Attachment Graphic with Alt Text.docx added by kevinchao89 on 2012-04-29 15:16
Description:

Comment 1 by mdcurran on 2012-04-30 06:48
Seems pretty simple to support this. However a few questions:

• How do other screen readers announce the existance of the image. MS Word's alternativeText property on the inlineShape object for this example outputs:

Image starts.\nA man with a suit sitting at a desk with laptop.\nImage ends.

We would prefer not to have that junk at the start and end, though if this changes for particular MS Word versions, or with different locales this will be hard to remove.
If soeone could test this on MS Word 2010, and with possible also a non-english MS Word as well that would be very helpfull. Steps are a sfollows:

1. Move the review cursor to the image (/ character).

2. Open the Python console with NVDA+control+z

3. type (and press enter):

   review.expand("character")
   

4. type (and press enter):

review._rangeObj.inlineShapes.item(1).alternativeText

You should then see the alt text for the image.

Comment 2 by orcauser on 2012-04-30 09:23
Please also see closely related ticket #1541
Thanks.

Comment 3 by jhomme (in reply to comment 1) on 2012-04-30 11:01
Replying to mdcurran:

Hi,
The other screen reader I use does not indicate that alternative text in Word 2007 is actually from a graphic, if alternative text exists. Unless you assume that the image description is a sentence or paragraph, you can't absolutely know for sure where it ends. You have to assume that the context will tell you that. On the web, though, it tells you that something is an image before it speaks the alt text. As far as I can remember, no screen reader tells you where an image description ends with some sort of end tag. And all of the Word accessibility advice I've read says that the user should indicate that it's a graphic in their alt text. So perhaps the start and end tag would be superfluous.

========= end comment ==========

Seems pretty simple to support this. However a few questions:

• How do other screen readers announce the existance of the image. MS Word's alternativeText property on the inlineShape object for this example outputs:

Image starts.\nA man with a suit sitting at a desk with laptop.\nImage ends.

We would prefer not to have that junk at the start and end, though if this changes for particular MS Word versions, or with different locales this will be hard to remove.

If soeone could test this on MS Word 2010, and with possible also a non-english MS Word as well that would be very helpfull. Steps are a sfollows:

1. Move the review cursor to the image (/ character).

2. Open the Python console with NVDA+control+z

3. type (and press enter):

   review.expand("character")
   

4. type (and press enter):

review._rangeObj.inlineShapes.item(1).alternativeText

You should then see the alt text for the image.

Comment 4 by kevinchao89 on 2012-05-01 20:29
The image starts and image ends are part of the alt text in this specific Word docx. Most alt text will not, nor should they have these start/end tags included. Of the three screen readers I tested: NVDA, System Access, and JAWS; only the ladder, JAWS is able to read the alt text.
In Word 2010 (2007 is reported to not say "description:" but i do not have access to it" with JAWS 12:
"description:Image starts. man with a suit sitting at a desk with laptop.\nImage ends. 6.46 inches wide by 7.49 inches high"

In Firefox/HTML and Reader/PDF, screen readers, like NVDA will say "graphic or image" before/after reading a graphic/image, but that is not the case in Word, which is why image start/end was added.

Comment 5 by mdcurran on 2012-05-02 10:25
Implemented in 80af921. I have tested with Word 2003 and 2007. Please test with Word 2010 and close if fixed.
Changes:
Milestone changed from None to 2012.2

Comment 6 by mdcurran on 2012-05-02 10:32
Should also be noted that dimensions will not be reported for graphics as this info can easily be viewed from the Picture formatting dialog found on the context menu of any graphic. As far as reading is concerned, graphic dimensions do not play a mandatory part of communication.

Comment 7 by kevinchao89 on 2012-05-02 17:25
Excellent! Thanks, confirmed as fixed in Main,5155. It's excellent that NVDA now handles alt text properly and does so better than any other screen reader I have tried. It reads "graphic followed by alt text", which means the image start and end image can be removed, which I did to test, and it works very well. As a result, we have modified our production work flow, specifically for alt text within Word to only include alt text and image start/end will be removed.
Changes:
State: closed