http://blog.gmane.org/gmane.text.tei.general hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://permalink.gmane.org/gmane.text.tei.general/15512 <pre>In a project I am supervising this summer, a Northwestern student will do an edition of "Fair Em" based on the EEBO-TCP text. I hope that this will be the first of many (or quite a few) Young Scholar Editions in a project I call Shakespeare His Contemporaries. We haven't settled on all the details, but I'm strongly inclined to keep all the stuff done by the editor in a separate &lt;text&gt; element, whether it's an introduction, short glosses, or long interpretive notes. Maintaining a clean and easily seen division between 'editorial' and 'authorial' content seems to me to matter more than whether the authorial stuff and the editorial stuff are bundled in &lt;group&gt; or are separate documents (though I prefer the former). Ultimately the important question is how to articulate and maintain the unity of a single "work." Eduard Fraenkel's Agamemnon, one of the monuments of mid-twentieth century scholarship, was published in three separate volumes (text, introduction, commentary), but it is a single "work" with a title of its own (Agamemnon edited with a commentary by John Eduard Fraenkel) and a single catalogue number. To the extent that &lt;teiHeader&gt; and &lt;text&gt; relate to each other like a book and its catalogue record, it would seem to be desirable to have a single teiHeader to underscore the unity of such a work if that unity is claimed or exists, which is not always the case. From: Michelle Dalmau &lt;michelle.dalmau&lt; at &gt;GMAIL.COM&lt;mailto:michelle.dalmau&lt; at &gt;GMAIL.COM&gt;&gt; Reply-To: Michelle Dalmau &lt;michelle.dalmau&lt; at &gt;GMAIL.COM&lt;mailto:michelle.dalmau&lt; at &gt;GMAIL.COM&gt;&gt; Date: Saturday, May 18, 2013 4:22 PM To: &lt;TEI-L&lt; at &gt;LISTSERV.BROWN.EDU&lt;mailto:TEI-L&lt; at &gt;LISTSERV.BROWN.EDU&gt;&gt; Subject: Re: Introductory material for a primary source text Students provide critical introductions, biographies, and annotations for the Victorian Women Writers Project (http://www.dlib.indiana.edu/collections/vwwp/). The intros and bios are encoded in a separate TEI document and linked via the Header (see below). We did this for several reasons, not least maintaining the source text intact, but also to provide index access to these intros and bios irrespective of the source text, though they are linked from the source text and vice versa. We also have bibliographic metadata specific to these contributions and wanted to keep the Header as singular as possible. Finally, in the cases of the bios, they can be attached to multiple source texts. Sometimes those working on the source texts are not the same ones providing commentary so separating these contributions in terms of workflow is a smoother process for the editors. To illustrate, if you search for Cholmondeley, you'll see a bio, an intro, and two books by the author in the results. As you view each document, you'll see the appropriate references on the right to the related materials. This is how we cross-reference the files in the Headers (nothing earth-shattering, just wanted to share): The source text, The Danvers Jewels: &lt;notesStmt&gt; &lt;note type="relatedItem" subtype="critical_intro" resp="#inglezak"&gt; &lt;bibl&gt; &lt;title ref="http://purl.dlib.indiana.edu/iudl/vwwp/VAB7387_intro"&gt;Introduction to The Danvers Jewels&lt;/title&gt; &lt;author&gt;Inglezakis, Mara L&lt;/author&gt; &lt;pubPlace&gt;Indiana University&lt;/pubPlace&gt; &lt;date&gt;2011&lt;/date&gt; &lt;/bibl&gt; &lt;/note&gt; &lt;note type="relatedItem" subtype="bio" resp="#inglezak"&gt; &lt;bibl&gt; &lt;title ref="http://purl.dlib.indiana.edu/iudl/vwwp/cholmondeley"&gt;Mary Cholmondeley&lt;/title&gt; &lt;author&gt;Inglezakis, Mara L&lt;/author&gt; &lt;affiliation&gt;Indiana University, Bloomington&lt;/affiliation&gt; &lt;date&gt;2011&lt;/date&gt; &lt;/bibl&gt; &lt;/note&gt; &lt;/notesStmt&gt; The bio for Cholmondeley: &lt;notesStmt&gt; &lt;note type="relatedItem" subtype="source_text"&gt; &lt;listBibl&gt; &lt;bibl&gt; &lt;title level="m" ref="http://purl.dlib.indiana.edu/iudl/vwwp/VAB7387"&gt;The Danvers Jewels&lt;/title&gt; &lt;/bibl&gt; &lt;bibl&gt; &lt;title level="m" ref="http://purl.dlib.indiana.edu/iudl/vwwp/VAB7143"&gt;Red Pottage&lt;/title&gt; &lt;/bibl&gt; &lt;/listBibl&gt; &lt;/note&gt; The intro for Danvers Jewels &lt;notesStmt&gt; &lt;note type="relatedItem" subtype="source_text"&gt; &lt;listBibl&gt; &lt;bibl&gt; &lt;title level="m" ref="http://purl.dlib.indiana.edu/iudl/vwwp/VAB7387"&gt;The Danvers Jewels&lt;/title&gt; &lt;/bibl&gt; &lt;/listBibl&gt; &lt;/note&gt; &lt;/notesStmt&gt; &lt;sourceDesc&gt; &lt;bibl&gt; &lt;title type="introduction"&gt;Introduction to The Danvers Jewels&lt;/title&gt; &lt;author&gt;Inglezakis, Mara L&lt;/author&gt; &lt;affiliation&gt;Indiana University, Bloomington&lt;/affiliation&gt; &lt;date&gt;2011&lt;/date&gt; &lt;/bibl&gt; &lt;/sourceDesc&gt; Because the annotations are related to chunks of text, we are leaning toward embedding these in the source text with attributes to distinguish the type and subtype of note, resp, etc. We haven't yet integrated the scholar-contributed annotations, but we have begun modeling it in the encoding: &lt;note type="scholarly" subtype="gloss" resp="#encoderusername" xml:id="note_017"&gt; Jocosely: “Playfully, in a waggish manner” (&lt;bibl&gt;&lt;title level="m"&gt;OED&lt;/title&gt;&lt;/bibl&gt;) &lt;/note&gt; We could maintain these notes in an external file and point to them via identifiers. Not sure what wold be the best approach at this point so I look forward to your suggestions. --Michelle ----- Michelle Dalmau, Interim Head Digital Collections Services ----- Herman B Wells Library 1320 East 10th Street, Rm W501 Bloomington, Indiana 47405 ----- Web: http://michelledalmau.com Twitter: &lt; at &gt;mdalmau On May 18, 2013, at 3:26 PM, Kevin Hawkins &lt;kevin.s.hawkins&lt; at &gt;ULTRASLAVONIC.INFO&lt;mailto:kevin.s.hawkins&lt; at &gt;ULTRASLAVONIC.INFO&gt;&gt; wrote: But if you need a surer way to pull out bibliographic descriptions of just source documents or just commentaries, then you will probably want to go with separate TEI documents as Martin Holmes has suggested. </pre> Martin Mueller 2013-05-18T23:43:29 http://permalink.gmane.org/gmane.text.tei.general/15511 <pre>Students provide critical introductions, biographies, and annotations for the Victorian Women Writers Project (http://www.dlib.indiana.edu/collections/vwwp/). The intros and bios are encoded in a separate TEI document and linked via the Header (see below). We did this for several reasons, not least maintaining the source text intact, but also to provide index access to these intros and bios irrespective of the source text, though they are linked from the source text and vice versa. We also have bibliographic metadata specific to these contributions and wanted to keep the Header as singular as possible. Finally, in the cases of the bios, they can be attached to multiple source texts. Sometimes those working on the source texts are not the same ones providing commentary so separating these contributions in terms of workflow is a smoother process for the editors. To illustrate, if you search for Cholmondeley, you'll see a bio, an intro, and two books by the author in the results. As you view each document, you'll see the appropriate references on the right to the related materials. This is how we cross-reference the files in the Headers (nothing earth-shattering, just wanted to share): The source text, The Danvers Jewels: &lt;notesStmt&gt; &lt;note type="relatedItem" subtype="critical_intro" resp="#inglezak"&gt; &lt;bibl&gt; &lt;title ref="http://purl.dlib.indiana.edu/iudl/vwwp/VAB7387_intro"&gt;Introduction to The Danvers Jewels&lt;/title&gt; &lt;author&gt;Inglezakis, Mara L&lt;/author&gt; &lt;pubPlace&gt;Indiana University&lt;/pubPlace&gt; &lt;date&gt;2011&lt;/date&gt; &lt;/bibl&gt; &lt;/note&gt; &lt;note type="relatedItem" subtype="bio" resp="#inglezak"&gt; &lt;bibl&gt; &lt;title ref="http://purl.dlib.indiana.edu/iudl/vwwp/cholmondeley"&gt;Mary Cholmondeley&lt;/title&gt; &lt;author&gt;Inglezakis, Mara L&lt;/author&gt; &lt;affiliation&gt;Indiana University, Bloomington&lt;/affiliation&gt; &lt;date&gt;2011&lt;/date&gt; &lt;/bibl&gt; &lt;/note&gt; &lt;/notesStmt&gt; The bio for Cholmondeley: &lt;notesStmt&gt; &lt;note type="relatedItem" subtype="source_text"&gt; &lt;listBibl&gt; &lt;bibl&gt; &lt;title level="m" ref="http://purl.dlib.indiana.edu/iudl/vwwp/VAB7387"&gt;The Danvers Jewels&lt;/title&gt; &lt;/bibl&gt; &lt;bibl&gt; &lt;title level="m" ref="http://purl.dlib.indiana.edu/iudl/vwwp/VAB7143"&gt;Red Pottage&lt;/title&gt; &lt;/bibl&gt; &lt;/listBibl&gt; &lt;/note&gt; The intro for Danvers Jewels &lt;notesStmt&gt; &lt;note type="relatedItem" subtype="source_text"&gt; &lt;listBibl&gt; &lt;bibl&gt; &lt;title level="m" ref="http://purl.dlib.indiana.edu/iudl/vwwp/VAB7387"&gt;The Danvers Jewels&lt;/title&gt; &lt;/bibl&gt; &lt;/listBibl&gt; &lt;/note&gt; &lt;/notesStmt&gt; &lt;sourceDesc&gt; &lt;bibl&gt; &lt;title type="introduction"&gt;Introduction to The Danvers Jewels&lt;/title&gt; &lt;author&gt;Inglezakis, Mara L&lt;/author&gt; &lt;affiliation&gt;Indiana University, Bloomington&lt;/affiliation&gt; &lt;date&gt;2011&lt;/date&gt; &lt;/bibl&gt; &lt;/sourceDesc&gt; Because the annotations are related to chunks of text, we are leaning toward embedding these in the source text with attributes to distinguish the type and subtype of note, resp, etc. We haven't yet integrated the scholar-contributed annotations, but we have begun modeling it in the encoding: &lt;note type="scholarly" subtype="gloss" resp="#encoderusername" xml:id="note_017"&gt; Jocosely: “Playfully, in a waggish manner” (&lt;bibl&gt;&lt;title level="m"&gt;OED&lt;/title&gt;&lt;/bibl&gt;) &lt;/note&gt; We could maintain these notes in an external file and point to them via identifiers. Not sure what wold be the best approach at this point so I look forward to your suggestions. --Michelle ----- Michelle Dalmau, Interim Head Digital Collections Services ----- Herman B Wells Library 1320 East 10th Street, Rm W501 Bloomington, Indiana 47405 ----- Web: http://michelledalmau.com Twitter: &lt; at &gt;mdalmau On May 18, 2013, at 3:26 PM, Kevin Hawkins &lt;kevin.s.hawkins&lt; at &gt;ULTRASLAVONIC.INFO&gt; wrote: </pre> Michelle Dalmau 2013-05-18T21:22:06 http://permalink.gmane.org/gmane.text.tei.general/15510 <pre> I was imprecise. This can already be accomplished by doing: &lt;sourceDesc&gt; &lt;p&gt;Encoded based on OCR of &lt;title&gt;An Old Edition&lt;/title&gt;, edited by Jane Smith. Modern commentary converted from Microsoft Word.&lt;/p&gt; &lt;/sourceDesc&gt; or: &lt;sourceDesc&gt; &lt;bibl target="#transcription"&gt;Encoded based on OCR of &lt;title&gt;An Old Edition&lt;/title&gt;, edited by &lt;editor&gt;Jane Smith&lt;/editor&gt;.&lt;/bibl&gt; &lt;bibl target="#commentary"&gt;Modern commentary converted from Microsoft Word&lt;/bibl&gt; &lt;/sourceDesc&gt; as long as you can point to some node with xml:id="transcription" and another with xml:id="commentary". But if you need a surer way to pull out bibliographic descriptions of just source documents or just commentaries, then you will probably want to go with separate TEI documents as Martin Holmes has suggested. --K. </pre> Kevin Hawkins 2013-05-18T19:26:36 http://permalink.gmane.org/gmane.text.tei.general/15509 <pre>On 18 May 2013, at 17:22, "McAulay, Elizabeth" &lt;emcaulay&lt; at &gt;library.ucla.edu&gt; wrote: you're adding an interesting new wrinkle here, that our modern intro is just the latest in a long series. which will bring us down to decisions about what our XML text claims to represent. the "full description in the TEI Header" is a difficult concept. yes, its an unequivocal statement which future human generations can read to understand what decisions you made, but its more or less opaque to machine parsing. I find it hard to escape the notion of genuinely interoperable digital texts. But we've been here before :-} -- Sebastian Rahtz Director (Research) of Academic IT University of Oxford IT Services 13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431 </pre> Sebastian Rahtz 2013-05-18T19:24:22 http://permalink.gmane.org/gmane.text.tei.general/15508 <pre>This has been an interesting thread, and the different views may to some extent be a function of where people are coming from. Thus Kevin and, I believe Elizabeth, are librarians, and Martin Holmes is (primarily) a developer. I look at it as a scholar, and from that perspective an edition with commentary etc. is "my" work, and I think of it as "one" thing, even though it may be internally subdivided. Martin H's use case, where the creator of the annotation is not the editor of the source text, is a different use case from what I envisaged. It may well be a use case that will become more common, as scholars use TCP texts as a point of departure for new editions, e.g. the planned edition of Hakluyt¹s Principal Navigations. But in that edition, if I understand it correctly, a lot of value will be added to the texts through editorial intervention, and the editors may well have a strong sense of ownership of the source texts and their commentary as "their own." So the encoding decisions rest on judgments that are not primarily technical. A comment on Kevin's remark that Martin H's model is "required" if your composite document needs to have more than one &lt;sourceDesc&gt;. Would it make sense to modify the content rules for &lt;sourceDesc&gt; so that you can say things like "this part of the document is born digital, while that part has the following printed source"? Martin Mueller Professor of English and Classics Northwestern University On 5/18/13 1:31 PM, "Kevin Hawkins" &lt;kevin.s.hawkins&lt; at &gt;ULTRASLAVONIC.INFO&gt; wrote: </pre> Martin Mueller 2013-05-18T19:15:12 http://permalink.gmane.org/gmane.text.tei.general/15507 <pre>If the source text being edited has its own front and back matter, I would have different divs in the &lt;front&gt; for each of them, much a 2013 print edition based on a 1971 edition contains the prefaces from both versions, one after another. That said, I don't actually think Martin Holmes's approach is wrong. It just feels more complicated than is required for most cases. But, for example, if you want to describe the source document (perhaps OCR'd text) and the source of the modern commentary (perhaps a Word document that you converted using OxGarage), you might want a separate &lt;sourceDesc&gt; for each, which requires Martin Holmes's model. --K. On 5/18/13 11:53 AM, Lou Burnard wrote: </pre> Kevin Hawkins 2013-05-18T18:31:36 http://permalink.gmane.org/gmane.text.tei.general/15506 <pre> That's an important issue, but it's by no means a clear distinction. Two separate standalone TEI files may be combined into a composite structure in a variety of ways (a teiCorpus which XIncludes them; a third TEI document which XPointers a set of selections from them to build something new; and so on). In this case, I meant standalone TEI files, one for the primary source and one for the editorial commentary. This is attractive in the case of the project I'm thinking of, because the encoders working on the primary source are not (usually) the scholars writing the editorial material, and we have different encoding guidelines for our primary source material as opposed to our born-digital content. Keeping them in separate files makes the distinction clearer and easier to manage. The rendered output is coming from an eXist database anyway, so at that level the distinction between actual documents is not so important. I imagine this sort of thing (in pseudo-code): if requested-doc is primarySource then if (exists(commentary for this doc)) and (user requested commentary) then return compositeEdition(primarySource, commentary) else return primarySource This could be much more sophisticated, of course; you could elect to include only some bits of the editorial material but not others, depending on what the user wants. I know we're not really supposed to be looking at an encoding question from the point of view of rendering, but for me it's an issue of separation of concerns. The rendering options become both easier to manage and potentially more powerful if the concerns are adequately separated in the first place, and separate TEI documents (stored in separate files, as it happens, for convenience) provide a natural way of doing that. Cheers, Martin </pre> Martin Holmes 2013-05-18T16:58:02 http://permalink.gmane.org/gmane.text.tei.general/15505 <pre>When you say "separate files" do you actually mean "different TEI documents"? For me a "file" is just a unit of physical storage, nothing to do with the XML element structure , which is what I thought we were discussing! On 18/05/13 17:35, Martin Holmes wrote: </pre> Lou Burnard 2013-05-18T16:38:34 http://permalink.gmane.org/gmane.text.tei.general/15504 <pre>Sounds like you have your solution.... That level of orderliness is beyond my capacity! :) ________________________________________ From: TEI (Text Encoding Initiative) public discussion list [TEI-L&lt; at &gt;LISTSERV.BROWN.EDU] on behalf of Martin Holmes [mholmes&lt; at &gt;UVIC.CA] Sent: Saturday, May 18, 2013 9:35 AM To: TEI-L&lt; at &gt;LISTSERV.BROWN.EDU Subject: Re: Introductory material for a primary source text I'm certainly not overloaded, but: When I asked this question I felt a bit embarrassed to do so, since it seemed such a novice-level thing to enquire about, but the range of disagreement amongst our most august authorities shows that even with such a simple scenario we still lack consensus on best practices. I think I'm with Sebastian at the moment: I don't like the idea of using &lt; at &gt;type to distinguish &lt;front&gt;s that are modern from those that are not, and I'm reluctant to rely on prose in the header to alert a putative text-analysis machine about which bits it should look at and which it shouldn't. I think the cleanest thing is separate files. Although traditional critical editions are gathered together into a single volume for physical practicality, there's no reason why digital resources should be so; we can present the original source text clean of editorial material or entangled with it, depending on our purposes. Cheers, Martin On 13-05-18 09:22 AM, McAulay, Elizabeth wrote: </pre> McAulay, Elizabeth 2013-05-18T16:38:05 http://permalink.gmane.org/gmane.text.tei.general/15503 <pre>I agree with Lou and Sebastian on these two points: 1. An edition of the kind discussed here is "composite object", consisting of a modern text created for this occasion and a modern version of a predigital one. 2. &lt;group&gt; is the least worst alternative. It is counter-intuitive to put into the &lt;front&gt; element of the source text an independent document that may well be longer and more complex than the source text. Consider a critical edition of Lycidas (~200 lines), which might have a 60 page introduction and 500 notes.Not to speak of the fact that the source text may have its own &lt;front&gt; matter. Perhaps this is a topic for dicussion and clarification in the Guidelines that say nothing about it. But the genre of the critical edition, where commentary may outweigh source text by a wide margin, is a fundamental scholarly genre. It is also a genre that may have a new life in the digital world, and the TEI needs to have some guidance for scholars. On 5/18/13 11:23 AM, "Sebastian Rahtz" &lt;sebastian.rahtz&lt; at &gt;IT.OX.AC.UK&gt; wrote: </pre> Martin Mueller 2013-05-18T16:37:51 http://permalink.gmane.org/gmane.text.tei.general/15502 <pre>I'm certainly not overloaded, but: When I asked this question I felt a bit embarrassed to do so, since it seemed such a novice-level thing to enquire about, but the range of disagreement amongst our most august authorities shows that even with such a simple scenario we still lack consensus on best practices. I think I'm with Sebastian at the moment: I don't like the idea of using &lt; at &gt;type to distinguish &lt;front&gt;s that are modern from those that are not, and I'm reluctant to rely on prose in the header to alert a putative text-analysis machine about which bits it should look at and which it shouldn't. I think the cleanest thing is separate files. Although traditional critical editions are gathered together into a single volume for physical practicality, there's no reason why digital resources should be so; we can present the original source text clean of editorial material or entangled with it, depending on our purposes. Cheers, Martin On 13-05-18 09:22 AM, McAulay, Elizabeth wrote: </pre> Martin Holmes 2013-05-18T16:35:37 http://permalink.gmane.org/gmane.text.tei.general/15501 <pre>On 18 May 2013, at 17:10, Lou Burnard &lt;lou.burnard&lt; at &gt;retired.ox.ac.uk&gt; wrote: i agree that a &lt;text&gt; in a &lt;group&gt; is definitely preferable to &lt;front&gt;. It would fine, if we had a better way than &lt; at &gt;type for distinguishing between source and editors commentary. I resign myself to not being agreed with about &lt; at &gt;type :-} that does seem to be confusing as well, so preferably not. So &lt;group&gt; is the least worst alternative. -- Sebastian Rahtz Director (Research) of Academic IT University of Oxford IT Services 13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431 </pre> Sebastian Rahtz 2013-05-18T16:23:07 http://permalink.gmane.org/gmane.text.tei.general/15500 <pre>I have seen texts that include multiple introductions, each from a different time period. I would simply use divisions to separate different introductions or preliminary material from each other. And I would use a full description in the TEI Header to articulate that this is a new edition with new front matter, and therefore, if anyone wants to analyze the textual components, they are alerted via the TEI Header. So, I'm still lobbying for Kevin's approach. Normally I would say we've totally overloaded the poor questioner, but as it's Martin H., I think he'll survive! He's used to this listserv. Lisa ________________________________________ From: TEI (Text Encoding Initiative) public discussion list [TEI-L&lt; at &gt;LISTSERV.BROWN.EDU] on behalf of Sebastian Rahtz [sebastian.rahtz&lt; at &gt;IT.OX.AC.UK] Sent: Saturday, May 18, 2013 8:58 AM To: TEI-L&lt; at &gt;LISTSERV.BROWN.EDU Subject: Re: Introductory material for a primary source text On 18 May 2013, at 16:53, Lou Burnard &lt;lou.burnard&lt; at &gt;RETIRED.OX.AC.UK&gt; wrote: you're just putting the problem up a stage - what if the source is composed of several &lt;group&gt;s of texts? -- Sebastian Rahtz Director (Research) of Academic IT University of Oxford IT Services 13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431 </pre> McAulay, Elizabeth 2013-05-18T16:22:04 http://permalink.gmane.org/gmane.text.tei.general/15499 <pre> Well, what would you do? The truth of the situation is that we're dealing with a composite object, one part of which is a modern text and one part of which is a modern version of an pre-digital one. Using &lt;group&gt; still seems like the right solution to me. If the source &lt;text&gt; has its own &lt;group&gt; that would still be distinct from the modern &lt;text&gt;. &lt;text type="composite"&gt; &lt;!-- the outer group corresponding with the digital thingie --&gt; &lt;group&gt; &lt;text type="commentary"&gt;&lt;front/&gt;&lt;body/&gt;&lt;back/&gt; &lt;!-- of my modern commentary --&gt;&lt;/text&gt; &lt;text type="source"&gt; &lt;group&gt; &lt;text/&gt; ... &lt;text/&gt; &lt;/group&gt; &lt;/text&gt; &lt;/group&gt; &lt;/text&gt; Are you voting for &lt;teiCorpus&gt;? </pre> Lou Burnard 2013-05-18T16:10:44 http://permalink.gmane.org/gmane.text.tei.general/15498 <pre>On 18 May 2013, at 16:53, Lou Burnard &lt;lou.burnard&lt; at &gt;RETIRED.OX.AC.UK&gt; wrote: you're just putting the problem up a stage - what if the source is composed of several &lt;group&gt;s of texts? -- Sebastian Rahtz Director (Research) of Academic IT University of Oxford IT Services 13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431 </pre> Sebastian Rahtz 2013-05-18T15:58:05 http://permalink.gmane.org/gmane.text.tei.general/15497 <pre>Yes, I was disagreeing with Martin, as you surmise. However, I also think I disagree with you. What if the source text being edited has its own front and back matter? You don't want to confuse that with your modern front and back. Hence, I think the &lt;group&gt; approach is generally to be preferred, if the modern parts are too big to be treated just as annotations. On 18/05/13 15:34, Kevin Hawkins wrote: </pre> Lou Burnard 2013-05-18T15:53:21 http://permalink.gmane.org/gmane.text.tei.general/15496 <pre>Kevin is recommending, as I understand it, &lt;front type="editorial" xml:id="documentation"&gt; which I, as a consumer, find quite unsettling, because it relies solely on an arbitrary value for the notoriously private &lt; at &gt;type to convey to the user that this text is not in the original. If I am doing an analysis of the language used in the original text, I have no information here which will let me filter out the commentary section. I think &lt; at &gt;type has its place, but is really _only_ useful for the project or person who put it in. For the rest of accessing the TEI text in that great infinite archive we talk about, it doesn't convey anything. And I would claim that, by default, &lt;front&gt; represents the front matter of the _original text_. the same applies to the &lt;group&gt; solutions, of course. -- Sebastian Rahtz Director (Research) of Academic IT University of Oxford IT Services 13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431 </pre> Sebastian Rahtz 2013-05-18T15:40:06 http://permalink.gmane.org/gmane.text.tei.general/15495 <pre>I like Kevin's approach as described below. If there are footnotes or annotations, I would integrate those into the body and use either or both the "cert" and "resp" attributes to note the modern commentary. Best, Lisa ------------------------------------- Elizabeth "Lisa" McAulay Librarian for Digital Collection Development UCLA Digital Library Program http://digital.library.ucla.edu/ email: emcaulay [at] library.ucla.edu ________________________________________ From: TEI (Text Encoding Initiative) public discussion list [TEI-L&lt; at &gt;LISTSERV.BROWN.EDU] on behalf of Kevin Hawkins [kevin.s.hawkins&lt; at &gt;ULTRASLAVONIC.INFO] Sent: Saturday, May 18, 2013 7:34 AM To: TEI-L&lt; at &gt;LISTSERV.BROWN.EDU Subject: Re: Introductory material for a primary source text When Lou says that "this approach" seems exactly right, I assume he is referring to what Martin Mueller and Rob suggested (and not what Martin Holmes suggested). I am more inclined to put the modern commentary in &lt;front&gt; and, if desired, also in &lt;back&gt;, with the encoded source document in &lt;body&gt;. This is analogous to a scholarly edition in print, which would have the commentary at the front, likely with page numbers in roman numerals, and possibly also an index or two at the back. I personally don't think you need separate headers: most of the header describes the digital edition, whereas the &lt;sourceDesc&gt; describes the source document. If your edition is of more than one text, as in Rob's case, you could have: &lt;text xml:id="digitalTemple"&gt; &lt;front type="editorial" xml:id="documentation"&gt; &lt;p&gt; &lt;ptr target="#textIntro"/&gt; &lt;ptr target="#fileStructure"/&gt; &lt;ptr target="#encodingScheme"/&gt; &lt;ptr target="#acknowledgments"/&gt; &lt;ptr target="#bibliography"/&gt; &lt;/p&gt; &lt;/front&gt; &lt;group&gt; &lt;text type="primary" xml:id="textA"&gt; &lt;body&gt; &lt;p/&gt; &lt;/body&gt; &lt;/text&gt; &lt;text type="primary" xml:id="textB"&gt; &lt;body&gt; &lt;p/&gt; &lt;/body&gt; &lt;/text&gt; &lt;text type="primary" xml:id="textC"&gt; &lt;body&gt; &lt;p/&gt; &lt;/body&gt; &lt;/text&gt; &lt;/group&gt; &lt;/text&gt; --Kevin On 5/17/13 6:25 PM, Lou Burnard wrote: </pre> McAulay, Elizabeth 2013-05-18T15:29:23 http://permalink.gmane.org/gmane.text.tei.general/15494 <pre>When Lou says that "this approach" seems exactly right, I assume he is referring to what Martin Mueller and Rob suggested (and not what Martin Holmes suggested). I am more inclined to put the modern commentary in &lt;front&gt; and, if desired, also in &lt;back&gt;, with the encoded source document in &lt;body&gt;. This is analogous to a scholarly edition in print, which would have the commentary at the front, likely with page numbers in roman numerals, and possibly also an index or two at the back. I personally don't think you need separate headers: most of the header describes the digital edition, whereas the &lt;sourceDesc&gt; describes the source document. If your edition is of more than one text, as in Rob's case, you could have: &lt;text xml:id="digitalTemple"&gt; &lt;front type="editorial" xml:id="documentation"&gt; &lt;p&gt; &lt;ptr target="#textIntro"/&gt; &lt;ptr target="#fileStructure"/&gt; &lt;ptr target="#encodingScheme"/&gt; &lt;ptr target="#acknowledgments"/&gt; &lt;ptr target="#bibliography"/&gt; &lt;/p&gt; &lt;/front&gt; &lt;group&gt; &lt;text type="primary" xml:id="textA"&gt; &lt;body&gt; &lt;p/&gt; &lt;/body&gt; &lt;/text&gt; &lt;text type="primary" xml:id="textB"&gt; &lt;body&gt; &lt;p/&gt; &lt;/body&gt; &lt;/text&gt; &lt;text type="primary" xml:id="textC"&gt; &lt;body&gt; &lt;p/&gt; &lt;/body&gt; &lt;/text&gt; &lt;/group&gt; &lt;/text&gt; --Kevin On 5/17/13 6:25 PM, Lou Burnard wrote: </pre> Kevin Hawkins 2013-05-18T14:34:41 http://permalink.gmane.org/gmane.text.tei.general/15493 <pre> This was true once. However when sourceDoc et al were added we also made it possible for surface zone and friends to carry coordinate information directly. So the mapping Alexander wants should be quite straightforwardif he maps to a soyrceDox rather than to a text. ----- Reply message ----- From: "Conal Tuohy" &lt;conal.tuohy&lt; at &gt;VERSI.EDU.AU&gt; To: "TEI-L&lt; at &gt;LISTSERV.BROWN.EDU" &lt;TEI-L&lt; at &gt;LISTSERV.BROWN.EDU&gt; Subject: ALTO=&gt;TEI experience or advice Date: Sat, May 18, 2013 04:07 In METS/ALTO, the cartesian coordinates are attached to content directly, whereas in TEI the the textual content (transcript) and coordinates are kept separate (within the tei:text and tei:facsimile elements, respectively). onal </pre> Lou Burnard 2013-05-18T08:27:54 http://permalink.gmane.org/gmane.text.tei.general/15492 <pre>I wrote an XSLT for this some years ago, which, I'm sorry to say, I have since lost in my travels. The XSLT was an exercise at generating TEI "facsimile" markup. In METS/ALTO, the cartesian coordinates are attached to textual content directly, whereas in TEI the the textual content (transcript) and coordinates are kept separate (within the tei:text and tei:facsimile elements, respectively). From memory, my XSLT used two XSLT "modes" to make two passes over the METS/ALTO content; in one pass (xsl:apply-templates mode="facsimile") it generated TEI facsimile, surface, and zone elements, and in the second pass (mode="transcript") it created a TEI text, etc, and within that,TEI ab elements, each of which it linked to the corresponding zone by URI reference. I hope that helps! Conal </pre> Conal Tuohy 2013-05-18T01:57:17 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.text.tei.general