How do screen readers read HTML entities?
Home / Site Building & Testing / How do screen readers read HTML entities?
| dotjay wrote: | ||
I've run some tests with JAWS, Home Page Reader and Window-Eyes. My results can be found here on my site: Screen Readers and HTML Character Entities As I suspected, JAWS isn't all that bad, but doesn't know the entity for minus, so the best you can do I guess is use a dash I suppose. Also, I think there's a newer version of Window-Eyes that might do better than the version I tested with - perhaps I'll see at some point. |
Thanks Jon, that's very interesting.
As you say, hopefully a newer version of Window-Eyes will perform better. It's surprising to see that Home Page Reader does not not read out the dash in test 12. I suppose that attests to the difficulty of distinguishing between the various functions of the dash character.
I just tested the entities on your page with (a) the builtin browser in my Sony Ericsson W800i mobile, and (b) the Opera mini simulator. All of the characters displayed correctly except for:
Builtin browser
− displays as −
numeric equivalent displays as []
⁄ displays as ⁄
numeric equivalent displays as []
Opera mini
− displays as [empty square box]
numeric equivalent displays as [empty square box]
_________________
Nick
Share: 
| coplanar wrote: |
| Builtin browser
− displays as − numeric equivalent displays as [] ⁄ displays as ⁄ numeric equivalent displays as [] Opera mini − displays as [empty square box] numeric equivalent displays as [empty square box] |
This kind of shows why I decided to test both named and numeric references. Screen readers will more often than not sit on top of your browser, which should interpret them all the same way, but you can't rely on it.
I'll be adding tests with more browser/screen reader combinations as I get to them. I'll try latest version of Window-Eyes for starters I guess.
Cheers all.
_________________
Jon Gibbins :// blogs at dotjay.co.uk, works with Analog.
Share:
| eatyourgreens wrote: | ||
There's also the question of how to make charts, graphs and plots of scientific data accessible too. Longdesc might work for that, but some charts can contain an awful lot of information eg. the stellar spectra from the Sloan Digital Sky Survey on this tutorial: http://outreach.atnf.csiro.au/... |
Hi All,
Window-Eyes does not read these entities with its default settings. For Window-Eyes to read these entities, change the punctuation settings to enable it to read Text, Math and general punctuations.
Regards,
Priti
Accessibility Consultent
Net Systems Informatics (India) Pvt. Ltd.
Url: www.n-syst.com
Share:
| eatyourgreens wrote: |
| Just playing devils advocate, and pointing out that alt text for maths is difficult. |
There's a nice article about making dynamically generated graphs accessible at the Accessible Content magazine site. It's a short case study of work done by The National Cancer Institute, and while I'm not sure their approach would provide a solution for your stella spectra, it does work for reasonably complex data.
_________________
Dan Champion, Champion IS, Mooch Marketing, Revish
Share:
| priti wrote: |
| Window-Eyes does not read these entities with its default settings. For Window-Eyes to read these entities, change the punctuation settings to enable it to read Text, Math and general punctuations. |
Thanks, Priti. I'm a Window-Eyes newbie, so useful to know. I just can't get used to the interface in Window-Eyes!
_________________
Jon Gibbins :// blogs at dotjay.co.uk, works with Analog.
Share:
| danchamp wrote: | ||
There's a nice article about making dynamically generated graphs accessible at the Accessible Content magazine site. It's a short case study of work done by The National Cancer Institute, and while I'm not sure their approach would provide a solution for your stella spectra, it does work for reasonably complex data. |
That is interesting. I wonder how it scales as the number of data points gets larger? I imagine it's difficult to listen to all of the x and y points, for a large plot, and build a picture of what's going on. Naming the maximum and minimum points is a good idea, but those charts have unambiguous maxima and minima. In spectroscopy, you'll find that the exact position of the minima depend on the measurement technique that you apply to the data, but that could be explained in the main body of the paper. Similarly, if you fit curves to experimental data points, you'll probably find there are a choice of curves which fit the data and this is probably beyond automated description of the graph.
_________________
Jim O'Donnell
work: Royal Observatory Greenwich
play: eatyourgreens
Share:
| eatyourgreens wrote: |
| That is interesting. I wonder how it scales as the number of data points gets larger? I imagine it's difficult to listen to all of the x and y points, for a large plot, and build a picture of what's going on. Naming the maximum and minimum points is a good idea, but those charts have unambiguous maxima and minima. |
It's a useful starting point, and I'd imagine that listening to the data serially will allow you to build up a mental image of the trends. It could be improved though IMHO - having data to 4 decimal places as they do is probably not the best idea, and the data is currently presented as one long paragraph with points separated by line breaks.
I'm just finishing work on a dynamic service for some of our statistics (things like mid-year population estimates), and have been using tabular equivalents. The next step I'm considering is to use some heat-map styling on the tables (a bit like Jeff Veen's rainfall tables at @media) but the difficulty is how to convey that to non-visual browsers. These are simple bar graphs, so one possibility is to use hidden contextual information like distance from the mean for each value, which with standard deviation would provide enough contextual information to take a value in isolation and still get an idea of its significance.
| Quote: |
| In spectroscopy, you'll find that the exact position of the minima depend on the measurement technique that you apply to the data, but that could be explained in the main body of the paper. Similarly, if you fit curves to experimental data points, you'll probably find there are a choice of curves which fit the data and this is probably beyond automated description of the graph. |
Yeah, well, you fancy-pants astro-physicists never take the easy option, do you?
_________________
Dan Champion, Champion IS, Mooch Marketing, Revish
Share:
A blind astro-physicist might want the data to four decimal places for use in their research. However, a less specialised visitor might just want to know an general overview of the data and maybe some of the more interesting outliers?
You might give a longdesc description of the charts as you would a photograph: summarise the scene and pick out a few interesting details rather than create a table of the RGB colour values used for each pixel.
I guess the key thing is to understand how much data your visitors actually want and balance that against what is practical to access from a web browser?
_________________
My CV type thing and my Life of Ben (Blog). Nigel Peck's Accessify Forum Requirements.
Share:
Apologies for dragging this one up again, but I've just been reading up about Fire Vox, Charles Chen's speech extension for Firefox.
Fire Vox will read MathML back to users with no additional plugins to install. Fire Vox reads MathML using the "Mathspeak rules", a vocabulary developed for reading formulae to screen reader users.
Check out Mathspeak at:
http://www.mathspeak.org/
and
http://www.rit.edu/...
_________________
Jon Gibbins :// blogs at dotjay.co.uk, works with Analog.
Share:
Sorry to dredge this topic up.
I was curious how wirds such as "cliche" are read. Will just typing the word be enough or would I need to write "Cliché" Or would it be better to wrap either of those in <span lang="fr">
_________________
Jonathan Worent
The early bird gets the worm, but the second mouse gets the cheese.
Share:
| jw_developer wrote: |
| Sorry to dredge this topic up.
I was curious how wirds such as "cliche" are read. Will just typing the word be enough or would I need to write "Cliché" Or would it be better to wrap either of those in <span lang="fr"> |
I've not tested this (I may when I get time), but definitely wrap the word in a span with the lang attribute. I'm guessing that this should get JAWS to read that in French, but I'd say you'll probably need to use Cliché as well.
_________________
Jon Gibbins :// blogs at dotjay.co.uk, works with Analog.
Share:
| jw_developer wrote: |
|
I was curious how wirds such as "cliche" are read. Will just typing the word be enough or would I need to write "Cliché" Or would it be better to wrap either of those in <span lang="fr"> |
In this case, the word has been adopted into English and so doesn't require anything special.
According to my dictionary, the correct way to spell it is with an accent on the 'e' -- but it's commonly seen without.
Either way, a screen reader should have no difficulty in pronouncing the word correctly.
.
Share:
| Jake wrote: |
| In this case, the word has been adopted into English and so doesn't require anything special. |
Valid point - as "chiche" has been appropriated into the English language, it should be read correctly. But you can't always rely on "should", so I decided I'd test it quickly.
Testing with JAWS 7.10 (and Firefox 1.5)
Just the word without the accent: cliche
Read correctly, "klee-shey"
The word with the accent: cliché
Read incorrectly as "klich-ey" (like "itch a")
Without accent, wrapped in span and lang: <span lang="fr">cliche</span>
Read incorrectly as "kleesh" (like "quiche")
With accent, wrapped in span and lang <span lang="fr">cliché</span>
Read correctly, faster with better accent pronunciation as "klee-shey"
So, I recommend not using the accent if you are not going to define a lang attribute... It's a valid spelling even though it would be correct to use the accent.
But if you are going to use the accent, I guess that the accent is handled as a separate character unless you use the lang attribute to specify that it is French. And if it's French, you should, of course, use the accent.
_________________
Jon Gibbins :// blogs at dotjay.co.uk, works with Analog.
Share:
| dotjay wrote: |
|
Without accent, wrapped in span and lang: <span lang="fr">cliche</span> Read incorrectly as "kleesh" (like "quiche") |
Being nitpicky, but 'cliche' (no accent) is English, not French.
_________________
Jim O'Donnell
work: Royal Observatory Greenwich
play: eatyourgreens
Share:
| eatyourgreens wrote: | ||
Being nitpicky, but 'cliche' (no accent) is English, not French. |
Fair enough. But tested anyway!
_________________
Jon Gibbins :// blogs at dotjay.co.uk, works with Analog.
Share:
All times are GMT
You cannot post new topics in this forumYou cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum