<title>, <title type="alt"> and other options

[arlogriffiths]

@michaelnmmeyer — The EGD model that we have converged to for encoding inscriptions is not fully suitable to editions of texts that actually have their own titles, like <title>Digital Critical edition of the Meghadūta with Malinātha's commentary</title>, because we lose the ability to encode the original text title as such. This problem does not arise for inscriptions which never have their own title. The situation is somewhat analogous to that of <author>, which we have reserved for EGC usage in case of known original authors.

We could deploy <title type="alt"> for recording the original title. But it would be semantically underspecified.

Would it be a nuisance for you if the EGC model allowed <title type="original" xml:lang="und"> for recording original titles, besides the elements <title> and <title type="alt">?

[michaelnmmeyer]

Why do you need the original text title to be stored separately, in a different <title> element? In <title>Digital Critical edition of the Meghadūta with Malinātha's commentary</title>, the original title is already given. Why do you also want a <title type="original">Meghadūta</title>?

[arlogriffiths]

For the same reason that we want to be able to state the <author> in a manner that a machine can index. Even if the machine can distinguish English from foreign words, how would it know that the title of the work being edited is Meghadūta and not Mallinātha? Don't we want to be able to draw up an alphabetically ordered list of original titles of works edited in our database?

Am I missing something?

[michaelnmmeyer]

Don't we want to be able to draw up an alphabetically ordered list of original titles of works edited in our database?

Given that this would concern only a few files, I most likely won't use the original title at all. I think this should be put on the backburner for DHARMA itself.

[arlogriffiths]

Thanks @michaelnmmeyer !

It seems that we need to distinguish two things:

1. What 'I' deem essential from a semantic point of view and hence insist on including in one or more of our encoding models;
2. What 'you' will use for the system you are building for the DHARMA project — NB: from <encodingDesc> #368 I gather than you don't have a principled objection to the presence in source code of certain elements that 'your' system will ignore.

I use quotation marks because we are talking about our collective undertaking and I would prefer us to speak about it in corresponding terms.

1. To me it is unthinkable not to encode as such the original title of a work being edited digitally, if it has one (just as we encode its <author> if his name is known). This is a different thing from a title assigned by an editor to a work of digital scholarship, like the edition of an inscripton or the edition of a manuscript-transmitted text. Our system as a whole must be able to represent this difference. Hence, I would now suggest:
   -- use @type="editorial" on the main <title> elements of our EGD-compliant editions.
   -- keep the optional @type="alt".
   -- offer (in the EGC model at least) the optional @type="original" .
2. As for what the dharmalekha.info will do with any <title type="editorial"> encoded in EGC-compliant files, I can live with it doing nothing, as seems to be your intention. Still, I wonder why I should not be allowed to expect of our database that it enables a reader to find a list of all original titles in a logical order. We have several dozens of texts with original titles already encoded, and the number will keep growing. I don't know where you draw the line between 'few' and 'many' and I think as PI I should have a say in what functionality is to be built into our system.

[michaelnmmeyer]

from #368 I gather than you don't have a principled objection to the presence in source code of certain elements that 'your' system will ignore.

Basically, yes. What should be avoided is 1) encoding the same piece of information differently (as is the case for the apparatus, which is structured very differently in inscriptions and in manuscript editions); and 2) using the same encoding for representing semantically different things (this one is harder to spot, I have no example in mind).

As concerns titles, if you really want a @type="original", why not. Still, all editions (EGC-based or EGD-based) should have a mandatory plain <title>, and only one such element. We can assume that, when @type is not given, it defaults to editorial (so that EGD-based editions remain valid).

Still, I wonder why I should not be allowed to expect of our database that it enables a reader to find a list of all original titles in a logical order.

We have a "sort by title" functionality already. If you also want the ability to sort files by original title, we will need to have both "sort by editorial title" and "sort by original title". We can reasonably expect that the reader will not know what the terminology refers to, so we will also need to make the difference explicit in the display, as in:

Editorial title: Digital Critical Edition and Translation of an Old Javanese Penal Code
Original title: Kuṭāra Mānava

The reader will notice that most texts do not have an original title, so he will likely try both sorting options. And he will also not know the sort order of records that do not have an original title. Indeed, when a record does not have an original title, should it appear at the beginning of the sorted list or at its end? Or maybe we should use the editorial title as fallback, so records would be sorted, first by original title, then by editorial title?

By the way, what information should the editorial title record? For the above example, should it just be "Digital Critical Edition and Translation of an Old Javanese Penal Code", or should it also include the original title, as in "Kuṭāra Mānava: Digital Critical Edition and Translation of an Old Javanese Penal Code"? If we assume editorial titles should make sense on their own, without looking at the original title, then the second option is better. But maybe you want the original title and the editorial title to be mechanically combined in a single value? And what if there is an author name, should it go in the generated title string as well?

[danbalogh]

I don't really have a solution for this. Since we are apparently dealing with a relatively small number of files, and since TEI apparently provides out-of-the-box solutions only for encoding the title of the current digital work (in <title>) and the identifier of a particular manuscript, but nothing in between, I would prefer if we didn't have to come up with custom encoding solutions. Would it not be possible to change the <title> of the editions concerned so that it always begins with the original title? E.g. "Kuṭāra Mānava: Digital Critical Edition and Translation of an Old Javanese Penal Code", as suggested by Michaël above. Likewise, "Meghadūta: Digital Critical Edition with Mallinātha's Commentary". This would allow us to stick to a single option to "sort by title" and yet show the original titles in alphabetical order.
Interestingly, most of the actual examples in the TEI documentation seem to use <title> for the original title, which is pretty confusing. There's also an example there which used <title type="sub"> for a subtitle, so I'm wondering if we could go with

<title>Kuṭāra Mānava</title>
<title type="sub">Digital Critical Edition and Translation of an Old Javanese Penal Code</title>

and sort only by the main title, which would then have to be the original title for works that have one.

Then again, I can also sympathise with Arlo's concern for the separate and clear encoding of an original title and the title of the edition. In that case I could imagine <title type="original"> or, perhaps, <title xml:lang="san-Latn">, with the understanding that a title without @xml:lang (and hence in English, as the rest of the TEI header), is the editorial title, and if another title with @xml:lang is present, than that is the original title in whatever language that the value indicates. For sorting, I then assume that if an original/language-specified title exists, then that title should take precedence over the editorial title. I.e. instead of allowing separate sort options for original and editorial title, just use the original title when present. I assume that end users would not be interested in sorted lists where e.g. "digital edition of y" precedes "electronic edition of x".

[arlogriffiths]

I would like to propose the following conclusion.

1. all files must have <title>
2. in the absence of @type this will be understood by the machine to be = <title type="editorial">
3. when such an editorial title applies to a work that has an original title, all effort must be made to formulate the contents in such a way that this title-of-work occupies the first place in the <title type="editorial>, and that it also include a mention of the name of any known <author>. Examples:

• "Kuṭāra Mānava: Digital Critical Edition and Translation of an Old Javanese Penal Code"
• "Meghadūta or Cloud Messenger, by Kālidāsa: Digital Critical Edition"
• "Meghadūta: Digital Critical Edition with Mallinātha's Commentary"

5. all files may additionally contain <title type="original" xml:lang="xxx-Latn"> to serve specifically for recording the original title in romanized form
6. neither from @michaelnmmeyer nor from @ajaniak is it expected that the respective display routines for inscriptions and for manuscript-based editons do any processing on any <title type="original" xml:lang="xxx-Latn"> that is present: in other words, it is filled in only for the sake of completeness of metadata, by analogy with
7. all files may additionally contain <title type="alt"> for representing additional information in free-text
8. no markup is allowed on the contents of any of the <title> elements

Please give your thumbs-up so that Axelle can implement these conclusions.

[danbalogh]

Fine by me, except that I don't understand why it is important for the editorial title to include a mention of the name of a known author. I assume that the <author> will continue to be encoded, and to participate in search and perhaps sorting, right?

[arlogriffiths]

Thanks. Yes, <author> will continue to be encoded and it would be nice if this metadata could participate in searc and perhaps sorting. Inclusion of any author's name in <title type="editorial"> is important for the same reason that a title page of an edition of any non-anonymous work will feel empty if its author's name is not displayed. In other words, requesting its inclusion is a way to avoid needing to represent <author> as such in display.

[michaelnmmeyer]

Sounds good to me.

[arlogriffiths]

Thanks. So this is now implemented in the EGC.

The EGC template for <titleStmt> now reads as follows:

<titleStmt>
<title type="editorial"><!-- Title given by the editor to this digital edition--></title>
                <!--change the title of the edition starting, if possible, with the title of your text-->
		<!-- include a translation of the title and the author's name if relevant; form a subtitle after a colon -->
<!--<title type="alt"></title>-->
                <!-- fill in any alternative title if necessary; it may also form a comment on the editorial title -->
<!--<author></author>-->
                <!-- fill in the name of the author if relevant; repeat author is there is more than one -->
<respStmt>
                    <!-- repeat the respStmt when necessary; you can also provided a free text explanation of the process in the bibliography part-->
         <resp>author of digital edition</resp>
                    <!-- work done in the file, e.g. editor, intellectual authorship of edition ...-->
                    <persName ref="part:jodo"><!-- update the id -->
                        <forename>John</forename><!-- fill in your  names — it is possible to use only <name> instead of <foreame> and <surname> -->
                        <surname>Doe</surname>
                    </persName>
</respStmt>
</titleStmt>

@ajaniak : I have modified it after our discussion this morning.

[arlogriffiths]

Oops, now with <title type="original">:

<titleStmt>
                <title type="editorial"><!-- Title given by the editor to this digital edition--></title>
                <!--change the title of the edition starting, if possible, with the title of your text-->
		<!-- include a translation of the title and the author's name if relevant; form a subtitle after a colon -->
                <!--<title type="alt"></title>-->
		    <!-- fill in any alternative title if necessary; it may also form a comment on the editorial title -->
		<!--<title type="original" xml:lang="...-Latn"></title>-->
                    <!-- fill in the Romanized original title if relevant, and apply the relevant three-letter ISO language code -->
                <!--<author></author>-->
                <!-- fill in the name of the author if relevant; repeat author is there is more than one -->
                <respStmt>
                    <!-- repeat the respStmt when necessary; you can also provided a free text explanation of the process in the bibliography part-->
                    <resp>author of digital edition</resp>
                    <!-- work done in the file, e.g. editor, intellectual authorship of edition ...-->
                    <persName ref="part:jodo"><!-- update the id -->
                        <forename>John</forename><!-- fill in your  names — it is possible to use only <name> instead of <foreame> and <surname> -->
                        <surname>Doe</surname>
                    </persName>
                </respStmt>
            </titleStmt>