SIG:Correspondence/task-force-correspDesc

From TEIWiki

Jump to: navigation, search

<- back to SIG:Correspondence

Contents

Task force "correspDesc"

  • Members: Marcel Illetschko, Sabine Seifert, Peter Stadler
  • "Pusher": Mariana Gomes

We will try to make our work as transparent and open as possible so others can chime in and contribute at every stage. Please feel free to make use of all the various comment/edit/share/fork functions or to contact us directly.

Goal

Create a formal proposal for a <correspDesc> (or whatever name) element for approval by the TEI council as a new element of the TEI standard. This work shall be based on the "second draft for a correspondence module and a wrapper element for correspondence meta data" as it was started during the 2012 SIG meeting.

Background

There have been several P4 projects with a dedicated correspondence schema (DALF probably being the renowned). While several correspondence tags (<postscript> etc.) are available from the TEI P5 standard, there is no specific way of dealing with correspondence meta data (e.g. sender, addressee) in the current TEI header. Since that's an impediment on the interoperability and interchange of these texts and since there are good practice models working with a <correspDesc>, the SIG Correspondence started to draft such a proposal during its 2012 meeting and now set up a task force for elaborating this draft.

Repositories

https://github.com/TEI-Correspondence-SIG/correspDesc This will be the main place for bringing together the proposal.

Recent discussion

The SIG's proposal already caused a lot of traffic at the mailing list. See the SIG:Correspondence/task-force-correspDesc/summary of the recent discussion here - sorted by elements and furnished with the task force's view on the topics.

Meetings

Conf Call Nov 13, 2014

The schema was updated with the new changes, examples were encoded according to new schema, the GitHub repository was tidied up.

Conf Call Nov 04, 2014

After the discussions at the TEI conference, several changes and adaptions were and will be made to <correspDesc>. The schema will be slimmed down to reach a consensus on the proposal at the next TEI Council meeting (Nov 17-19). Focus is on the proposed elements <participant> und <context>. For now, <transmission> is being discarded but will be added again later.

Changes within <correspDesc>: The proposed elements <sender> and <addressee> are replaced by a proposed element <participant>. To denote the role the attribute @role will be used (role="sender", role="recipient"). The name of the agent, the place and date are child elements of <participant> (rather than on a parallel level as it was before). The name of <context> is too generic, so it is renamed <correspContext>.

There are still some open questions concerning the content model of <participant> and <correspDesc>. Trying out of schema with examples.

PAPER at TEI conference in Evanston 2014, Oct 22

Paper entitled "Towards a correspondence module in the TEI" held at the annual TEI conference to present the conceptualization and implementation of the proposed <correspDesc>. Thanks for the questions, discussions, examples!

The paper can be found here, the Prezi is here, the abstract is here.

Conf Call Oct 16, 2014

And another conf call in preperation of the TEI conference discussing our paper on the new <ct:correspDesc> and preparing the SIG Meeting at the conference.

TEI Tweet Chat, Oct 9, 2014

There was a TEI Tweet Chat held, the first one ever! A summary can be found here.

Conf Call Oct 03, 2014

A German holiday does not stop the task force from working: Discussing and refining structure and content of Chicago paper.

A tweet chat on Twitter will be held on Thursday, October 9, 4 pm (GMT+2) with questions and information on the SIG's activities for everybody to join and participate in. Preparation of possible topics to tweet about and to discuss.

Conf Call Sept 25, 2014

Discussion of structure and content of Chicago paper (presentation of work and <correspDesc>).

Preparation of a SIG correspondence Tweet Chat, suggested by Paul O'Shea. Started thinking about possible topics that might be of interest:

  • general focus of the SIG
  • current work on „correspDesc“
  • showcase projects
  • favourite letter project?
  • people know of any digital editions of correspondence/working with historic letters?
  • correspSearch and letter repositories/search engines, etc.

Conf Call Sept 12, 2014

Just some organizational stuff in preparation for the TEI2014 conference in Chicago, discussing content and structure for our paper, preparing SIG correspondence meeting.

Conf Call Sept 05, 2014

Peter created an example for an alternative <ct:correspDesc> context and structure (<correspDesc> as part of <profileDesc>).

To-Do: Looking for examples and trying them out with "both" conceptions of <correspDesc>.

Conf Call Sept 03, 2014, task force with TEI council members

Participants: Syd Bauman, Lou Burnard, James Cummings, Martin Holmes, Sabine Seifert, Peter Stadler.

Discussions about:

1. <correspDesc> within <profileDesc> instead of <sourceDesc>

  • Basically a question of whether one considers <correspDesc> as containing bibliographic information describing the source (<sourceDesc>) or as containing information describing the text (<profileDesc>).
  • As part of <sourceDesc>, all overlap with <msDesc> needs to be avoided (this done by concentrating on actors and act of transmission).
  • General consensus of understanding <correspDesc> as text-describing element --> as direct child of <profileDesc>, next to <settingDesc>, <textDesc> etc.
  • Other ideas thrown around: allowing <correspDesc> in both <sourceDesc> and <profileDesc>; <correspDesc> looks very much like <channel>

2. <correspDesc> and model.biblLike

  • If <correspDesc> is a member of model.biblLike, then it can be used anywhere what might lead to confusion.
  • <correspDesc> is supposed to be element for the metadata in the TEI header, not for text encoding in <body>. Therefore, it won't be part of model.biblLike. (And with <correspDesc> within <profileDesc>, model.biblLike not needed.)

3. name of wrapper element within <profileDesc>

  • suggestions: correspondenceDesc/correspondence, communication and listCommunication, etc., will be decided later

4. conception of <ct:sender> vs. <author>

  • The need to differentiate between sender and author is based on the conception of <correspDesc> as dealing not with an object but with an act of communication
  • More details on this conception can be found in the proposal and the summary of recent discussion.

5. element <participant> as an alternative?

  • Discussion of <sender> vs. <participant role=”sender”> --> <sender> does not equal <persName role=”sender”>, is really new (container) element.
  • @resp for intellectual content so it cannot be used for things related to transmission
  • right now within <sender> and <addressee> allowed: structured content + prose and text content + empty element --> advice to better not allow prose and text content (problem of <msDesc>) and to give information only in a structured way (persName, orgName, name instead of just

    or sth.), provide for "anonymous"

  • When <participant> instead of <sender>/<addressee>, then other information on dates and places also within <participant> as a container element or do they remain outside of <participant>?

6. to-do list

  • model should be tried out and tested with different kinds of correspondence
  • examples put on a wiki page or somewhere so that TEI community can play with them and try out
  • more publicity for more feedback

Conf Call Sept 03, 2014, just task force

Revision of recent discussions via TEI mailing lists and uploading a summary with pros and cons, etc..

Discussion of several elements and ideas, amongst others:

  • The element <ct:transmission> will need more structure.
  • The conception of <correspClass> was not convincing and its function can sufficiently be replaced by textClass/keywords mechanism in the TEI header. <correspClass> is therefore given up.
  • Instead of one element for sender and addressee perhaps element <participant>, that might be also useful for other (not correspondence-related) contexts? Or just one element <participant> that also contains places and dates?

Conf Call Aug 21, 2014

Discussion of <participant>/<party role=”...”> vs. author-sender. A TEI element <participant> may be very useful in other contexts, too. But sth. like <persName role="addressee"> has the disadvantage that the addressee may be an organization rather than a person. Giving up of developing an element <correspClass>. Discussion whether <msDesc> is really the right place for <correspDesc>, but putting it in <msDesc> to avoid redundancy is not satisfying either (more than one source for text possible, electronic communication, etc.). Discussion of James Cummings' helpful notes via email on the element <transmission>, it needs more structure(e.g. multiple stages of transmission, changes in format/material, method of delivery, agency responsible, changes in dateTime or location). In <transmission> link to text witnesses (with pointers?), respStmt needed.

Conf Call July 11, 2014

Peter reported from TEI Council meeting in Oxford. General tendency there: introduction of new element <correspDesc> convincing but some of the child elements seem to double with what can be done by already existing elements, e.g. question raised whether <sender>/<addressee> are just "syntactic sugar" or really better than <persName type=”sender”>/<name type=”sender”> etc. (For us, the task force: yes, it is better!) With changes in <correspDesc>, CIF needs to be reconsidered as well (and maybe made independent from <correspDesc>.

Conf Call July 07, 2014, task force & Syd Bauman

Syd had a closer look at our proposal and suggested to put information on places/dates concerning addressee or sender within these two elements rather than having them next to them on the same level. This solution would also suit for several senders and differing dates/places, etc. <correspClass> seems not convincing and looks close to <textClass>. Discussion of <context> and whether this information could be put into <correspDesc>, with pointers or sth.

Meeting June 17-18, 2014

Two days of intense work on the proposal, mainly reworking and adding of prose description with encoding examples in the chapter "Implementation of the correspondence description":

  • differentiation between sender and author
  • differentiation between the intended and the actual (place of) addressee
  • repeating of direct child elements of <correspDesc>
  • use of <transmission> in contrast to the TEI element <channel>
  • use of <context> with varying evaluations of the position within a correspondence
  • text classification, classification schemes (and the lack thereof) and materiality aspects.

Adding of examples to formal specification of various elements.

Preparing proposal for handing in as feature request soon. All updates available at GitHub, with additional HTML page for easy reading.

Conf Call June 12, 2014

The main part of this conf call was held with Stefan Dumont (BBAW) who set up the web service "correspSearch" that analyzes beacon files based on the task force's <correspDesc> format from several digital correspondence editions and thus providing the possibility for searching through meta data of these editions.

Discussion of information needed from participating digital editions/projects: responsible person (preferably with e-mail address, in <titleStmt>); name of the beacon file (in <titleStmt>); latest update (in <revisionDesc>); information on licensing (with URL etc., <publicationStmt>); bibliographic reference (<sourceDesc>, <bibl>).

There is also the question of how to handle duplicates, i.e. one letter in more than one participating digital edition and thus resulting in two entries. Especially in these cases, a correspSearch ID for each entry is desirable (referencing to them with @ref).

<correspClass> is not yet part of the web service but could be implemented.

The web service "correspSearch" will be developed further and soon be publicly available online in a beta version, with corresponding API.

Conf Call May 28, 2014

It's been our fastest conf call ever! We reported on last weeks activities (not much, but some additions to the GitHub repo) and agreed on a f2f meeting on June 17/18 in Berlin. Next conf call is scheduled for June 12, 10 AM.

Conf Call May 20, 2014

Reports on two conferences:

  1. Berlin: "Datenmodellierung in digitalen Briefeditionen und ihre interpretatorische Leistung. Ontologien, Textgenetik und Visualisierungsstrategien"

Peter presented on "interoperability of letter collections" and touched on the work of the task force. A lot of people are planning (dreaming) of connecting letter collections.

  1. Vienna: Editorenkolloquium

"The old world" – a strong bias against digital editions and only few letter projects presented at the conference. Nevertheless Marcel waved the correspDesc flag during his talk.

correspSearch

Stefan Dumont from the BBAW created a web service for querying correspDesc data. This web service will be presented as an open beta at the end of June. During his work on correspSearch he encountered several bugs and missing features in the CIF standard that need to be addressed now. (see next point)

proposed schema changes

  • rename addressee to receiver? We decided not to change that right now but we will wait for the natives to chime in.
  • CIF: @when-iso instead of @when: For the web service there is a strong need for standardized date formats. --> Yes, will implement
  • CIF: clearly defined meta data with information about origin, person in charge, last change date etc. --> Yes, will implement
  • CIF: shall @notBefore be required to have a counterpart @notAfter? --> No, rejected

Roadmap

  • Next conf call on May 28, 10 AM
  • f2f meeting on June 17/18 (date needs to be finalized)
  • feature request due before next Council f2f, June 30
  • correspSearch will go public at the end of June

Conf Call March 17, 2014

Peter gave a short summary of the workshop “Digitising experiences of migration” (Mellon Centre for Migration Studies, Omagh, Northern Ireland, March 14-15) – and we discussed the difficulties with lacking authority-file-data for our Correspondence Interchange Format (CIF). Marcel (located at the Austrian National Library) will have a look at Europeana’s interchange format to find out whether it could be combined with the task force’s CIF; Peter will get in contact with the BBAW in order to create some sort of web service, which will facilitate for data interchange. Roadmap: June 2014 TEI-Council in Oxford – proposal should be ready for feature request – October 2014: TEI-Meeting (SIG meeting and/or <correspDesc>-workshop)

Conf Call Feb. 18, 2014

We discussed the <correspType> element, decided to change its name to <correspClass> (according to <textClass>). Then we dealt with the preparation of the presentation in Berlin, Feb. 26, 2014 (workshop: “Briefeditionen um 1800 – Schnittstellen finden und vernetzen”). To-dos: enhancing second draft ODD, enhancing text of proposal (theoretical considerations, preceeding work). Inform the communtiy via mailing list to get some feedback.

Presentation Feb. 02, 2014

Short presentation of the element <correspDesc> (as developed so far) at workshop on editions of correspondence around 1800 at Berlin-Brandenburg Academy of Sciences in Berlin. The slides of the presentation (in German) are available here and at GitHub.

Next conf call will be on March 17, 2014.

Conf Call Jan. 30, 2014

Discussing new changes in schema (<correspDesc> now repeatable, <note> and @ref in <correspDesc> possible, probably new name for <ct:correspType> needed, etc.). Latest version of the proposal can be found on GitHub. Proposal fits well for encoding e-mails. Looking at real example of encoding a letter with our schema, discussing detailed structure and flat structure (CIF).

Meeting Jan. 13-14, 2014

Meeting of task force in Berlin. We pushed forward the work on an ODD file for the correspondence module with the new element <correspDesc> (and several child elements). The result can be found at GitHub: proposal.xml. While the formal specification can already be considered alpha (i.e. you can build a schema and try it out), we are still behind enhancing the documentation with information on background, theoretical considerations, notes on implementation, etc.

Next conf call will be on Jan. 30 (09:30 CET), 2014.

Conf Call Jan. 8, 2014

This conf call mainly dealt with the preparation of the face-to-face meeting in Berlin on Jan. 13-14. To-do-list for this meeting: reading of Mariana's e-mail and files, and of literature mentioned during conf calls (editio, Emert, etc.).

General to-do-list: enhancing second draft ODD, enhancing text of proposal, have a look at Genetic Editions proposal.

We decided to put together a bibliography of literature on editing letters that will help writing our proposal.

All minutes from TEI conferences in TEI now online here.

Conf Call Dec. 18, 2013

We closely dealt with the question, what information we want to encode within the new <correspDesc> element and what elements will be needed. Child elements of/information given within <correspDesc> will be: <sender>, <addressee>, <placeSender>, <placeAddressee>, <dateSender>, <dateAddressee>, Korrespondenzstelle, type of correspondence, bearer and method of message, information whether there is an envelope or not (description of envelope in <msDsc> and transcription in <body>).

What will not be encoded within <correspDesc> because it is not letter-specific: <incipit>, printed things like form fields or letterheads, <stamp> and postage stamps - all this information is encoded within <msDesc>.

There will be a short presentation of the work of the SIG Correspondence and especially the <correspDesc> task force at the meeting on technical interfaces in German digital editions of letters around 1800 at the Berlin-Brandenburg Academy of Sciences in Berlin on February 26th.

Next conf call will be on Jan. 8, 10AM.

Conf Call Nov. 26, 2013

Second meeting of the task force. Main topic was bringing together correspondence specific metadata (that should be encoded within a new <correspDesc> - or otherwise called - element but not necessarily): sender, addressee, place of sender, date of sending, place of addressee, date of receiving, type of correspondence (letter, postcard, etc.), incipit, envelope, context (previous/next letter), postmark, stamp, perhaps transmitter of the message and method of delivery.

We had a closer look at the DALF Preliminary P5 Proposal and the <dalf:letDesc> element as well as the <correspDesc> element used by Weber-Gesamtausgabe. We dealt more closely with the problems of how and where to encode envelopes, the type of message, postmarks, etc. Seals are not specific for letters but for other manuscripts as well. Therefore, changing their encoding (in <sealDesc>) is not necessary.

Next steps: Peter starts reworking/extending the 'second draft', Marcel turns the minutes of the SIG Correspondence meetings into TEI, and starts writing goal and background information, DALF and WEGA, development of this proposal, etc. Sabine starts putting together information on the letter specific metadata and their description, place, examples, etc. as well as building up a structure for our proposal.

Next conf call will be on Dec. 18, 10AM. At the face-to-face 'boot camp' on January 13-14 in Berlin, we will work on the customization ODD file.

Conf Call Nov. 6, 2013

First meeting of the task force. We summed up the SIG meeting at Rome and made sure that we all shared (more or less) the same ideas about the goal of the task force. Had a look at the "second draft" from last year which we decided to take as a starting point for future work.

Furthermore we decided on GitHub as a central repository for carrying out the work on our proposal and envisioned a face2face meeting at the beginning of next year as sort of a "boot camp" (scheduled for Jan. 13/14 in Berlin).

Next conf call shall be on Nov. 26, 12PM were we will look more closely at DALF and try to narrow down the "correspondence specific meta data".

Literature and previous work

Personal tools