search

benetech / mde

Math Description Engine - The core of MathTrax. MathTrax is a type of graphing calculator for blind and vision impaired students.
http://math.bookshare.org
3 stars 1 forks source link

Create a delimited feature description template #10

Open TerryHodgson opened 12 years ago

TerryHodgson commented 12 years ago

Currently, mde returns its text descriptions in a block of natural language text that is a concatenation of feature descriptions. Create a prototype template that returns delimited feature descriptions instead so that the client application can have more flexibility in display of the description(s).

TerryHodgson commented 12 years ago

Perhaps the delimiter should be for compatibility with SVG, or maybe for compatibility with Daisy? What do you think?

TerryHodgson commented 12 years ago

Need to work with Benetech/other potential clients to determine best scheme for delimiting features. Experimented with using html headers as delimiters. The thing to remember when choosing is that the delimiters should not appear in a request for "text" output. They should appear in a request for "xml" output. "html" output may or may not include delimiters, depending on whether html is chosen as the delimiter. The addition of a delimiter should be done such that html code within the descriptions (and there's not much at this stage) is not broken.

TerryHodgson commented 12 years ago

email thread discussing implementation options (hmm, the xml and html samples are rendering in this comment, and I can't seem to get them to show up as plain text - well, contact me if you want to see it):

Yeah, I would use class instead of id and let the user parsing the results set ids as needed. A heading followed by a paragraph with the description seems to make more sense.

Meghan

From: Terry Hodgson terry.r.hodgson@gmail.com Date: Friday, September 21, 2012 11:02 AM Subject: Re: delimited results

Good points. Now that you mention it, I think a heading with the feature name followed by the feature description would be better than what I currently have.

I chose h1 because then screenreader users can easily jump to a particular feature if they want to hear it decribed again, rather than reading through all the descriptions.

As for applications that want more control than the default html, I believe they could still parse based on that delimiter.

However, it was just one idea. I'm open to others. ;-)

Terry

On Fri, Sep 21, 2012 at 12:02 AM, Anh Bui wrote:

Hi Terry,

I'd love to get Meghan's opinion on it as well. My main concern is about those id values. What happens when someone tries to insert multiple descriptions into a single document? Don't ids have to be unique? Also,if they are being used semantically is there another another attribute that would be more appropriate? Finally, I'm not sure about the use of

as opposed to something more like

. What drew you to

?

Thanks!

Anh

From: Terry Hodgson terry.r.hodgson@gmail.com Date: Thursday, September 20, 2012 3:13 PM Subject: delimited results

Hi again. What do you think about this approach to delimiting the feature descriptions?

I use a simple html header with an id.

With mde output format set to "xml" (delimited) it looks like:

<?xml version="1.0" encoding="UTF-8"?>

Cartesian.

The X axis is labeled x.

The Y axis is labeled y.

x = -10.0 to 10.0 and y = -10.0 to 10.0.

Equation(s) provided: y = 1.0*x +1.0.

a line.

It rises from left to right.

and with it set to "text" (not delimited):

Cartesian. The X axis is labeled x. The Y axis is labeled y. x = -10.0 to 10.0 and y = -10.0 to 10.0. Equation(s) provided: y = 1.0*x +1.0. a line. It rises from left to right.

and with it set to "html" (auto rendering delimited)

Cartesian.

The X axis is labeled x.

The Y axis is labeled y.

x = -10.0 to 10.0 and y = -10.0 to 10.0.

Equation(s) provided: y = 1.0*x +1.0.

a line.

It rises from left to right.