gmane.text.xml.relaxng.discuss

New validator

Dave Pawson — 2012-04-19T05:03:56

Copied from xml-dev

The purpose of the Serene project is to create an open source
validation engine implementing the JAXP 1.3 Validation Framework API
for RELAX NG, that concentrates on good messages and a clear handling
of ambiguity and conflicts.

For this release, Serene has been extensively reviewed and refactored
in order to improve performance. Some bugs were fixed, messages were
slightly tweaked, one feature was added.

For more information, please come to
 http://code.google.com/p/serene/

Thank you,
Radu Cernuta

Re: Is SVG 1.1 RNG schema valid?

Michael[tm] Smith — 2012-04-13T11:46:54

Murata-san,

OK, thanks. Now I'll try to figure why jing isn't handling it correctly...

  --Mike

MURATA Makoto <EB2M-MRT-uxaRg46wHOZxShC0m9Ra0A< at >public.gmane.org>, 2012-04-13 17:08 +0900:

Re: Is SVG 1.1 RNG schema valid?

MURATA Makoto — 2012-04-13T08:08:57

Michael[tm],

I  prefer A, since it allows further customization by another
define having combine="interleave".  But A and B are identical
if there is no further customization.

Regards,
Makoto

2012/4/13 Michael[tm] Smith <mike-Pl0VvzL1eo4< at >public.gmane.org>:

Re: odd result in trang-generated RNG

Eric van der Vlist — 2012-04-13T07:36:01

Hi,

You can't generate schemas from XML instances without heavily relying on
design conventions such as the one mentioned by James (and many others,
implicit or explicit).

An option to give users more control over how the schema is generated is
to add annotations to the instances: that's what I have proposed with
Examplotron (http://examplotron.org/).

A more elaborated (and complex) mechanism has also be proposed by OASIS
CAM (http://www.oasis-open.org/committees/tc_home.php?wg_abbrev=cam)
with their "templates".

Eric

Le vendredi 13 avril 2012 à 13:05 +0700, James Clark a écrit :

Re: odd result in trang-generated RNG

James Clark — 2012-04-13T06:05:56

IIRC, Trang aims to generate very simply schemas, in particular ones that
can be expressed exactly as XML DTDs, so it won't ever generate schemas
with &.

James

On Thu, Apr 12, 2012 at 10:45 PM, Terry Catapano <thc4ster-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

Is SVG 1.1 RNG schema valid?

Michael[tm] Smith — 2012-04-13T05:55:28

http://www.w3.org/Graphics/SVG/1.1/rng/svg-basic-font.rng has the following
(if you prefer to read it in RNC, see the RNC equivalents at the end)

  <define name="SVG.font-face.extra.class">
    <notAllowed/>
  </define>

  <define name="SVG.font-face.content">
    <zeroOrMore>
      <ref name="SVG.Description.class"/>
    </zeroOrMore>
    <optional>
      <ref name="font-face-src"/>
    </optional>
    <ref name="SVG.font-face.extra.class"/>
  </define>

So SVG.font-face.extra.class is defined as notAllowed at that point.

Then, http://www.w3.org/Graphics/SVG/1.1/rng/svg-font.rng has the following
(let's call it A):

  <include href="svg-basic-font.rng"/>

  <define name="SVG.font-face.extra.class" combine="interleave">
    <optional>
      <ref name="definition-src"/>
    </optional>
  </define>

Is that valid? Or should it instead be the following (let's call it B):

  <include href="svg-basic-font.rng">
    <define name="SVG.font-face.extra.class" combine="interleave">
      <optional>
        <ref name="definition-src"/>
      </optional>
    </define>
  </include>

I ask because if I keep it as in A above, it causes jing to throw an NPE
(which is a bug in jing regardless, and I already raised an issue for it:
http://code.google.com/p/jing-trang/issues/detail?id=156), but if I change
it to B, jing works as expected.

  --Mike

[1] Same bits above as RNC instead:

In file svg-basic-font.rnc:

  SVG.font-face.extra.class = notAllowed
  SVG.font-face.content =
      SVG.Description.class*,
      font-face-src?,
      SVG.font-face.extra.class

Then in svg-font.rnc:

(A)
  include "svg-basic-font.rnc"
  SVG.font-face.extra.class &= definition-src?

(B)
  include "svg-basic-font.rnc" {
    SVG.font-face.extra.class &= definition-src?
  }

Re: odd result in trang-generated RNG

Terry Catapano — 2012-04-12T21:11:51

John,

That does make things much clearer. I think the key concept is that "your
only guarantee is that all the documents will be valid", while I was
expecting that Trang would infer more from the set of instances, especially
with regard to required elements.

Thanks much,

Terry

On Thu, Apr 12, 2012 at 1:38 PM, John Cowan <cowan-PrmTNUR8zL8< at >public.gmane.org> wrote:

Re: odd result in trang-generated RNG

John Cowan — 2012-04-12T17:38:42

Terry Catapano scripsit:


The key to understanding your results is to realize that Trang creates
a model based on the first example it sees and then loosens it based on
more experience.

Your first document is abstractly (A, B), so the model is also (A, B).

Your second document is (A, B, B), so the model is loosened to (A, B+).

Your third document is (B, A), so the model must allow A and B to come
in any order, and Trang chooses to loosen it to (A | B)+.  It could go
with A & B+, but it doesn't.

Your fourth document adds no new requirements, so the model remains
the same.

(You can test these assertions by running Trang repeatedly with more
documents each time.)


Your only guarantee is that all the documents will be valid.  The more
different-looking documents you have, the looser the model will become,
as Trang never tightens the model.

(Disclaimer:  What I'm saying here is based on my experience, not from
examining Trang's code.  YMMV.)

odd result in trang-generated RNG

Terry Catapano — 2012-04-12T15:45:45

I'm using Trang to generate a RelaxNG schema from a set of XML
documents. I've run into a result I don't understand and which I hope
someone can explain.

Here are 4 xml documents:

case1.xml: One A element followed by one B element

<root>
  <A>xxxxx</A>
  <B>yyyyy</B>
</root>

case2.xml: One A element followed by two B elements

<root>
   <A>xxxxx</A>
   <B>yyyyy</B>
   <B>yyyyy</B>
</root>

case3.xml: One B element followed by one A element

<root>
   <B>yyyyy</B>
   <A>yyyyy</A>
</root>

case4.xml: Two B elements

<root>
   <B>yyyyy</B>
   <B>yyyyy</B>
</root>


Using these 4 documents as input, Trang (both version 20081028 and
20091111) generates the following RNG schema:

<grammar ns="" xmlns="http://relaxng.org/ns/structure/1.0"
datatypeLibrary="http://www.w3.org/2001/XMLSchema-datatypes">
 <start>
   <element name="root">
     <oneOrMore>
       <choice>
         <element name="A">
           <data type="NCName"/>
         </element>
         <element name="B">
           <data type="NCName"/>
         </element>
       </choice>
     </oneOrMore>
   </element>
 </start>
</grammar>

or in compact syntax:

default namespace = ""

start =
 element root {
   (element A { xsd:NCName }
    | element B { xsd:NCName })+
 }

What is surprising to me is that the resulting pattern does not
require a B element so that the following document:

<root>
  <A>xxxxx</A>
</root>

is valid, though I'd assume it would be invalid, given that a B
element is present in each of the input instances.

Is this a bug or mistake on Trang's part or something which is to be
understood as a reasonable result? I can imagine that inferring
patterns from a set of sample documents must be difficult going and
that what one gets from Trang might not be optimal. I would like to
understand it better, however, so I can adjust my expectations for
what I'll get from it and what additional work I might need to do to
use what it produces.

Thanks,

Terry

Re: CFP: TEI tutorials & workshops (Mon 11-05 to Wed11-07 < at > TAMU)

John Cowan — 2012-03-25T18:49:44

Syd Bauman scripsit:


If anyone wants to reuse my tutorial slides from various Extreme Markup
tutorials, they are available at:

http://ccil.org/~cowan/schelangs1.odp (Open Document Format)
http://ccil.org/~cowan/schelangs1.ppt (PowerPoint 97)
http://ccil.org/~cowan/schelangs1.pdf (PDF)

I used them for half-day non-hands-on tutorials to teach DTDs, RELAX NG
(using the compact syntax), and a little bit of Schematron.

They are available under the GNU GPL, so they can be reused and modified
freely for any purpose, but if you use them in a presentation, you must
make your modified slides available, at minimum to the attendees, but
preferably to everyone via the Web (I can host if necessary).

Please forward this as appropriate.

(By replacing 1 with 2 above, you get my XSD tutorial, which teaches
XSD in a compact syntax due to Kilian Stilhard, using the method of
differences -- that is, you must already know RELAX NG compact syntax.
The XSD syntax converter is available at
http://ccil.org/~cowan/xsconv-1.0.1.zip .)

CFP: TEI tutorials & workshops (Mon 11-05 to Wed 11-07 < at > TAMU)

Syd Bauman — 2012-03-25T15:41:57

 I think a 0.5 or 1.0 day hands-on tutorial on RELAX NG, or a 1.0 or
 1.5 day hands-on tutorial on Schematron would be quite apropos for
 this group. (What would be best, IMHO, is a 3-day hands-on tutorial
 on TEI customization that covers TEI, ODD, RELAX NG, and Schematron,
 but the call limits tutorials to 2 days.)  -- Syd

===================================================

Call for pre-conference workshop and tutorial proposals

TEI and the C(r|l)o(w|u)d
2012 Annual Conference and Members? Meeting of the TEI Consortium
Texas A&M University, Initiative for Digital Humanities, Media, and Culture

* Workshop proposals due Wed 15 May 2012
* Workshop dates: Mon 5 November to Wed 7 November, 2012
* Meeting dates: Wed 7 November to Sat 10 November, 2012 (see separate call)

The TEI Conference and Members' Meeting will be preceded by
educational tutorials or workshops. The goal of the tutorials is to
give an opportunity to learn more about the use of TEI markup under
the guidance of experienced instructors and practitioners, whereas
workshops are an opportunity for specific groups to meet and work
together on a TEI related subject.

Workshops and tutorials range in length from a single morning or
afternoon to a maximum of two days. Tutorials are run on a
cost-recovery basis: a separate fee is charged of participants that
is intended to cover the costs of running the tutorial. Workshops are
expected to be free of charges.

If you are interested in proposing either a workshop or a tutorial
for the 2012 Members? Meeting and Conference, please submit your
proposal as early as possible and before 15 May 2012 via conftool,
the availability of which will be announced shortly. Expressions of
interest should include as much as possible of the following
information (the committee is willing to work with proposers in
developing their proposals):
   * A proposed topic
   * A rationale explaining why this topic is likely to draw sufficient attention to the TEI community
   * Preferred length of the event
   * Infrastructural requirements
   * (In the case of a tutorial) A proposed instructor or slate of
     instructors including brief discussion of relevant experience,
     as well as a preliminary budget of your anticipated costs (if
     any).
   * (In the case of a workshop) A core list of people who are likely
     to participate, keeping in mind that workshops are by essence
     open for participation
Organisational and infrastructure costs (e.g. coffee breaks and the
like) will be determined later in conjunction with the local
organising committee.

Tutorial proposals will be evaluated by the programme committee
primarily on the basis of their likely appeal to the TEI community,
the quality of the proposed instructors and method of instruction,
and cost. The committee will work with selected organizers after this
date to refine the details of their proposals.

Please send queries to meeting-Ezmyqz1S860EbZ0PF+XxCw< at >public.gmane.org

For the International Programme Committee,

Elena Pierazzo (chair)
-------
Dr Elena Pierazzo
Lecturer in Digital Humanities
Chair of the Teaching Committee
Department of Digital Humanities
King's College London
26-29 Drury Lane
London WC2B 5RL

Phone: 0207-848-1949
Fax: 0207-848-2980
elena.pierazzo-3vygKqVCdQxaa/9Udqfwiw< at >public.gmane.org
http://www.kcl.ac.uk/ddh

CFP: TEI 2012 (Wed 11-07 to Sat 11-10 < at > TAMU)

Syd Bauman — 2012-03-22T02:59:37

Call for papers and proposals
==== === ====== === =========

TEI and the C(r|l)o(w|u)d
2012 Annual Conference and Members¹ Meeting of the TEI Consortium
Texas A&M University, Initiative for Digital Humanities, Media, and Culture

* Deadline for submissions: May 15, 2012
* Meeting dates: Wed 7 November to Sat 10 November, 2011
* Workshop dates: Mon 5 November to Wed 7 November, 2012 (see separate call)

The Programme Committee of the 2012 Annual Meeting of the Text
Encoding Initiative (TEI - http://www.tei-c.org/) Consortium invites
individual paper proposals, panel sessions, poster sessions, and tool
demonstrations particularly, but not exclusively, on digital texts,
scholarly editing or any topic that applies TEI to its research.

Submission Topics
---------- ------
Topics might include but are not restricted to:
- TEI and Google Books
- Handicraft vs. Large Scale Digitization: a False Dichotomy?
- TEI and massive digital collections
- TEI and Recording Document Corrections
- TEI and 'Dirty' OCR
- TEI Schemas and Document Publication History
- Text vs. Document: Can the TEI semantics express both?
- TEI and text corpora
- The relation between representation (encoded text) and presentation
  (visualisation, user-interface)
- TEI encoded data in the context of quantitative text analysis
- Integrating the TEI with other technologies and standards
- TEI as metadata standard
- TEI as interchange format: sharing, mapping, and migrating data (in
  particular in relation to other formats or software environments)

In addition, we are seeking proposals for 5 minute micropaper
presentations focused on experiences with the TEI guidelines gained
from running projects and discussing one specific feature.

Submission Types
---------- -----
Individual paper presentations will be allocated 30 minutes: 20
minutes for delivery, and 10 minutes for questions & answers.

Panel sessions will be allocated 1.5 hours and may be of varied
formats, including:

  * three paper-panels: 3 papers on the same or related topics

  * round table discussion: 5-8 presenters on a single theme. Ample
    time should be left for questions & answers after brief optional
    presentations.

Posters (including tool demonstrations) will be presented during the
poster session. The local organizer will provide flip charts and
tables for poster session/tool demonstration presenters, along with
wireless internet access. Each poster presenter is expected to
participate in a slam immediately preceding the poster session.

Micropapers will be allocated 5 minutes.

Submission Procedure
---------- ---------
All proposals should be submitted via conftool, the availability of which
will be announced shortly.  Please submit your proposals by May 15, 2012.

If you don¹t have already one, you will need to create an account (i.e.,
username and password) in order to file a submission. For each submission,
you may upload files to the system after you have completed filling out
demographic data and the abstract.

  * Individual paper or poster proposals (including tool
    demonstrations): Supporting materials (including graphics,
    multimedia, etc., or even a copy of the complete paper) may be
    uploaded after the initial abstract is submitted. Submission
    should be made in the form of an abstract of 750-1500 words (plus
    bibliography).

  * Micropaper: The procedure is the same as for an individual paper,
    however the abstract should be no more than 500 words. Please be
    sure the abstract mentions the TEI feature to be presented!

  * Panel sessions (three paper panels): The panel organizer submits
    a proposal for the entire session, containing a 500-word
    introduction explaining the overarching theme and rationale for
    the inclusion of the papers, together with a 750-1500 words
    section for each panel member.

  * Panel sessions (round table discussion): The panel organizer
    submits a proposal of 750-1500 words describing the rationale for
    the discussion and includes the list of panelists. Panelists need
    to be contacted by the panel organizer and have expressed their
    willingness in participation before submission.

Information
-----------
All proposals will be reviewed by the program committee and selected
external reviewers.

Those interested in holding working paper sessions outside the
meeting session tracks should contact the meeting organizers at
meeting-Ezmyqz1S863YtjvyW6yDsg< at >public.gmane.org to schedule a room.

Please send queries to meeting-Ezmyqz1S860EbZ0PF+XxCw< at >public.gmane.org

Conference submissions will be considered for conference proceedings,
edited as a special issue of the Journal of the Text Encoding
Initiative. Further details on the submission process will be
forthcoming.

For the International Programme Committee,

Elena Pierazzo (programme committee chair)

Re: jing + RELAX NG xsd:token on attributes: spaces coalesced?

Syd Bauman — 2012-02-27T19:26:50

Well, isn't that what xsd:token means? 
See http://www.w3.org/TR/xmlschema-2/#token

If you want internal whitespace preserved before comparison, I think
you should be using either "string" or "xsd:string".

jing + RELAX NG xsd:token on attributes: spaces coalesced?

Jason Sachs — 2012-02-27T19:09:57

I'm confused. If I use the schema attrbug.rnc (see below) it seems to
let through multiple spaces in a row, even though that shouldn't be
legal.

I'm using this pattern:
  xsd:token { pattern="[A-Za-z_][A-Za-z_0-9]*( [A-Za-z_][A-Za-z_0-9]*)*" }

Below is a test file that I'm using with "java -jar jing.jar -c
attrbug.rnc attrbug.xml".

What's going on here? Does jing coalesce spaces in attributes before
matching against xsd:token? Or am I missing something?

--Jason

--- begin file attrbug.xml ---
<?xml version="1.0" encoding="UTF-8"?>
<model>
  <item class="foo bar baz">we want a class attribute with multiple
    identifiers separated by single spaces.
  </item>
  <item class="foo    bar   baz">why does this not fail?</item>
  <item class="3 4 5">this fails as expected</item>
</model>
--- end file ---

--- begin file attrbug.rnc ---
# RELAX NG compact-syntax grammar

grammar {

# contains a model element
# the model element contains one or more item elements

start = element model { model-content }

model-content =
item+

item =
(element item { item-content })

item-content =
  attribute class { identifier-array },
  arbitrary-text-description

############### Text patterns ###############

arbitrary-text-description =
  text

identifier =
  xsd:token { pattern="[A-Za-z_][A-Za-z_0-9]*" }

identifier-array =
  xsd:token { pattern="[A-Za-z_][A-Za-z_0-9]*( [A-Za-z_][A-Za-z_0-9]*)*" }

}
--- end file ---

Re: Generating _1 = element * { notAllowed } any ideas why?

Alex Muir — 2012-02-17T17:44:29

Okay thanks very much,, created a little script to adjust that at the end
of the process.

Regards

On Fri, Feb 17, 2012 at 5:32 PM, Tara Athan <taraathan-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

Re: Generating _1 = element * { notAllowed } any ideas why?

Tara Athan — 2012-02-17T17:32:24

Re: Generating _1 = element * { notAllowed } any ideas why?

Alex Muir — 2012-02-17T10:27:05

Something I still don't understand is whether this is a solution that I
should use and what is it doing.

thanks

On Thu, Feb 16, 2012 at 2:24 PM, Alex Muir <alex.g.muir-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

Is this the final recommended fix for the problem?

Regards

Re: Generating _1 = element * { notAllowed } any ideas why?

Alex Muir — 2012-02-17T07:12:48

Yeah so what's involved to code this. Has anyone spent some time looking
into it?

On Thu, Feb 16, 2012 at 6:33 PM, Tara Athan <taraathan-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

Re: Generating _1 = element * { notAllowed } any ideas why?

Tara Athan — 2012-02-16T18:33:32

Re: Generating _1 = element * { notAllowed } any ideas why?

Alex Muir — 2012-02-16T15:19:27

Yeah as it turns out,,

_3 = element * { text }
_4 = element * { xsd:anyURI }
_5 =
  element * {
    "Collection"
    | "Dataset"
    | "Event"
    | "Image"
    | "MovingImage"
    | "StillImage"
    | "InteractiveResource"
    | "Service"
    | "Software"
    | "Sound"
    | "Text"
    | "PhysicalObject"
  }
_6 = element * { xsd:language }

_7 =
  element * - (ns2:* | ns3:*) {
    (attribute * { text }
     | text
     | _7)+
  }

I'm wondering why these _# names are created?

I don't find I have this text _1 & _1 ?

On Thu, Feb 16, 2012 at 2:56 PM, MURATA Makoto <EB2M-MRT-uxaRg46wHOZxShC0m9Ra0A< at >public.gmane.org>wrote:

Re: Generating _1 = element * { notAllowed } any ideas why?

MURATA Makoto — 2012-02-16T14:56:45

If this change does not remove the error message, do you have
another occurrence of "element *" elsewhere in your schema?
Or, do you have _1 & _1 ?

Regards,
Makoto
2012/2/16 Alex Muir <alex.g.muir-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org>