Dear Elisa and Piotr,

Regarding ‘ʿmq’, it is not a coma but the Ugaritic letter ʿ, “ayin”. So that is why it is permitted—am I right Michael Sperberg-McQueen?

As for @ana, of course, my mistake. I had copied/pasted the same line in my Dita… I had checked several times, but you know how it is when you check a text yourself… So thank you very much! I’ll do the modification today. 

Regarding @type vs @pos, I doubt it will be possible to change in my chapter; however, I’ll change in the guidelines, XML (and XSL). Thank you very much for your kind advice.

Last, thank you for the compliment Piotr, I hope my doctoral supervisors will do the same… but above all, my goal was to do the guidelines as useful as possible for those who want to use taxonomies in order to prepare to a hermeneutics of action.
Anyway, these guidelines (always) need improvements, and I think the first will be to add ODD—I have decided to add my customized ODD, not only for the custom @attribute. Thank you Elisa for having convinced me. It makes me realized that a customized ODD is very useful (and interesting too). 
Just wondering why a a comprehensive ODD course is not always given before TEI encoding course. I truly believe we have to bridge this lack of ODD training—at least in France, I don’t know for other countries.


Le 5 mai 2018 à 05:47, Elisa Beshero-Bondar <[log in to unmask]> a écrit :

Hi Piotr— I hate to open Pandora’s box here, but can you make a concise summary of Michael’s explanation about the string “ʿmq” and why it’s okay? You’ve piqued my curiosity here! 
Thanks for the thoughtful review—I quite agree about modifying the values of @ana to make them more consistent (and about the vagueness of anyURI in practice if not in principle). 


On Fri, May 4, 2018 at 11:39 PM, Piotr Bański <[log in to unmask]> wrote:
PS. I hasten to correct one of the points I've made, about what I misinterpreted as an inverted comma and the nature of NCNames. And that with many thanks to Michael Sperberg-McQueen, who has kindly tested the string "ʿmq" and explained to me why and how I was wrong suggesting that it's not legal. Briefly: I was wrong.

So that point is fine, but you might want to re-scan your @ana attributes. A plain "subjectiveVar" will qualify as anyURI (because almost everything, sadly, will), but I suspect that your intentions could have been different there, and you probably don't want to mix labels and references in a single attribute.


On 05/05/18 03:28, Piotr Bański wrote:
Dear Vanessa,

 > As for @type vs @pos, I’m a little worried

Well, your hands are clean -- processing the ticket that produced @pos and friends took many months, and processing the additional documentation took some extra time as well. It looks like the page you're pointing at will be recreated after another official release only. It's not that @type is illegal -- it's simply been released to perform other duties (or fancies), while a specialized attribute took over, for those who need it. If you can smuggle in a footnote into your publication, your readers will have a chance to get updated, and if not, well, we live in a dynamic world, where some things fortunately move forward, even if that takes time... No one is going to blame you, I'm sure.

Your documentation looks very impressive! Notice that in some of your @ana, you use pointers at fragments (essentially, at xml:ids):

ana="#Character #ANT #v-ANT-ktu1-3_ii_l5b-6a"

but in others, you use plain labels:


That may eventually bite, if you expect the same behaviour of both.

Also, I am not checking this against the spec right now, but I would be cautious in defining IDs such as "ʿmq" (note this little inverted comma at the beginning) -- this is most probably illegal from the XML point of view (others will correct me if I'm wrong on this point, but I'm afraid I'm not).

Still, these are minor friendly nitpicks of the sort that I am sure you'd expect of this list, but, overall, I'm impressed by the documentation and by the scale of your endeavour -- chapeau bas!

Best regards,


On 05/04/18 22:26, Vanessa Bigot Juloux wrote:
Dear Piotr,

Thank you very much for your useful answer. I’m currently looking to ODD based on your suggestion and with the kind help of Elisa.

As for @type vs @pos, I’m a little worried since I just wrote a chapter in a forthcoming volume (Brill) that I’m co-editing ( I have used @type following the example on the TEI guidelines:

Regarding @ana, it’s not for grammatical features, but to set up a hermeneutics of action based on taxonomies, see my guidelines:
- taxonomies:
- transliteration:
In fact, this is the main topic of my chapter.

Of course, your comments/remarks are most welcome.

One again, thank you.

Le 4 mai 2018 à 16:35, Piotr Bański <[log in to unmask] <mailto:[log in to unmask]>> a écrit :

Dear Vanessa,

> Later today, I should be able to send my argumentation about adding a new attribute needed for cuneiform script.

Thanks, but -- do take your time. You can just add that attribute to your ODD and work with it right away, there is no special approval process necessary for you to be able to customize your own TEI schema. (On the other hand, we would love to learn about the results, when it's convenient to you.)

A piece of relevant documentation is located here:

What you want to see in your ODD looks a bit like the following:

<elementSpec ident="w" module="analysis" mode="change">
<attDef ident="myNewAttribute" mode="add">
<desc>This is a test attribute to hold a normalized value for my words.</desc>
  <dataRef key="teidata.text"/>

(No need to bother about the namespace, at this point -- you can play with that later. The datatype is also very permissive.)

The above would allow you to do:

<w pos="verb" myNewAttribute="your-text">etc.</w>

Please note the @pos (part-of-speech) attribute used instead of @type. I also don't know what you want to use the attribute @ana for -- if for further grammatical features, then there exists the @msd attribute ready and willing to do that job, see

If you need @ana for something else, then please bear in mind that its datatype is URI.



On 05/04/18 14:51, Vanessa Bigot Juloux wrote:
Dear Radu,
Dear TEI List,

thank you for having transmitted my message to TEI listServ. I have just added a comment to my post on StackExchange:

I have looked at other |element| and |attribute| but they don't work. I am currently working on an argumentation to explain why they are not relevant to the cuneiform script, and why |@correspUnic|should be preferred.

Later today, I should be able to send my argumentation about adding a new attribute needed for cuneiform script.

All the best,

Vanessa Bigot Juloux | Ph.D. candidate
» Ecole Pratique des Hautes Etudes <>  (EPHE), Paris Sciences et Lettres <> (PSL)
» Chair Membership and Outreach Sub-committee for Europe (American Schools of Oriental Research <>)
Mobile + WhatsApp: +33 (0) 6 98 97 02 02
Academia <>, <>, [log in to unmask]eu <mailto:[log in to unmask]>

Le 4 mai 2018 à 14:14, Radu Coravu <[log in to unmask] <mailto:[log in to unmask]>> a écrit :


I just spotted a question on StackExchange about adding a custom attribute to the TEI vocabulary:

I do not know how to do that but if any of you knows maybe they can answer there.


Radu Coravu
<oXygen/> XML Editor

Elisa Beshero-Bondar, PhD
Director, Center for the Digital Text | Associate Professor of English
University of Pittsburgh at Greensburg | Humanities Division
150 Finoli Drive
Greensburg, PA  15601  USA
E-mail: [log in to unmask]
Development site: