Blockquotes?

Hi,
a new feature was added to FF - "q" and Shift+q
move between blockquotes. 
Fine, but - what for heaven sake are those
"blockquotes"?
Yesterday evening I was strolling through my bookmarks, and
I couldn't find 
a single page containing blockquotes. So my Question: What
are blockquotes 
and what is this feature for?
A similar thing are those famous "large objects".
Till today i wasn't able 
to figure out what it is all about. Moving between that
objects seems to 
result in stochastical behavior of Orca, e.g. you land
anywhere, and you 
cannot predict where.
So what about replacing those two feature by some more
common thins like 
moving between frames, textblocks >=x, buttons, edit
fields and - not so 
important - paragraphs.
There should also be a feature implemented that
speaks/brailles background 
information of an object; such a feature is imlemented in
Firevox.
Hermann
_______________________________________________
Orca-list mailing list
Orca-listgnome.org
http
://mail.gnome.org/mailman/listinfo/orca-list
Visit http://live.gnome.org/Orca
 for more information on Orca

Hi there,
For once, and hopefully not the only time i'm agreeing with
you. A
question though, what on earth is "background
information"? Is it font,
style, color etc? Isn't there a key implemented for this
already?
Jumping between frames is surely more important than
blockquotes i
think. Perhaps, but maybe that's what you suggested, a
feature to jump
to first non-linked text, would be great for skipping large
segments
with menu links etc. I think that's what the chunks feature
was meant to
do, and i have had no trouble with it, but that's me.
/Krister

Hermann wrote:
> Hi,
> a new feature was added to FF - "q" and
Shift+q move between blockquotes. 
> Fine, but - what for heaven sake are those
"blockquotes"?
> Yesterday evening I was strolling through my bookmarks,
and I couldn't find 
> a single page containing blockquotes. So my Question:
What are blockquotes 
> and what is this feature for?
> A similar thing are those famous "large
objects". Till today i wasn't able 
> to figure out what it is all about. Moving between that
objects seems to 
> result in stochastical behavior of Orca, e.g. you land
anywhere, and you 
> cannot predict where.
> So what about replacing those two feature by some more
common thins like 
> moving between frames, textblocks >=x, buttons, edit
fields and - not so 
> important - paragraphs.
> There should also be a feature implemented that
speaks/brailles background 
> information of an object; such a feature is imlemented
in Firevox.
> Hermann
> _______________________________________________
> Orca-list mailing list
> Orca-listgnome.org
> http
://mail.gnome.org/mailman/listinfo/orca-list
> Visit http://live.gnome.org/Orca
 for more information on Orca
_______________________________________________
Orca-list mailing list
Orca-listgnome.org
http
://mail.gnome.org/mailman/listinfo/orca-list
Visit http://live.gnome.org/Orca
 for more information on Orca

Hi Krister,
I know this blockquotes from Jaws and Window Eyes, but I
never ever used 
it.
Regarding "lage objects": I think it could be
rebuild to just moving to 
"large blocks of non-link text"; then it would be
useful. But actually you 
land sometimes on links, headings etc., and I couldn't
figure out why. 
Perhaps because they have long names...
Regarding frames: There are FF hotkeys for frames (F6 and
Shift+F6), but 
the problem the developers have to solve here is the
identification of the 
frames. This seems a bit tricky since they often are not
labeled properly. 
Maybe further accessibility development by the FF team is
needed, but I 
think it should be done as soon as possible.
Another item are anchors. Either there is a new keystroke
implemented, or, 
since they are links, pressing "u" and
"shift+u" should announce them as 
they appear. This variant is implemented in Firevox, it
announces "external 
link" and "internal link" when you come upon
a link.
Hermann
_______________________________________________
Orca-list mailing list
Orca-listgnome.org
http
://mail.gnome.org/mailman/listinfo/orca-list
Visit http://live.gnome.org/Orca
 for more information on Orca

In HTML the <blockquote> element should be used to
markup a long 
quotation from an external source.  The name comes from a
typographic 
convention where longer quotations are set in a new,
indented text block 
rather than inlined with the main text. For documentation,
see:

h
ttp://www.w3.org/TR/html4/struct/text.html#h-9.2.2

You can look up other HTML elements from the index at:

http
://www.w3.org/TR/html401/index/elements.html

Like all HTML elements, <blockquote> is subject to a
certain degree of 
misuse, classically to produce indented text in visual
browsers even 
when not quoting an external source. My impression is that
the rise of 
cascading stylesheets to add presentational hints has made
that misuse 
much less common. Conversely, (again like all HTML elements)

<blockquote> is frequently subject to people's failure
to use it when 
they should.

Properly used <blockquote> elements are rare on the
commercial and 
technical web, but common in the sort of material where you
would expect 
to find quotations, e.g. blogs, forums, academic documents,
HTML email. 
Some examples in the wild:

h
ttp://crookedtimber.org/2007/05/11/rupturerapture/

http://uk.promotions.yahoo.com/forgood/feat
ured-charities/joinup.html

http://etext.lib.virginia.edu/journals/EH/EH42/Swan42.
html

Now it may be that your bookmarks happen not to contain any
<blockquote> 
elements or it may that Firefox and Orca are malfunctioning
somehow. The 
only way to be sure is to inspect the DOM to see if any
<blockquote> 
elements are present.

The same q and shift + q key bindings are used by JAWS and
Window-Eyes 
to jump between <blockquote> instances:

http://www.freedomscientific.com/fs_products
/Surfs_Up/Quick_Keys.htm

http://www.gwmicro.com/Window-Eyes/Manu
al/HTML/19_13quotesblockquotes.htm

I would have thought the same key combination could be apply
to jumping 
between instances of the <q> element sometimes used
for short 
quotations. But without knowing more about how this
navigational aid is 
currently used in practice, it's hard to be sure.

A <blockquote> element may contain multiple text
blocks or paragraphs, 
and widget navigation (frames, buttons, fields) is another
matter 
entirely. It doesn't make sense to speak of replacing one
feature with 
other rather different features, unless they happen to be
competing for 
binding to the same key combinations. That's not replacement
of 
features, but removal of features.

There are already standard Firefox key bindings for moving
between 
frames (F6 and Shift + F6):

http:
//www.mozilla.org/support/firefox/keyboard

If there is no further frame, pressing F6 moves you on to
the address 
bar. So there is no need for Orca to duplicate this
feature.

With regard to your other feature suggestions, what defines
a 
"textblock" and what is "background
information"?

I'm not sure what "large objects" we're talking
about are though. When 
you say you cannot predict where you land, do you mean that
if you move 
to the next large object, then another large object, then
press the 
command to go back an large object, you do not arrive at the
first large 
object? Or do you just mean that you don't know where the
next large 
object will be? That seems to be an inherent aspect of
cycling between 
instances of any webpage component, unless you know the
webpage in 
question very well.

--
Benjamin Hawkes-Lewis

Krister Ekstrom wrote:
> Hi there,
> For once, and hopefully not the only time i'm agreeing
with you. A
> question though, what on earth is "background
information"? Is it font,
> style, color etc? Isn't there a key implemented for
this already?
> Jumping between frames is surely more important than
blockquotes i
> think. Perhaps, but maybe that's what you suggested, a
feature to jump
> to first non-linked text, would be great for skipping
large segments
> with menu links etc. I think that's what the chunks
feature was meant to
> do, and i have had no trouble with it, but that's me.
> /Krister
> 
> Hermann wrote:
>> Hi,
>> a new feature was added to FF - "q" and
Shift+q move between blockquotes. 
>> Fine, but - what for heaven sake are those
"blockquotes"?
>> Yesterday evening I was strolling through my
bookmarks, and I couldn't find 
>> a single page containing blockquotes. So my
Question: What are blockquotes 
>> and what is this feature for?
>> A similar thing are those famous "large
objects". Till today i wasn't able 
>> to figure out what it is all about. Moving between
that objects seems to 
>> result in stochastical behavior of Orca, e.g. you
land anywhere, and you 
>> cannot predict where.
>> So what about replacing those two feature by some
more common thins like 
>> moving between frames, textblocks >=x, buttons,
edit fields and - not so 
>> important - paragraphs.
>> There should also be a feature implemented that
speaks/brailles background 
>> information of an object; such a feature is
imlemented in Firevox.
>> Hermann
>> _______________________________________________
>> Orca-list mailing list
>> Orca-listgnome.org
>> http
://mail.gnome.org/mailman/listinfo/orca-list
>> Visit http://live.gnome.org/Orca
 for more information on Orca
> _______________________________________________
> Orca-list mailing list
> Orca-listgnome.org
> http
://mail.gnome.org/mailman/listinfo/orca-list
> Visit http://live.gnome.org/Orca
 for more information on Orca
> 

_______________________________________________
Orca-list mailing list
Orca-listgnome.org
http
://mail.gnome.org/mailman/listinfo/orca-list
Visit http://live.gnome.org/Orca
 for more information on Orca

Hi,
Hmm, ok now i think i see where you're coming from. The big
question is
should this announcement of external and internal links be
optional?
I agree that it can make sence with links and anchors, but i
don't know
whether i only should be annoyed at it or not, were it
optional, i.e via
an app-specific setting for FF, then i could choose whether
to use it or
not...
Maybe, as a thought to the developers, just maybe the
verbosity levels
should be revamped. Right now as i understand it, all the
brief
verbosity does is removing reading of shortcut keys, maybe
user profiles
would be another way? That way one could have a dialog
reachable from
the prefs where you could change what to read and what not
to read and
have profiles or levels, rather like Jaws or Mobile speak,
where you
have beginner, intermediate and advanced profiles wher you
have certain
default values, but you can still customize the behaviour of
these profiles.
/Krister

Hermann wrote:
> Hi Krister,
> I know this blockquotes from Jaws and Window Eyes, but
I never ever used 
> it.
> Regarding "lage objects": I think it could be
rebuild to just moving to 
> "large blocks of non-link text"; then it
would be useful. But actually you 
> land sometimes on links, headings etc., and I couldn't
figure out why. 
> Perhaps because they have long names...
> Regarding frames: There are FF hotkeys for frames (F6
and Shift+F6), but 
> the problem the developers have to solve here is the
identification of the 
> frames. This seems a bit tricky since they often are
not labeled properly. 
> Maybe further accessibility development by the FF team
is needed, but I 
> think it should be done as soon as possible.
> Another item are anchors. Either there is a new
keystroke implemented, or, 
> since they are links, pressing "u" and
"shift+u" should announce them as 
> they appear. This variant is implemented in Firevox, it
announces "external 
> link" and "internal link" when you come
upon a link.
> Hermann
_______________________________________________
Orca-list mailing list
Orca-listgnome.org
http
://mail.gnome.org/mailman/listinfo/orca-list
Visit http://live.gnome.org/Orca
 for more information on Orca

I think we should keep things easy, so I'm not sure wherther
it is a good 
idea to have so many options.
What's so disturbing when Orca, for example, announces
"internal link" when 
you land on one by pressing "u" or shift+u? If it
is an external link, It's 
sufficient to announce it as "link" like it
already does. I don't think 
that 
this is too much verbosity. And for braille users I prefere
abrieviations 
such as "hx" to announce the heading level.
Of course we can demand more and more optional features,
but, on the other 
hand, I remember well the complains about bloating the
software. So we 
should think carefully to hold the ballance.
Hermann
_______________________________________________
Orca-list mailing list
Orca-listgnome.org
http
://mail.gnome.org/mailman/listinfo/orca-list
Visit http://live.gnome.org/Orca
 for more information on Orca

Hello Benjamin,
to make a long story short: I think you've confirmed my
point of view that 
the blockquotes feature is not really necessary, since this
styling of 
webpages is often missused.
Asking for different features, I referred to developping
capacities; I 
dodn't want to stress the fatct, whether they are comparable
or not. Or, to 
put it in a simple image: The team sometimes tends to serve
the pie before 
they 
have served the beef; I prefere the other way.
When I talk about textblocks, I suggest you download and
install any 
Windows screen-reader and check out how Jaws, Window-Eyes or
the 
Webformator handle this feature. It is simple a navigation
helper.
Large Objects: It's not the point whether I can reproduce
the navigation, I 
simply don't know what those "objects" are. It
seems to me they can be 
anything, which means, in the end, they mean nothing. A link
is a link is a 
link, and a heading is a heading is a heading. But I've
absolutely no idea 
about the systematics of that odd "objects";
that's why I call the "large 
object" navigation stochastical.
Background information: Download and install the mentioned 
screen-readers/web-tools. Place the cursor on, let's say a
link, Press 
insert+shift+f1 in Jaws, insert+e in Window-Eyes or f11 in
Webformator. 
Then you know what I mean. And BTW: In Firevox the keystroke
is 
control+shift+q; press it several times and you know what
I'm talking 
about.
I'm an everyday scren-reader user and a common web user and
my point of 
view is the user's 
perspective, that's the whole story.
Hermann
_______________________________________________
Orca-list mailing list
Orca-listgnome.org
http
://mail.gnome.org/mailman/listinfo/orca-list
Visit http://live.gnome.org/Orca
 for more information on Orca

With regards to long objects, "block" is a very
confusing term in this 
context. It can be used to refer to arbitrary selections of
text, 
semantic segments of text, new paragraphs, or indented text.
If I 
understand you correctly, when you talk of a "block of
non-link text" 
you mean a segment of text that doesn't contain a hyperlink
anchor. For 
instance you might have a long paragraph with two links
interspersed 
amongst the text, such that it would consist of three
non-link text 
segments: one before the first link, one between the links,
and one 
after the second link. If I've got this right, that gives
rise to two 
questions:

1. How long is a large segment as opposed to a small
segment?

2. What is the utility of cycling between large non-link
segments? Is 
the real use-case here to skip over navigation areas
containing long 
lists of links and little or no non-link text?

With regards to frames, when you say the
"identification of the frames", 
do you mean announcing some sort of title for each frame as
you cycle 
through them rather than "HTML container" which is
what Orca currently says?

The Window-Eyes HTML manual has two frames:

http:
//www.gwmicro.com/Window-Eyes/Manual/HTML/

Each frame actually is given a title attribute. One is
"Navigation 
frame"; the other is "Content frame". The
Firefox Accessibility 
Extension provides a frame list feature and the beta can be
used with 
Firefox Minefield:

http://fi
refox.cita.uiuc.edu/index.php#beta

It correctly lists these title attributes as the frame
titles, following 
WCAG:

h
ttp://www.w3.org/TR/WCAG10-HTML-TECHS/#frame-names

I had a look at the frames in Accerciser and Firefox is
putting the 
title attribute into the description property of the Frame
object. So 
this should certainly be possible to implement with Orca.

An interesting question is what heuristics to use, if any,
to identify 
frames in the absence of such a title attribute or
description property. 
For example, we might wish to use the title element, the
first heading 
element, or failing that the first text from a document
embedded using a 
frame. I don't know whether Firefox already inserts any of
those into 
the description property.

I didn't follow what you were saying about anchors.
Typically, external 
and internal links are not presented differently by visual
user agents, 
so they do not "appear" differently to most users.
The heuristics to 
determine whether a link is genuinely external to a given
resource can 
be complicated by a web resource being referenced by more
than one URI, 
so it's worth thinking about precisely what we mean by
"external" in 
this context. If by internal means a fragment of the current
URI as 
opposed to the current resource, then that should be
possible.

--
Benjamin Hawkes-Lewis

Hermann wrote:
> Hi Krister,
> I know this blockquotes from Jaws and Window Eyes, but
I never ever used 
> it.
> Regarding "lage objects": I think it could be
rebuild to just moving to 
> "large blocks of non-link text"; then it
would be useful. But actually you 
> land sometimes on links, headings etc., and I couldn't
figure out why. 
> Perhaps because they have long names...
> Regarding frames: There are FF hotkeys for frames (F6
and Shift+F6), but 
> the problem the developers have to solve here is the
identification of the 
> frames. This seems a bit tricky since they often are
not labeled properly. 
> Maybe further accessibility development by the FF team
is needed, but I 
> think it should be done as soon as possible.
> Another item are anchors. Either there is a new
keystroke implemented, or, 
> since they are links, pressing "u" and
"shift+u" should announce them as 
> they appear. This variant is implemented in Firevox, it
announces "external 
> link" and "internal link" when you come
upon a link.
> Hermann
> _______________________________________________
> Orca-list mailing list
> Orca-listgnome.org
> http
://mail.gnome.org/mailman/listinfo/orca-list
> Visit http://live.gnome.org/Orca
 for more information on Orca
> 

_______________________________________________
Orca-list mailing list
Orca-listgnome.org
http
://mail.gnome.org/mailman/listinfo/orca-list
Visit http://live.gnome.org/Orca
 for more information on Orca

Re: blockquote navigation

<blockquote> is not a styling, although it is
sometimes misused for 
styling. Table and heading elements are often similarly
misused and 
authors frequently fail to use <p> for paragraphs, but
I assume you're 
not opposed to navigation by table, heading, and paragraph
elements.

It's one thing to say another feature should have been
implemented 
first, and another thing to call for the removal of a
feature. Taking 
features out also requires development time.

I suspect the team's priorities are partly determined by
ease of 
implementation, partly by personal interests, and partly by
what feature 
requests have been made on the bug tracker. Quite possibly I
missed 
something, but I can't find a request to read the titles of
frames 
there, for example, even though it's a good feature idea.
It's probably 
more efficient to request urgent features explicitly than
complain about 
the addition of different features after the event.

Re: Text blocks

The Window-Eyes manual mentions "Next text" (X)
and "Previous text" 
(Shift + X):

http://www.gwmicro.com/W
indow-Eyes/Manual/HTML/index.html?19_15morenavigationoptions
.htm

When I tried it out, it ignored the size of text segments
entirely, 
happily reading out even tiny snippets of text. Oddly,
although it 
skipped links per se, it also read out title attributes from
<abbr> 
elements inside links. Presumably that's a bug not a
feature.

But JAWS has a similar feature that seems closer to your
"large text 
blocks". According to the manual:

> Next Non Link Text use N. Moves the focus to the next
block of
 > non-link text that is larger than 25 characters.
> 
> Prior Non Link Text use SHIFT+N. Moves the focus to the
previous
> block  of non-link text that is larger than 25
characters.
> 
> Note: This value can be adjusted in the default.jcf or
browseui.jcf
 > (Internet Explorer 5 and 6) JAWS configuration files
under Set
 > OptionsHTML.

Dolphin's HAL manual says you can jump to the next text,
skipping links, 
with Left Control + Left Shift + Cursor Right. It's not
clear from the 
manual whether the text has to be a certain length.

Which of these models is better?

Some users at least seem to use the N key purely to skip
groups of 
links. e.g. a request was made to implement this feature in
NVDA phrased 
in the following terms:

> Also, in Internet Explorer, can you please add a quick
key to jump 
> to next non-link text like in JAWS with the N key, so I
can step over a large 
> groups of links without repeatedly tapping down arrow
again and again? It just 
> makes me tired!

http://www.freelists.org/archives/nvda/04-2007/msg00
343.html

If the use-case for this feature is basically to skip long
navigation 
lists, we might (I suppose) be able to devise more efficient
algorithms 
for doing so and create a "Skip navigation"
command. Just a thought.

Re: background information

The Window-Eyes manual specifies precisely what properties
are read out:

> * Description – i.e. Link, Picture, Edit Box, etc
> * Font – i.e. Arial, Verdana, Tahoma, etc.
> * Font Size – i.e. 10pt, 12px, larger, etc.
> * Font Style – i.e. Normal, Bold, Underline, etc.
> * Foreground – Color, either named (i.e. Purple) or
in RGB (i.e. 255, 255, 0)
> * Background – Color, either named (i.e. Purple) or
in RGB (i.e. 255, 255, 0)
> * ALT text if available
> * TITLE text if available
> * Filename if the element is an image

http://www.gwmicro.com/Window-Eyes/
Manual/HTML/19_15morenavigationoptions.htm

Judging from:

http://developer.mozilla.org/en/docs/Accessibili
ty/ATSPI_Support

It should be possible to extract most of this information
from AT-SPI.

Would we want to add anything to this list (for example
axis, scope, and 
headers attributes for table cells, or language
information)? Do we want 
heuristics to try to put approximate names to colour values
specified in 
hexadecimal? How about special treatment for microformats
using the 
title attribute, e.g. to store an ISO date? Would we want
key bindings 
for particular properties or groups of properties, or just
one key 
binding for all of them? What order should they come in?

Re: large objects

I guess what these mysterious "large objects" are
is something that 
should be explained in the Orca manual. Without knowing what
it is 
supposed to be, it's hard to assess the feature.
"Large" doesn't seem to 
be the right word for what it's doing, which based on a bit
of 
experiment seems to mainly involve skipping over lists.

--
Benjamin Hawkes-Lewis

Hermann wrote:
> Hello Benjamin,
> to make a long story short: I think you've confirmed my
point of view that 
> the blockquotes feature is not really necessary, since
this styling of 
> webpages is often missused.
> Asking for different features, I referred to
developping capacities; I 
> dodn't want to stress the fatct, whether they are
comparable or not. Or, to 
> put it in a simple image: The team sometimes tends to
serve the pie before 
> they 
> have served the beef; I prefere the other way.
> When I talk about textblocks, I suggest you download
and install any 
> Windows screen-reader and check out how Jaws,
Window-Eyes or the 
> Webformator handle this feature. It is simple a
navigation helper.
> Large Objects: It's not the point whether I can
reproduce the navigation, I 
> simply don't know what those "objects" are.
It seems to me they can be 
> anything, which means, in the end, they mean nothing. A
link is a link is a 
> link, and a heading is a heading is a heading. But I've
absolutely no idea 
> about the systematics of that odd "objects";
that's why I call the "large 
> object" navigation stochastical.
> Background information: Download and install the
mentioned 
> screen-readers/web-tools. Place the cursor on, let's
say a link, Press 
> insert+shift+f1 in Jaws, insert+e in Window-Eyes or f11
in Webformator. 
> Then you know what I mean. And BTW: In Firevox the
keystroke is 
> control+shift+q; press it several times and you know
what I'm talking 
> about.
> I'm an everyday scren-reader user and a common web user
and my point of 
> view is the user's 
> perspective, that's the whole story.
> Hermann
> 

_______________________________________________
Orca-list mailing list
Orca-listgnome.org
http
://mail.gnome.org/mailman/listinfo/orca-list
Visit http://live.gnome.org/Orca
 for more information on Orca

The textblock scheme is the very same in Jaws and
Window-Eyes, and WE's 
behavior is not at all a bug. It has been set up to odd
default values, 
while Jaws uses a reasonable value. I don't care about the
keystroke, I 
want 
to see it work. And if you know a better algorithm to
implement it, please 
contact the Orca team.
The background information provided by WE seems sufficient,
I think we 
should not extend it.
The frame information is under Windows either provided by
the DOM 
mechanism or by MSAA; the latter seems to give better
results (WE). Since 
there is no MSAA in Linux, the information can be obtained
either by FF's 
DOM or by AT-SPI, if possible. I think it should be checked
out what's 
better. Properly labeled frames are no problems (WE manual),
but that's not 
usual on many pages, so the task is to get as much useful
information as 
possible.
If the team, with your assistance, manage to refactor the
"large object" 
stuff, so that a common user understands it, OK!
An BTW: Recognizing headings and tables is more important
then block quotes 
- they belong to the "beef".
Hermann
_______________________________________________
Orca-list mailing list
Orca-listgnome.org
http
://mail.gnome.org/mailman/listinfo/orca-list
Visit http://live.gnome.org/Orca
 for more information on Orca