Print

Print


No, such constraints ("if attribute x has value a, then attribute y must 
have value b') are not expressable in either DTDs or W3C XML Schema 
version 1.0 (the version of W3C XML Schema used by the current version 
of the EAD Schema). W3C XML Schema 1.1, RelaxNG, and Schematron, 
however, can express these "co-occurrence constraints". Personally, I'd 
like to see the next version of EAD tighten things up and address the 
"if "@level="otherlevel", then @otherlevel is required' and similar 
situations. If this is not already a feature request, I'll submit it.

/Terry



On Thu, 10 Feb 
2011, Ethan Gruber wrote:

> Correct me if I'm wrong, but doesn't the schema require a value set for @otherlevel if @level is set to 'otherlevel'?
> 
> Ethan
> 
> On Thu, Feb 10, 2011 at 3:43 PM, Michael Rush <[log in to unmask]> wrote:
>       Nathan,
>
>       For the record, there is no "right" way of encoding EAD - you should
>       do what works best in the context of your local situation, consortium,
>       etc. �That ambiguity, as you observe, is a result of EAD's
>       flexibility, which was an intentional design decision intended to
>       encourage adoption. �It's premature to say exactly what will change
>       with the upcoming EAD revision, but it is safe to say that
>       constraining some of that flexibility will be a goal of the Technical
>       Subcommittee for EAD (I'm a co-chair of TS-EAD).
>
>       There are two issues to unravel in your reply. �First is numbered
>       components. �The only meaning that you should associate with a
>       numbered component is what its valid parent and child elements are.
>       There is no difference between <c02> and <c03> other than their
>       parent-child relationship. �So the level attribute is how you indicate
>       what the archival nature of a given component is - series, subseries,
>       file, etc.
>
>       Second is the use of level="otherlevel" without the otherlevel
>       attribute. �Encoding <c level="otherlevel"> doesn't tell you much that
>       is useful, so if you aren't going to specify otherlevel="something",
>       I'd just omit it. �I'm not of the opinion that you need to supply
>       @level for every component - here at Yale we only require @level for
>       series, subseries, subgroups, and otherlevel="accessions". �Our best
>       practices allow file and item, but they are optional. �But if your
>       local practices require @level and you have a level that can't be
>       properly assigned to one of the closed list of values, then I think it
>       best if you include a value for @otherlevel.
>
>       Mike
>
>       Accessioning Archivist / EAD Coordinator
>       Beinecke Rare Book & Manuscript Library
> 
> 
> On Thu, Feb 10, 2011 at 2:43 PM, Nathan Tallman <[log in to unmask]> wrote:
> > Mike,
> > I hear you. �But, I don't think I explained my self very well.
> > The way I was taught, you don't specify the otherlevel attribute when using
> > level="otherlevel", it just left at that. �No mater where it is in the
> > hierarchy.
> >
> > For example:
> > <co1 level="series>....
> > <co2 level="subseries>....
> > <c03 level="otherlevel">...
> > or
> > <co1 level="series>....
> > <co2 level="subseries>....
> > <c03 level="otherlevel">...
> > <c04 level="otherlevel">....
> > In other words, the placement within the�hierarchy�is indicated by the
> > numbered component, not the otherlevel attribute. �So, no extra encoding
> > work. �I know that not everyone uses numbered components, so it might not
> > work for all. �I'm not trying to be argumentative, I just want to make sure
> > I am encoding things correctly!
> > Gmail is telling me that Barbara has an update, let me see what she says...
> > �Ah, I see she uses the otherlevel attribute. �Perhaps I should start using
> > it... �I agree with Barbara that both are probably acceptable, as long as
> > you know the affect on your stylesheets. �EAD was designed to be flexible
> > after all.
> > So here's my�follow up�question: �If I'm using level="otherlevel", should I
> > also be using the otherlevel attribute and not relying on numbered
> > components?
> > Thanks!
> > Nathan
> >
> > On Thu, Feb 10, 2011 at 1:32 PM, Michael Rush <[log in to unmask]>
> > wrote:
> >>
> >> Nathan,
> >>
> >> That level of specificity in the level attribute adds too much
> >> complexity to the markup. �What is a sub-subseries if it isn't a <c
> >> level="subseries"> within a <c level="subseries">? �By encoding <c
> >> level="otherlevel" otherlevel="sub-subseries"> you use two attributes
> >> where you could use one and don't gain any semantic meaning. �What if
> >> you have a third subseries? �Do you encode <c level="otherlevel"
> >> otherlevel="sub-sub-subseries">? �Where does it end?
> >>
> >> Encoding <c level="otherlevel" otherlevel="sub-subseries"> creates two
> >> thing - more encoding work and the need for a more complicated
> >> stylesheet - neither of which I like. �Since I can't see what you gain
> >> by it, I argue for sticking with @level="subseries".
> >>
> >> This is not to say there aren't times to use @otherlevel - we often
> >> use @otherlevel="accession" - I just don't think there's any reason to
> >> specify sub-sub-sub-sub-subseries when you can easily identify them as
> >> such given the encoding context.
> >>
> >> Mike
> >>
> >> On Thu, Feb 10, 2011 at 12:57 PM, Nathan Tallman <[log in to unmask]>
> >> wrote:
> >> > Mike,
> >> > When I was trained on EAD, I was instructed to use level="otherlevel"
> >> > for
> >> > sub-subseries. �May I ask why you don't�recommend�this approach?
> >> > Thanks!
> >> > Nathan
> >> >
> >> > On Thu, Feb 10, 2011 at 11:09 AM, Michael Rush <[log in to unmask]>
> >> > wrote:
> >> >>
> >> >> Christine,
> >> >>
> >> >> I don't recommend that approach. �I when I have sub-subseries, I just
> >> >> encode them with level="subseries". �Keep it simple, as always. �If
> >> >> for some reason you need to identify the actual sub-subseries - in a
> >> >> stylesheet for example - it's a simple xpath: *[@level='subseries' and
> >> >> parent::*[@level='subseries']].
> >> >>
> >> >> Mike
> >> >>
> >> >> On Thu, Feb 10, 2011 at 10:53 AM, De Catanzaro, Christine D
> >> >> <[log in to unmask]> wrote:
> >> >> > Hi everyone,
> >> >> >
> >> >> > This is hopefully a rather straightforward question for the list: For
> >> >> > a
> >> >> > collection with sub-subseries, when entering the level, would the
> >> >> > correct
> >> >> > attribute for level be "otherlevel" and then to specify the level
> >> >> > would it
> >> >> > be "sub-subseries" (with the hyphen)?
> >> >> >
> >> >> > Thanks,
> >> >> > Christine
> >> >> >
> >> >> > --
> >> >> > Christine D. de Catanzaro, Ph.D., M.L.I.S.
> >> >> > Access Archivist
> >> >> > Subject Librarian - Music
> >> >> >
> >> >> > Georgia Tech Archives
> >> >> > Library and Information Center
> >> >> > 704 Cherry Street
> >> >> > Atlanta, GA 30332-0900
> >> >> >
> >> >> > Phone: 404-385-0107
> >> >> > Fax: 404-894-9421
> >> >> >
> >> >> > E-mail: [log in to unmask]
> >> >> >
> >> >
> >> >
> >
> >
> 
> 
> 
>

Terry Catapano
Special Collections Analyst/Librarian
Columbia University Libraries Digital Program
212-854-9942
[log in to unmask]