Shouldn’t the notBefore= and notAfter= attributes on ＜education＞ in the example in its tagdoc be instead from= and to=?
I don't see why this is wrong. It would depend on the evidence, which isn't presented in the example. If you know from external evidence that he definitely wasn't at school before 1986 and he had definitely left by the end of June 1990, @notBefore and @notAfter would be the right attributes, surely? On the other hand, if you happen to have access to the school records and you can see exactly when he started and finished, you'd use @from and @to.
I agree with Martin. (This happens so rarely that I feel the need to record the fact)
I disagree. It happens all the time.
While I think your point, Martin, that it is possible to want notBefore= and notAfter= on ＜education＞ is quite reasonable, I think on this example we want from= and to=.
The content reads “Attended Cherwell School”. Cherwell school, if I read the website correctly, is what I would call a high school, covering 5 years of schooling from 7th to 11th year. The dates are 4 years. So attending Cherwell for 4 years is quite reasonable. If the intent was “attended for 1 year sometime between” then I think the content would reflect that.
The case requiring from= and to= on ＜education＞ is far more common, and thus what we should be exemplifying.
I find this argument unpersuasive. It's not in dispute that someone might attend a school for four years; the issue is that (a) don't know WHICH four years (b) we're not sure if the person actually did manage the full four years.
I don't know on what evidence you assert that "the case requiring @from and @to on ＜education＞ is far more common" but feel free to add another example.
The point isn’t to add a good example, but to get this bad example out. (Or fix it.)
My point is that ordinary people (i.e., those who have not internalized the difference between notBefore= & notAfter= and from= & to=) reading this example will interpret it to mean “at Cherwell for 4 years”, not that the person was at Cherwell for one or more minutes sometime in this 4-year span.
As for evidence for the frequency of cases, I’m not sure what evidence would convince you. I don’t have anything other than toy personographies that use ＜education＞, not even in all the TAPAS test data. That said, taking a quick scan through some existing personographies I find that for the vast majority of cases we either
a) know, or can find out, exactly when this person went to which school, or
b) have no idea, and don’t even have a way to find out, when this person went to school (other than “when she was alive, probably in her youth”).
For cases of (a) we would use from= and to=; for cases of (b) we would most likely not use ＜education＞ at all.
I imagine that there are cases where we would have evidence that a person was at a particular school sometime between X and Y. But if that were the case, we would have more than just “attended Z” as the content. (E.g., “was at Z during W’s senior year, 1896” or whatever.)
This is absolutely not a bad example. It's common to change schools in the middle of a term, or to finish on an unspecified date. For instance, I know that my last day of school was before my 17th birthday, and probably some time in July 1976, but there's no way of knowing on exactly which day I walked down the school drive for the last time and set fire to my school tie on the street outside. I would absolutely have to say @notAfter="1976-07-31". Similarly, I know I started some time in September 1970, but who could know which day? @notBefore="1970-09-01". I maintain that this is the norm, not the exception. And I absolutely would use <education> for this.
Working with early modern texts as I do a lot these days, I find myself using @notBefore and @notAfter all the time; dates are typically imprecise, and the best you can do is establish the best-knowable boundaries in this way.
Syd to add a @from @to example.
This was fixed 2014-08-06 at revision r12960