2007-114 MinutesPDF-UA 2007-114 April 18-19, 2007 PDF/UA Working Group Teleconference PDF/UA (Universal Accessibility) Working Group Meeting Agenda Wednesday, April 18, 2007, 9:00 A.M. - 5:00 P.M. EST Thursday, April 19, 2007, 8:30 A.M. - 12:30 P.M. EST Boston Convention Center, Room 207
Conference Call Information International Dial-in: +1-303-928-2600 U.S. Dial-in: 1-888-742-8686 Conference ID: 7083945 NOTE: The meeting was conducted "live" on the Wiki. Day 1: Wednesday, April 18 Roll Call- Duff Johnson - DJ, Chair, Doc Solutions
- Nobuichi Aoki - NA, Ricoh
- Cherrie Ekholm - CE, Microsoft
- Mark Gavin - MG, Appligent
- Dick Herring - DH, Usable, Accessible & Critical
- Karen McCall - KM, Karlen Communications (phone particpant)
- Greg Pisocky - GP, Adobe Systems
- Leonard Rosenthol - LR, Adobe Systems
- Neil Soiffer - NS, Design Science
- Cassandra Wilson - CW, HW Wilson Publishing
Approval of AgendaGavin / Soiffer No objections Rosenthol / Soiffer No objections Welcome New Participants- Cassandra Wilson - HW Wilson Publishing - Many questions concerning the accessibility of PDF based educational texts arise.
- Cherrie Ekholm - Microsoft - Introduced to the working group by Karen McCall. Cherrie has been working on the accessibility of PDF produced from Microsoft products.
Action Items will be updated by DJ later during the face to face Framework is the element of the specification into which the various modules fit. Sets the stage for additional issues in a PDF file if and when they occur. Provides users of the document an easy way to refer to items within their particular area of concern. DJ read the Framework outline. LR: I would like to recommend that we switch from Adobe PDF Reference PDF 1.6 to PDF Reference 1.7. Adobe PDF Reference 1.7 will be the source for the PDF version that will be submitted to ISO for formal adoption. Adobe will be providing AIIM with a properly ISO formatted version (ISO 32000) of the PDF Reference Manual 1.7 MG: Could you go over the differences between the two? LR: Differences between PDF 1.6 and 1.7. - Additions to tagged PDF, Ability to identify roles of form fields in non-interactive PDF documents. In essence non interactive PDF forms will behave as if they are "read only" PDF form fields.
- Ability to provide table summaries associated with table structures.
- Ability to identify background page artifacts.
- Ability to differentiate pagination artifacts: watermarks, headers and footers.
NS: Was anything done to change other elements? LR: Added that any string used in a PDF file can use UTF-8 (Unicode). LR: Other items, improvements in annotation, printing, 3D GP: Is the item an explanation of the TaggeePDF representation of nested table of contesnts entries or list entries also an improvement? LR: Not a change to PDF functionality, improved description on how one does nested table of contents entries or list entries. DJ: You say 1.7 is being ISOized. Does that mean we will have to go through another examination again. NS: You are not submitting another PDF spec? LR: We are not. Adobe is submitting a standard that is technically unmodified. This committee need not wait for ISO 32000 to be adopted in order to proceed with its work on PDF/UA. As an example, PDF/A 2 is using PDF Reference 1.7 as well. NS: I am generally in favor of staying current with what is the most recent standard. Proposal: That PDF/UA adopt PDF 1.7 as the reference for the PDF/UA Specification. Rosenthol / Soiffer Approved unanimously by those in attendance. KM: Document requires a base language should be in the Framework.
DJ: I recall that it was placed in the text section, because of situations that Joe Clark identified in Canada for instance. KM: I only bring this up because language is one of the parameters you can choose. It seems important that in document properties the document have a base language. LR: In the absence of a tag, treat it as this... DJ: What do we want to say about Base Language at this point? Right now the framework says every document has to define a base language. NS: How does that interact with the text module? DJ: Natural language shall be declared as in section Section 10.8.1 and/or as described in Section 3.8.1. Changes in natural language shall be declared. Language less documents shall not declare a language. KM: Instead of saying base language say default language. DJ: What about the Framework? LR: I have another item for within the Framework. DH: Statement of conformance, are we clear about what that means? MG: I would model it after PDF/A's statement of conformance. LR: Flickering elements are prohibited. MG: There are requirement boundaries that need to be clarified. DJ: I just added, note boundaries to the flickering requirement. We need to be more specific about that. MG: Copy and paste the definition of the Reader and Writer from other specs. LR: Yes, that would work. DJ Updated the Framework accordingly. MG: Start out by making it more of a specification, there's an introduction, then definitions, etc. NS: Notion of a whole spec and the framework and then the individual modules for it. MG: Framework is a misnomer. It needs to be removed and retitled, "Introduction." DJ made the change. LR: There are a standard list of required components for an ISO standard: Forward, introduction, scope, normative references, and if you like terms and definitions. DJ made the changes. BreakDJ recommended modifying the agenda to beat the lunch time crowd. Begin the discussion, break MG: Does everyone know what annotations are in PDF? Annotations are things that are not in the PDF file typically, They are layers on top of the PDF. Form fields, post it notes, digital signatures, dimensions on a CAD drawing, multi-media. DJ: It's a huge wide topic. Things that can reside "on top of" the PDF page. We have listed a numbe of annotation types in the list of modules. Comments, links, notes, movies. We have put something generic regarding Lunch BreakAnnotation discussion has resumed. LR has updated the Annotation General Section. Included entries for the various types of annotations Annotations Markup BreakNS presented and demonstrated authoring and viewing math in PDF. Math should be spoken with speech chosen by user which is a better approach than using alt text. The amount of experience and frequency of use a user has will determine the verbosity. There are also local customs on how math is spoken. Other important considerations: Magnification, Synchronized highlighting, Navigation, and Braille. There is a separate Braille math code that is different from the literary Braille code. Also there is more than one Braille code per county / language. NS is demonstrating examples of Math in a document. Word + MathType to author. Acrobat + MathPlayer to view. Embedding Math in PDF- All math should be tagged with formula
- All non lineear math should be tagged with MathML
- All characters inside of formula should mapt to Unicode
2 problem tags: <semantics> and <maction> LR: Proposal for a solution to the Problem Tags, Create a new property type to solve the issue of the problem tags. Recommend use existing marked content operator to point to a property which is a stream of the XML data. Another problem is large characters. LR: Tag called the actual text tag for a series of characters that you wish to be treated as some other entity. Hyphenated word is an example. DH: Does this work for noncontiguous content? MG: That's what I was going to ask. DH: Could only the beginning and end tags be marked? The solution is actual text. Refer to Section 10.8.3 Replacement Text in the PDF Reference 1.7 Action Item: NS write an item relating to SVG and accessibility. Break Topic 4: Programmatic Validation A Review of the role of programmatic validation in PDF/UA took place in the previous meeting of the working group and is noted in the minutes. DJ: Discussion of a status flag to indicate whether a PDF is in an accessible workflow and its progress in that workflow. LR: 2 levels of programmatic validation. Some class of items can be determined programmatically. DJ: What we said last time. 1) There would be a validator approach for each module which would discuss a testing approach for each module. Another notion is that of conformance level. LR: Is the issue conformance levels around readers vs. conformance levels for documents. GP: Conformance levels provide us with a way to get useful accessible content published without waiting for perfection. DH: We're talking about conformance levels and we are also talking about programmatic validation. We can at least determine which modules can be programatically validated and note. The section on Math will be used to illustrate how to document the validation of a framework item. Wrapup, revise plan for next day Adlourned at 5:05 pm 5:30 pm: PDF/UA DinnerDay 2Thursday, April 19Roll Call Day 2Duff Johnson - DJ, Chair, Doc Solutions Nobuichi Aoki - NA, RicohCherrie Ekholm - CE, MicrosoftFerass Elrayes - FE, Netcentric Technologies (phone participant)Betsy Fanninbg - BF, AIIMMark Gavin - MG, AppligentDick Herring - DH, Usable, Accessible & CriticalKaren McCall - KM, Karlen Communications (phone participant)Greg Pisocky - GP, AdobeNeil Soiffer - NS, Design ScienceYoshiyuki Sorimachi - YS, Fuji XeroxReview previous day's work, add Action Items, establish schedule until the next F2F meeting May 2, 16, 30, June 13, 27, July 18, August 1, 15, 29 Sept. 12, 26 Oct. 10 (Face to Face) DJ included the NIMAS description of table mark up in the Table module FE Noted notion of repeating headers not pertinent to PDF discussion due to relationship between tagging and layout in PDF. Solution to repeating headers could be to make the subsequent headers artifacts and add the subsequent cells to header tags attached to first table sheet. <Thead> can be repeated on each page but not redundantly tagged (treat as an artifact). NS the table constructs in the section should be tested. NS demoed problem tables. FE would like to add a column to the test table called module that would indicate what the sample is indicating. DJ made the change. DJ Made the point that authors are free to tag beyond the alogorithm in order to make sense of the table. The obligation of the algorithm is to correctly resolve the table. Does the algorithm serve our purposes well for the majority of cases? RH suggests an Action Item: Compare PDF tables to OASIS table model. AI: NS will look into the CALS table model. Look into feasability of specifying a should statement regarding the tagging of tables. FE: Yes we should take advantage of the fact that the structure in PDF is separate from the presentation. AI: FE to describe how for example something that appears to be a column header actually functions as a row span header. BreakA number of the attributes in multi-media are not an attribute of PDF, but are a property of the multi-media format (i.e., captioning). KM: Multimedia objects in PDF should require alternative text that describes the multi-media object. It was determined a multi-media tag is needed to enable assistive technology to distinguish multi media items from other types of annotations. It was also determined that it is of importance to indicate the nature of the multi media object to assist users in determining what controls or multi media players will be at their disposal. Updates to other modules, other businessDJ updated the Action Items page. Upcoming Meeting Schedule, Summer / Fall 2007All times 3:00 pm to 5:00 pm Eastern with the exception of the face to face meetings. May 2, 16, 30, June 13, 27, July 18, August 1, 15, 29 Sept. 12, 26 Oct. 10 (Face to Face tentatively scheduled for Ottawa, Ontario, Canada) Wrapup and adjournBF: If the committee feels satisfied with the content and format of the standard, the AIIM Technical Editor will extract the content from the Wiki and turn it into a document. All: The committee feels the draft today is at a 4 on a scale of 10 for completeness. The committee's goal would be to use the October face to face as the opportunity to review the draft candidate. Action Item: DJ will add a status facility for the individual modules in order to gauge progress towards completion of the standard. Move to Adjourn 12:40 Eastern time, April 19, 2007 Gavin / Soiffer |