Looking at Julee’s example, even if you replace the identifier with one that starts with a letter the correspDesc doesn’t resolve correctly. The error it’s throwing is that the element is simply not allowed, and that the allowed elements are “creation,” “langUsage,” and “textClass.” And swapping out correspDesc with any of those three allows it to resolve just fine. Everything else mentioned under “may contain” in the TEI guidelines gives the same error regardless of whether or not it includes an attribute. Is it possible OxyGen is working from an older version of the guidelines? I note that the page states it was last changed on July 23rd.
Similarly, sourceDesc seems to only want to allow bibl, list, listBibl, p, or table. Anything else (biblStruct, biblFull, ab, msDesc, listEvent, listNym, listOrg, listPerson, listPlace, listRelation, recordingStmt, scriptStmt, listApp, and listWit) gives the same expected element error.
> On Oct 28, 2018, at 7:26 PM, Syd Bauman <[log in to unmask]> wrote:
> Julee Tanner --
> Relatively fast-and-lousy response, as I've got a lot of work to do
> tonight. :-)
> On the document fragment you posted:
> 1) <msDesc> doesn't go inside <bibl>, it *is* the bibl. I.e., put
> the <msDesc> directly inside <sourceDesc>.
> 2) The value of @xml:id on <correspDesc> must start with a
> letter, e.g. "t7660-7662". BTW, in this case I think this is a
> good thing, as it avoids the concern that some software will come
> along and think that the identifier is "-2".
> 3) A single <TEI> document has one and only one <teiHeader>. If
> your second <text> deserves its own complete set of metadata, it
> should be in a different <TEI> element with its own <teiHeader>.
> That <TEI> element could be in a different file, or could be
> combined with other <TEI> elements in a single file encoded as a
> If there are only some bits of metadata that are different, but
> lots of them are the same, you might do well to use some other
> mechanism to associate bits of metadata with sections of
> On the overall process:
> Personally, I think it is quite difficult to learn text encoding from
> the _TEI Guidelines_. They are more of a reference manual than a
> tutorial. I strongly recommend taking an introductory TEI
> workshop. Yes, I might be considered biased because I teach such
> workshops, but I teach them because I think they are very helpful.
> On asking questions:
> Generally much easier for those of us trying to provide answers if
> you send (a pointer to) the entire TEI file, and tell us to which
> schema you would like it to conform. (And send along (a pointer to)
> the schema, too, if it is not one of the standard ones. :-)
> Hope this is helpful!
>  This is a rule about identifiers in XML; if we (TEI) want to use
> XML identifiers, we are simply stuck with this constraint. Most
> of us feel it is much better to use XML identifiers and suffer
> some minor constraints than make up our own element-identifi-
> cation system.
>  And many of us would say it should have one and only one <text>,
>  See the tagdoc for <teiCorpus>
> as a starting point, but you would probably have to read sections
> of chapters 4 and 15, too.
>  See section 15.3
> for the main TEI mechanism, if interested.
>  Conveniently, you are quite close to Victoria, where just such a
> workshop is taught every summer at DHSI. But there are several
> other workshops or workshop series. The Oxford Summer School
> probably has one; DH@Guelph probably has one; and my own
> institution, Northeastern University, puts on a TEI workshop
> here every year or two, and also does them at other
> institutions. We did one at the University of Washington back in
> June of 2008. We do not have any scheduled for the immediate
> future, but updates would be posted on
>> I am working on a project to transcribe and encode a collection of
>> the correspondence of Armand Jean du Plessis, Cardinal Richelieu,
>> much of it previously unpublished. I am beginning with a series of
>> manuscript letters that I photographed this summer at the Arsenal
>> location of the Bibliothèque nationale.
>> I am teaching myself TEI, using largely the P5 guidelines, Appendix C, and
>> the TEI-Correspondence-SIG/correspDesc repository on GitHub. I have begun
>> to encode metadata and content for the first letter, but have a few errors
>> in my hierarchy.
>> I am hoping someone there might be able to check my work and tell me why
>> Oxygen will not validate <msDesc>, <idno>, <correspDescr>, <date>; and, as
>> I open a new section for encoding my second letter, the second <teiHeader>.
>> My coding is reproduced (without indents, non-validating elements bolded)
>> below this message.
>> Thank you in advance for any assistance you might provide!
>> Julee Tanner
>> MLIS candidate
>> University of Washington Information School
>> <repository>Bibliothèque nationale de France-Arsenal</repository>
>> *<idno type="BnF">Ms-2657</idno>*
>> *<correspDesc xml:id="7660-7662">*
>> <correspAction type="sent">
>> *<date when="1462-06-29"></date>*
>> <correspAction type="received">
>> <persName>de Noyers</persName>
>> <p>Sy Mr de Bouillon est pris il est question de faire voir . . .</p>
>> . . .
> Syd Bauman, NRP
> Senior XML Programmer/Analyst
> Northeastern University Women Writers Project
> [log in to unmask] or
> [log in to unmask]