clearer if it said something closer to that. It should also say clearly why the table will not always list everyone who is counted (and will often - as now - be seriously deficient). Makes no sense even to me - if the system can determine that someone qualifies enough to be counted, why can it not list the person?

I see that there's a similar table about killers. I expect it and others in the series  will need similar explanation, on the face of each page that contains such discrepancies, not hidden away on an explanatory page about SMW..

— Robin Patterson (Talk) 13:59, 14 June 2009 (UTC)

Arrangement
As the tables are alphabetical for column 1, we are presumably not suffering from the "omit if no birth year" problem. There must be other reasons for the number discrepancies, and they may be related to a flaw in the logic behind each heading, but I suspect that they are just the inability of this smw application to crawl everybody instantly after it has crawled the spouse.

Because of what the headings say, readers would expect to see every qualifying individual listed at least once in column 1, in each section corresponding to a marriage that was with another descendant. (And from what the second column subheadings say, I'm sure that that is the intention.) That would give almost twice the number of marriage rows unless there were a great proportion of descendants who married two or more descendants. Clearly that is not the current display, even if it is the intended ultimate display. Hugh the Great appears at least twice in column 1 but his third wife is not in that column at all. Readers who see discrepancies may wonder whether they have misunderstood the headings. I did at first.

Consider an example group: "Third spouse". Not all the people who will eventually fill column 2 were on their third marriage. ("No, Hadwig, I would love to marry you, but as we are both descended from a to-be-randomly-selected ancient king we would be upsetting some of our putative descendants a few centuries hence because I have been married twice but you have never been married.") Thus as Hugh the Great in column 1 took Hadwig, the person in column 2, as his third spouse but it was her first marriage, she should have an entry in column 1 of the "First spouse" section, with him as her "column 2" partner. She hasn't. Will she, when the system reaches her? If she will, the headings will eventually make sense and the numbers will add up until any new article has been created about a qualifying person (which will happen with increasing frequency, we hope). Until the system reaches Hadwig, any reader who cares about her enough to look will see that the tables are wrong or at least incomplete.

Readers could be forgiven for wondering about the current wording of the headings. We don't want to lose such readers when we could explain better. Clarification, please. Something at the top of the page can probably cover this problem and the "numbers" problem.

— Robin Patterson (Talk) 13:59, 14 June 2009 (UTC)