[tei-council] [TEI-L] SV: TEI MIME type (fwd)

[Martin Holmes]

I hope it's OK to post even though I'm not officially on Council till 
January. Smack me down if it's not.

The W3C document would seem to strongly support the creation of a TEI 
MIME type:

"In general, a representation provider SHOULD NOT assign Internet media 
types beginning with "text/" to XML representations."
<http://www.w3.org/TR/webarch/#xml-media-types>

This is partly to protect documents against transcoding by "Web 
intermediaries".

According to the RFC, it seems that the decision should depend on 
whether the XML is "readable by casual users":

"If an XML document -- that is, the unprocessed, source XML document
    -- is readable by casual users, text/xml is preferable to
    application/xml.  MIME user agents (and web user agents) that do not
    have explicit support for text/xml will treat it as text/plain, for
    example, by displaying the XML MIME entity as plain text.
    Application/xml is preferable when the XML MIME entity is unreadable
    by casual users."
<ftp://ftp.rfc-editor.org/in-notes/rfc3023.txt>

This seems to me a completely unhelpful distinction -- what does 
"readable" mean, and how would you define a "casual user"?

This contra argument on the feature request page:

"I think its especially complicated for us because we have no single
schema, but a series of application profiles."

would surely also apply to (for instance) SVG, which has SVG 1.0, SVG 
1.1, SVG 1.2, SVG Tiny 1.2, SVG Print, SVG Filters, and so on. A mime 
type is associated with a family of mutable recommendations rather than 
with a specific schema or DTD. In the case of TEI, ordinary users create 
mutations in addition to the standards body, which is slightly unusual, 
but surely not a disqualifying factor.

There's an interesting parallel in MathML version 2:

"[RFC3023] assigns MathML the MIME type application/mathml+xml. [...]

In MathML 1.0, text/mathml was given as the suggested MIME type. This 
has been superceded by RFC3023."
<http://www.w3.org/TR/2003/REC-MathML2-20031021/chapter7.html>

In other words, where MathML was previously text/mathml, the W3C seems 
to feel it's preferable to make it application/mathml+xml. If this is 
the general trend, then I think TEI should probably follow it.

One particular use-case might be this: If we have an application/tei+xml 
mime type, and web browsers are aware of it, an extension could easily 
be written for (say) Firefox which invokes or offers a conversion using 
Sebastian's stylesheets, or any other user-specified stylesheets. That 
would enable us to browse and read TEI documents without an intermediary 
web application to render them. That, in turn, would make it more likely 
that our own community would provide online access to their XML documents.

Finally, I can't actually think of any harm that it would do. Can anyone?

Cheers,
Martin

David Sewell wrote:
> Technically our Feature Request is still open:
> 
> http://sourceforge.net/tracker/index.php?func=detail&aid=2856036&group_id=106328&atid=644065
> 
> Sigfrid made a stronger case in his recent email to TEI-L for
> registering application/tei+xml than he had previously. Do we want to
> revisit our decision not to pursue registration?
> 
> If so, I would suggest that we ask Sigfrid to supply us with the
> information he thinks we would need to complete the IANA application (as
> he seems to be familiar with the process), namely
> 
> http://www.iana.org/cgi-bin/mediatypes.pl
> 
> and then we could decide who to authorize as the "Author" of the
> application.
> 
> David
> 
> ---------- Forwarded message ----------
> Date: Sun, 22 Nov 2009 16:05:26 +0100
> From: Sigfrid Lundberg <slu at KB.DK>
> To: TEI-L at listserv.brown.edu
> Subject: [TEI-L] SV: TEI MIME type
> 
> The docbook mime type is documented in an internet draft
> 
> http://tools.ietf.org/html/draft-walsh-app-docbook-xml-00
> 
> it expired a year or so ago. Don't know why it wasn't renewed. Perhaps the work with XProc and Calabash too much time.
> 
> In general, an XML application is a document format defined using XML as opposed to applications that are processing XML documents which are xml processors. In my view TEI is an XML application. Mime types for XML applications are described in RFC3023 (http://tools.ietf.org/html/rfc3023), which clearly states that mime types like
> 
> application/mathml+xml
> 
> is the preferred form. There are also a discussion of this in the recommendation from the W3C "Architecture of the World Wide Web" (http://www.w3.org/TR/webarch/#xml-media-types)
> 
> These practices are now very much integrated in the REST design patterns for web applications. I'd like to take a few examples:
> 
> One could imagine that someone would like to publish TEI documents using the atom publishing protocol. That is explicitely prohibited for other text/* mime types except text/html. XML documents published using this protocol must be application/*+xml.
> 
> Consider the situation that we have the same content available through a REST web service, where the representation is chosen through the accept http header, then what if there are more than one XML representation, should TEI then use text/xml and docbook application/docbook+xml? Well, one could live with that, but harvesters that where not prepared for this in advance couldn't.
> 
> The TEI community is really good at XML, but a registered applicaiton/tei+xml would demonstrate to the world that we regard ourselves as a part of the Wordwide web as well.
> 
> Yours
> 
> Sigfrid
> 
> 
> 
> 
> 
> ________________________________________
> Fra: TEI (Text Encoding Initiative) public discussion list [TEI-L at listserv.brown.edu] P&#229; vegne af Lou Burnard [lou.burnard at OUCS.OX.AC.UK]
> Sendt: 21. november 2009 18:34
> Til: TEI-L at listserv.brown.edu
> Emne: Re: TEI MIME type
> 
> Council discussed this recently, and reached the conclusion that there
> was no advantage (or indeed much sense) in registering a TEI-specific
> MIME type. It's XML all the way down; not an application as such.
> 
> Piotr Bański wrote:
>> Is there one?
>>
>> Docbook claims to have "application/docbook+xml", but I can't find it at
>>  IANA:
>>
>> http://www.iana.org/assignments/media-types/application/
>>
>> So I'm not sure whether the absence of application/tei+xml or
>> text/tei+xml or whatever ("tei-p5+xml"?) at IANA means that the TEI's
>> type is not registered or maybe merely not listed there.
>>
>> Thanks,
>>
>>   Piotr

-- 
Martin Holmes
University of Victoria Humanities Computing and Media Centre
(mholmes at uvic.ca)
Half-Baked Software, Inc.
(mholmes at halfbakedsoftware.com)
martin at mholmes.com