need input on ticket 1331, dating attributes
[Insert pun about the attributes required to get a ticket for a medieval match.com here.] [New folks -- e-mail from me should almost always be read with a fixed-width or monospaced font. This one is no exception.] For ticket 1331 (https://github.com/TEIC/TEI/issues/1331) I'm supposed to jigger the prose a bit, and develop a table of the semantics of the various combinations of @from, @to, @notBefore, and @notAfter. I've gotten a very good start on this, but am running into trouble with the table, and am seeking help and advice. Issues, in rough order of importance: 1. There are two combinations of these attributes that I don't know what to do with. As far as I'm concerned, it makes no sense to use the pairs @from & @notBefore or @to & @notAfter. That said, some *might* argue that using three of the four (either @from with @notBefore & @notAfter, or @to with @notBefore & @notAfter) does make sense. (Specifying a duration that starts at @from and ends sometime in the range specified by the other two, or that ends at 3A@to and starts sometime in the range specified by the other two.) So: a) What should I put for the makes-no-sense pairs? b) Should I be writing prose to support the ( (@from|@to), @notBefore, @notAfter ) encoding? Or is that too wild? Or is there some other interpretation we should be considering? 2. Since order of attribute specification is inconsequential, and an attribute may only be specified once, making a full 4x4 table of these attributes results in 4 cells that make no sense because they are for the same attribute twice, and each of the other unique 6 cells is duplicated (for a total of 16 cells). A B C D A XX ba ca da B ba XX cb db C ca cb XX dc D da db dc XX Should we display all 16 cells? Or make an abbreviated chart, e.g.: A B C D A XX -- -- -- B ba XX -- -- C ca cb XX -- D da db dc XX Or, noting that in the above chart row A and column D are useless, make a truncated chart, e.g.: A B C B ba XX -- C ca cb XX D da db dc If you look at the draft of the chapter I have at the moment (http://paramedic.wwp.neu.edu/~syd/temp/TEI_Council/html/ND.html#NDATTSda) you'll see that there several (ugly) versions of this table. Which one should we use? 3. The default HTML styling of <table>, <tr>, and <td> is not good for this purpose, and I see no easy way to special-case this. (Putting @style on <td>, and maybe <tr>, does not work; the attribute does not get propagated to the HTML.) Does anyone know of a way around this, or already know where in the Stylesheets I should be working to get @style working? If you take a look at http://paramedic.wwp.neu.edu/~syd/temp/TEI_Council/test-table-of-attrs-layou..., you'll see a set of tables closer to what I think they should look like.
Since I’ve not seen any comments so far I’ll add my thoughts inline:
1. There are two combinations of these attributes that I don't know what to do with. As far as I'm concerned, it makes no sense to use the pairs @from & @notBefore or @to & @notAfter. That said, some *might* argue that using three of the four (either @from with @notBefore & @notAfter, or @to with @notBefore & @notAfter) does make sense. (Specifying a duration that starts at @from and ends sometime in the range specified by the other two, or that ends at 3A@to and starts sometime in the range specified by the other two.) So: a) What should I put for the makes-no-sense pairs? b) Should I be writing prose to support the ( (@from|@to), @notBefore, @notAfter ) encoding? Or is that too wild? Or is there some other interpretation we should be considering? ad a) I’m not able to recall the whole discussion but I wonder if everyone was agreeing that these are nonsense? If we said they were nonsense we should add a schematron rule and make it explicit in prose. We should as well(?) restrict the combination of @when with any other attribute from att.datable.w3c ad b) Yes. Some James Cummings wrote on the ticket: "It is clear to me at least that as soon as you have @from or @to you have a duration and the other notBefore/notAfter then applies to the imprecision of the missing from/to attribute.“
2. Since order of attribute specification is inconsequential, and an attribute may only be specified once, making a full 4x4 table of these attributes results in 4 cells that make no sense because they are for the same attribute twice, and each of the other unique 6 cells is duplicated (for a total of 16 cells). A B C D A XX ba ca da B ba XX cb db C ca cb XX dc D da db dc XX Should we display all 16 cells? Or make an abbreviated chart, e.g.: A B C D A XX -- -- -- B ba XX -- -- C ca cb XX -- D da db dc XX Or, noting that in the above chart row A and column D are useless, make a truncated chart, e.g.: A B C B ba XX -- C ca cb XX D da db dc If you look at the draft of the chapter I have at the moment (http://paramedic.wwp.neu.edu/~syd/temp/TEI_Council/html/ND.html#NDATTSda) you'll see that there several (ugly) versions of this table. Which one should we use? The smaller, the better. And I’d add a note that the combination of any of these attributes with @when is nonsense …
3. The default HTML styling of <table>, <tr>, and <td> is not good for this purpose, and I see no easy way to special-case this. (Putting @style on <td>, and maybe <tr>, does not work; the attribute does not get propagated to the HTML.) Does anyone know of a way around this, or already know where in the Stylesheets I should be working to get @style working? If you take a look at http://paramedic.wwp.neu.edu/~syd/temp/TEI_Council/test-table-of-attrs-layou..., you'll see a set of tables closer to what I think they should look like.
I don’t know. Best Peter
On 15/02/16 13:45, Peter Stadler wrote:
Since I’ve not seen any comments so far I’ll add my thoughts inline:
1. There are two combinations of these attributes that I don't know what to do with. As far as I'm concerned, it makes no sense to use the pairs @from & @notBefore or @to & @notAfter. That said, some *might* argue that using three of the four (either @from with @notBefore & @notAfter, or @to with @notBefore & @notAfter) does make sense. (Specifying a duration that starts at @from and ends sometime in the range specified by the other two, or that ends at 3A@to and starts sometime in the range specified by the other two.) So: a) What should I put for the makes-no-sense pairs? b) Should I be writing prose to support the ( (@from|@to), @notBefore, @notAfter ) encoding? Or is that too wild? Or is there some other interpretation we should be considering? ad a) I’m not able to recall the whole discussion but I wonder if everyone was agreeing that these are nonsense? If we said they were nonsense we should add a schematron rule and make it explicit in prose. We should as well(?) restrict the combination of @when with any other attribute from att.datable.w3c ad b) Yes. Some James Cummings wrote on the ticket: "It is clear to me at least that as soon as you have @from or @to you have a duration and the other notBefore/notAfter then applies to the imprecision of the missing from/to attribute.“
Yes. I can't recall precisely what I was on about but I suspect I was saying that you might have something like: <date from="1932" to="1955"> which is clearly a duration and if you had: <date notBefore="1932-07" to="1955"> then you have a duration but with an uncertain start. If you have: <date from="1932" notBefore="1932-07" to="1955"> then the _only_ way I can read this so that it makes sense is that: "We know the date range is from 1932-1955 but additional evidence of some sort tells us that the from date is not before July of that year, but we aren't saying it started in July, just that we know it wasn't before then." In such a case <date from="1932-07" to="1955"> would be misleading since we can't know that it was July, just that it was after July. So really, following Syd's table, one should encode this as: <date notBefore="1932-07" to="1955"/> which makes sense to me. I'd use table 3, I don't mind the duplicated text. For @from/@notBefore and @to/@notAfter I would say something like: "Not allowed."
The smaller, the better. And I’d add a note that the combination of any of these attributes with @when is nonsense …
I think it would be good to have a short blurb of prose (just a couple more sentences) above where you have the table and a bit clearer content in the table. But yes, noting as Peter says that combination with @when is meaningless. With the examples given immediately after I'd try to have one of each of the major types of combination. Currently there is @when, @notBefore/@notAfter, @from/@to, @from/@notAfter. Although duplicating ideas I would include @notBefore/@to as well, as well as an uncertain from/to of YYYY-MM to YYYY or something. -James -- Dr James Cummings, James.Cummings@it.ox.ac.uk Academic IT Services, University of Oxford
participants (3)
-
James Cummings -
Peter Stadler -
Syd Bauman