Hi Keith, This sounds good. However, it leaves me uncertain what to develop in the document package. I may work on liblouisutdml and liblouis until things are more settled. Meanwhile, if BrailleBlaster can open an xml file and call StartDocument, I can test what I have already. That is an essential step. On another subject, I think that the UTD class might be folded into the Braille view. It contains a lot of stuff that needs to be stripped out, including a reference to the <p> element, which is not part of UTDML. John On Thu, Jan 24, 2013 at 05:50:56PM +0000, Keith Creasy wrote: > Hi John. > > These are mostly notes from this mornings meetings. > > Yes, for the browser view we just about have to use XSLT. I know there are > transformations for DTBook and NIMAS. I'm not sure about others but they > really are not that hard to create if they don't exist. > > The thing we talked about this morning is how to map segments of text by > offset to the DOM text nodes. I think what we decided is that we do need a > map of some kind and it needs three things. > > The style of the text node. We plan to do a proof-of-concept implementation > with just inline and block. Later the semantic action file can provide this. > The index position of the text node in the DOM. > The range of character offsets this text node represents in the text view. > > When the XML is parsed, or after it is parsed by walking the DOM tree, we > build the map or list of text nodes with the style and the index of each text > node. > > To get the actual text, or braille for that matter, you can use XPath to > extract the text from the text node at a given index. > The style field of the item in the map tells the view how to format it on the > screen. > > The view adds the text offsets to the map for each text node that is > displayed as it does the rendering. > > When the text changes in the view the map is used to make those changes in > the DOM. The document (DocumentBase) makes the changes and notifies the views > through the DocumentManager to update. > > Inserting text (typing) is easy as we only need to change the content of the > currently referenced text node. > > Deleteing is also easy if the text we are deleting is contained in a single > text node. > > Things get a little more complicated if a node becomes empty, a two nodes are > merged by a delete action, or a new node is created by an insertion. > We also have to decide things like what happens when a text node becomes > empty. Do we delete the parent element or just leave an empty element? > > Chuck is, in addition to working on the tool bar, going to do some research > into whether there is an open-source widget that takes care of any of this. > He's also going to poke around for something that can format text based on a > style sheet (css). Either of these might be useful and shorten development > time. > > Brandon is going to put together a proof-of-concept application that attempts > to use the mapping scheme we talked about using the StyledTextView and an XOM > DOM (how cute) to see if this idea will work. If the concept works then we > are very close to having like a wisiwyg XML editor. The only restrictions is > that it won't use mixed fonts initially and is pretty simple in its > presentation. > > > > > > > > > Keith Creasy > Software Developer > American Printing House for the Blind > KCreasy@xxxxxxx > Phone: 502.895.2405 > Skype: keith537 > > > -----Original Message----- > From: brailleblaster-bounce@xxxxxxxxxxxxx > [mailto:brailleblaster-bounce@xxxxxxxxxxxxx] On Behalf Of John J. Boyer > Sent: Thursday, January 24, 2013 12:23 PM > To: brailleblaster@xxxxxxxxxxxxx > Subject: [brailleblaster] Re: What will the text view show? > > Hi Keith, > > Having the text view handle semantic-action files is an interesting idea. > It's a matter of what is considered the best way to encapsulate things. My > thought was that the document class would encapsulate knowledge about the > vocabulary of the document and the text view would deal with an abstraction > by calling methids in DocumentBase. > Encapsulating this knowledge in the text view is another possibility. > > For the future, how will the browser view handle various tyoes of documents? > I would think that an xsl stylesheet would be appropriate for xml documents. > Are these available for diferent document types? > > John > > On Thu, Jan 24, 2013 at 01:52:37PM +0000, Keith Creasy wrote: > > John. > > > > I view this as more of a mental exercise because as far as I know we've > > decided to use your model, but... > > > > Why couldn't the text view and the braille view use the semantic action > > file? If its purpose is to define how the text and braille get formatted in > > the view then the view is where it belongs I think. > > > > No matter how we do it mapping text in the text view to nodes in the XML > > document won't be simple. Yes, it will involve offsets and some kind of > > mapper to map the text to its corresponding text node. I can't think of any > > way around that. > > > > > > -----Original Message----- > > From: brailleblaster-bounce@xxxxxxxxxxxxx > > [mailto:brailleblaster-bounce@xxxxxxxxxxxxx] On Behalf Of John J. > > Boyer > > Sent: Thursday, January 24, 2013 8:38 AM > > To: brailleblaster@xxxxxxxxxxxxx > > Subject: [brailleblaster] Re: What will the text view show? > > > > Hi Keith, > > > > I'm trying to understand your approach. Wouldn't the text view have to know > > the vocabulary of the particular xml document? The Braille view will of > > course use the UTDML vocabulary. The tree view essentially shows the > > structure of the document, with element names, etc. > > > > On using XPath expressions to locate text nodes, the search for a text node > > must start with the cursor location in the StyledText control. How is this > > related to XPath? > > > > I can see that the text view might have different methods or embedded > > classes for diferent document types, such as Daisy, epub3, heml5, etc. > > This is similar to using semantic-action files, but more cumbbersome and > > less flexible. > > > > John > > > > On Wed, Jan 23, 2013 at 05:45:47PM +0000, Keith Creasy wrote: > > > OK, I'll try again to explain this. > > > > > > I've always thought the format of the text in the views should be > > > controlled from within the views and that the view should just use the > > > XML and do the formatting. Semantics might still come into play but > > > anything that controls how the text or braille gets rendered on the > > > medium, in this case the screen, gets done by the view or one of it's > > > member objects. > > > > > > > > > This keeps the view and the rendering process encapsulated. > > > > > > The procedure might look like this: > > > > > > 1. DocumentManager creates an instance of DocumentBase and the views > > > associated with the document. > > > 2. XML is opened and processed through LibLouisUTD to create the internal > > > document model. > > > 3. Once the document is parsed and the model is complete DocumentManager > > > informs the views to perform an initial update. > > > 4. Each view calls methods that access the document model to pull > > > content, in the form of XML that has the UTDML markup included. > > > 5. The view, or one of it's members, processes a segment of the XML in > > > order to render content. In the case of the braille view, for example, it > > > would get braille elements and create the formatted braille and add the > > > newly formatted content to the StyledContent object and the view would > > > display it. > > > > > > > > > In this way any view gets it's content in the form of XML and knows what > > > to do with it. > > > > > > Does it make any sense? Please ask questions if it isn't clear. > > > > > > Keith Creasy > > > Software Developer > > > American Printing House for the Blind KCreasy@xxxxxxx > > > Phone: 502.895.2405 > > > Skype: keith537 > > > > > > > > > -----Original Message----- > > > From: brailleblaster-bounce@xxxxxxxxxxxxx > > > [mailto:brailleblaster-bounce@xxxxxxxxxxxxx] On Behalf Of John J. > > > Boyer > > > Sent: Wednesday, January 23, 2013 12:22 PM > > > To: brailleblaster@xxxxxxxxxxxxx > > > Subject: [brailleblaster] Re: What will the text view show? > > > > > > Hi Keith, > > > > > > Let me hear your approach, so it's a two-way street. > > > > > > John > > > > > > On Wed, Jan 23, 2013 at 04:53:58PM +0000, Keith Creasy wrote: > > > > Hi John. > > > > > > > > Well, since I'd have used a different approach to most of this I guess > > > > you are the driver here. What would you like? We are all anxious to see > > > > how the views get the content they need to display and how it is linked > > > > back to the elements in the XML. > > > > > > > > > > > > We can even add an additional view if that's what folks want. > > > > > > > > Keith Creasy > > > > Software Developer > > > > American Printing House for the Blind KCreasy@xxxxxxx > > > > Phone: 502.895.2405 > > > > Skype: keith537 > > > > > > > > > > > > -----Original Message----- > > > > From: brailleblaster-bounce@xxxxxxxxxxxxx > > > > [mailto:brailleblaster-bounce@xxxxxxxxxxxxx] On Behalf Of John J. > > > > Boyer > > > > Sent: Wednesday, January 23, 2013 11:50 AM > > > > To: brailleblaster@xxxxxxxxxxxxx > > > > Subject: [brailleblaster] What will the text view show? > > > > > > > > If I remember right, in the original spec what is now called the text > > > > view would present an appearance like that of a simple word processor, > > > > with paragraphs, headings, tables, etc. Alternatively, it could be > > > > switched to show the original print for each line in the Braille view. > > > > I would like to know the present "views" on this matter, so I know what > > > > to develop in the document package. > > > > > > > > Thanks, > > > > John > > > > > > > > -- > > > > John J. Boyer; President, Chief Software Developer Abilitiessoft, Inc. > > > > http://www.abilitiessoft.com > > > > Madison, Wisconsin USA > > > > Developing software for people with disabilities > > > > > > > > > > > > > > -- > > > John J. Boyer; President, Chief Software Developer Abilitiessoft, Inc. > > > http://www.abilitiessoft.com > > > Madison, Wisconsin USA > > > Developing software for people with disabilities > > > > > > > > > > -- > > John J. Boyer; President, Chief Software Developer Abilitiessoft, Inc. > > http://www.abilitiessoft.com > > Madison, Wisconsin USA > > Developing software for people with disabilities > > > > > > -- > John J. Boyer; President, Chief Software Developer Abilitiessoft, Inc. > http://www.abilitiessoft.com > Madison, Wisconsin USA > Developing software for people with disabilities > > -- John J. Boyer; President, Chief Software Developer Abilitiessoft, Inc. http://www.abilitiessoft.com Madison, Wisconsin USA Developing software for people with disabilities