[jawsscripts] Re: Object/element identification in html land? was The problematic highlight bar non-reading of an HTML-based "dropDown," that isn't identified as an editCombo box?

From: Scott Huey <shuey@xxxxxxxxx>
To: "jawsscripts@xxxxxxxxxxxxx" <jawsscripts@xxxxxxxxxxxxx>
Date: Wed, 4 Dec 2013 15:28:04 +0000
Left something out in that first paragraph which is rather important, sorry 
about that. The missing point regarding my justification on ClassName as a good 
anchor point for validation checks is that the className, since it arises from 
and is indicative of a development kit or creation tool, is usually not put in 
there by the DevelopER consciously. When class names in web apps stay static 
through updates to the app itself, that usually indicates the behavior hasn't 
changed radically because the development kit hasn't changed, so it's a 
relatively safe shorthand for detecting if scripting code is able to predict 
the behavior of the object it's looking at. 


-----Original Message-----
From: jawsscripts-bounce@xxxxxxxxxxxxx 
[mailto:jawsscripts-bounce@xxxxxxxxxxxxx] On Behalf Of Scott Huey
Sent: Wednesday, December 04, 2013 10:20 AM
To: jawsscripts@xxxxxxxxxxxxx
Subject: [jawsscripts] Re: Object/element identification in html land? was The 
problematic highlight bar non-reading of an HTML-based "dropDown," that isn't 
identified as an editCombo box?

Personally, I think the ClassName of an HTML object is a reasonable point to 
anchor validation checks. "Class" as an HTML attribute is effectively 
arbitrary, but in my experience (from the outside looking in, mind you) is that 
the existence of a class attribute is more or less always indicative of the web 
tool being created using a specific tool or web developer kit to produce the 
content 

Assuming we're talking about "Objects" in this context as HTML DOM components, 
then I can point you to the source I got my own crash-course on the subject. 
W3schools.com came up on my initial google search looking for basic web  coding 
information, and they have a beginner guide to DOM manipulations at 
http://www.w3schools.com/jsref/dom_obj_all.asp . It's written with Javascript  
in mind, but what I discovered very quickly was that once you have an object 
reference to the Document in JAWS (which is where all my code snippets used 
IEGetCurrentDocument() to produce an oFoc object), every single one of the 
basic manipulations works pretty much the same way as it would be written for 
JavaScript. The only limitations I've found so far are that JAWS cannot 
accommodate chaining methods of methods together as well as it seems Javascript 
can, and it has issues with data typing errors. In the former, Javascript can 
handle "Document.getElementByID("id").firstChild.nextSibling.ID" to ret  urn a 
value, but even if JAWS doesn't blow up with some variant of an "expected A , 
not B" compiler error, it fumbles the return and the code you might be using 
with that expected return doesn't work properly. In the second, JAWS throws 
periodic  data type mismatch compiler errors with the returns of the DOM 
methods if you try to use a return as part of certain statements (i.e. even 
though the method "element.tagName" returns a string, JAWS will complain about 
typing if you try to concatenate it with another string as part of a variable 
assignment). 

Quick heads up, the w3schools site isn't particularly well-designed for 
efficient Screenreader browsing, though I wouldn't categorize it as offensively 
inaccessible. There's just a large, panel bar with "chapter" quick links which 
you'll need to drill down through before reaching the main content of current 
"chapter" on the page. There's also a good solid source in the Mozilla 
Developer Network (https://developer.mozilla.org/en-US/docs/DOM) which I find 
myself referencing periodically.

The usage of the period as part of the assignments is actually a critical 
syntactic element of the statement. In Object Oriented Programming, the Objects 
themselves can contain full blocks of processing code, and have system 
resources of their own assigned with which they can execute these functions. In 
Object Oriented Programming, what we're doing would be called a "invoking" a 
"method", which conceptually can be likened to JAWS running a Function which 
JAWS does not own or understand, by way of requesting the Object to do so and 
then hand JAWS the return value.  So in one of the code snippets I presented 
earlier, the line...

let oFoc = oDocument.getElementByID(sElementID)

... is actually instructing JAWS to instruct the Document Object of the web 
browser itself (which JAWS has a reference to in oDocument) to run an internal 
process to find an element using the string sElementID, which JAWS passes. In 
the case of this method (getElementByID), the Document returns a "Node" Object, 
which we're assigning to oFoc in the  JAWS, and that Node has its own methods 
and properties which we can manipulate.  

As to the second question, you might need to experiment with it. In the context 
of my code, the user is assumed to be switching between VirtualPC and Forms 
mode due to the raw text content of the pages they need to navigate being 
important (it's a Call Center script program, so there's paragraphs of text 
reminders for speaking interspersed with only periodic controls). I'm not sure, 
but I believe turning OFF the virtual cursor will alter the states in regards 
to IsFormsModeActive()  as JAWS internally understands it. In that case, you 
probably could use a check for IsPCCursor() instead. Easy enough to test, I'd 
suggest just making a simple script (maybe even tie it to the 
FocusChangedEvent) with a test to the tune of "IF IsFormsModeActive() THEN 
Beep() ENDIF" 

In the context of your initial question, I'd suggest abandoning the idea of 
trying to capture the selected item with custom highlighting  entirely. JAWS is 
just too unreliable with web content when it comes to formatting, and in 
particular color coding. I have yet to run into a single case where Coloration 
in a web app is NOT the result of a stylesheet fragment or assignment via 
runtime javascript code (which confounds the JAWSCursor), and someone submitted 
a comment on this list a bit back where they noted that they'd found JAWS was 
generating RGB values for HTML BACKWARDS (as in "BGR"). I'm having trouble 
finding it, I just remember the conundrum as very odd, so I apologize to the 
thread's creator that I cannot cite by name. In most cases I've seen, rich 
internet applications store their selection choices for lists/dropdowns 
somewhere outside the main body content (basically a hidden list which the 
combobox references either front- or backend). I'd start by opening the drop  
down, selecting any available option, and making note of the text content. 
Then, search the page source for that exact phrase. If your application follows 
the pattern I've seen most, there will be a list, table, or possibly a div 
block (probably somewhere deep near the bottom) which contains that phrase plus 
all the other options in the dropdown. The pattern I've seen in tools I've 
worked with is that a web app often applies selection highlights by hooking a 
CSS style to something it changes in the hidden section as the selection 
changes. If you've found a list with the options, compare the nodes which 
contain the unselected options to the node which contains the selected one (be 
sure to check their parents as well, if they're nested, all the way back to the 
common ancestor). Something may jump out at you as different (usually a Class 
but it might be contained as a Style, Name, or Value as well), and if you can 
find that you'll have the key to the solution. All you'll need t  o do from 
there is hook the JAWS functions for next and previous selection, and include a 
subroutine that will hunt down the DOM element in that list/table/section based 
on that unique attribute the CSS is hooking. Then, you can assign it a 
reference, and pull back its text content with the innerText property so it can 
be spoken (i.e. SayMessage(OT_MESSAGE, oSelected.InnerText) ).



-----Original Message-----
From: jawsscripts-bounce@xxxxxxxxxxxxx 
[mailto:jawsscripts-bounce@xxxxxxxxxxxxx] On Behalf Of Geoff Chapman
Sent: Tuesday, December 03, 2013 9:53 PM
To: jawsscripts@xxxxxxxxxxxxx
Subject: [jawsscripts] Re: Object/element identification in html land? was The 
problematic highlight bar non-reading of an HTML-based "dropDown," that isn't 
identified as an editCombo box?

Oh my goodness Scott! This is all simply amazing stuf! Thank you so so so much 
for all the energy and time you've put into sharing all of this below to assist 
me!
absolutely incredible!
Wow!

It'll certainly take a little further digesting/sifting, for I'm unfortunately 
not a really experienced Scripter myself, and I can see, (as ai feared it would 
be,) that this is quite comprehensive and involved coding, at least for my 
experience level.
But I can't tell you how much I appreciate you having shared all these code 
segment examples!  Sooo helpful!

OK so I'm a bit torn here, and not sure exactly which way to jump re next 
protocol steps. In the sense that, I sooo know what it feels like, when you 
give your all, and put in a whole bunch of time/energy into trying to help 
someone out, only to find that it ends up entangling you in even more undesired 
spoon-feeding energy/work, when the person comes back with further basic 
questions about what you've already considerably spent yourself in putting out 
there! :( :)

So, If you'd like to have a break and let someone else perhaps have a shot at 
answering the below, please feel totaly free to do that.  As the last thing I 
wanna do is to engender that feeling of mental tedium/burden described above.
:)

However, unfortunately I do have a couple of what are first really basic 
queries I know, in relation to the syntax of working with objects as your code 
does below.

Then, after getting a handle on this, I feel I'm gonna have to glean a little 
more in-depth knowledge, on understanding, let alone effectively working with, 
the innards of html object element information, such as class, TagName, and 
style - etc, before I'm gonna have a hope of getting to my solution; if it 
turns out I need a scripting solution, which it's certainly looking 
unfortunately like it might be the case so far. :(

I presume from a small investigative study of your below, that you feel Class 
is a reasonable criteria validator to be utilizing for a conditional behaviour 
modification, "If ... then",  check?

So, I guess the first thing to ask is, to negate your having to tediously 
answer such low-level queries as to exactly how/when  the period is meant to be 
used, when coding involving objects, which at present, never having tried to 
work with objects before I'm totally unfamiliar with, might you/anyone have any 
tips on where i might go, to get a bit of a crashCourse tutorial on the basics 
of working with Objects?
Specifically, the usage of the period in the middle of names? Which I haven't 
seen before?

secondly, Assuming the Virtual Cursor is turned completely off with 
Jaws+Zee, this would presumably look the same to jaws as a
"FormsModeIsActive" state.  Would that be correct?
Because it looks like so far the most effective way for this user to traverse 
this form, and drop down the lists on each of these boxes with the downArrow 
key, is if we disable Virtual cursor, rather than him flip in and out of Forms 
mode, (as we've set it to manual at this point.)


Third, Would you feel up to sharing any more of your obviously considerable 
experience, as to how you might bring about the natural jaws behaviour we're 
shooting for here in such a situation? Assuming I can narrow down these objects 
reliably, so that I can put specific alternate behaviour for 
DownArrow/UpArrow/SayLine, when these particular elements gain focus?


Thanks again so much Scott for what you've already done below.
It's sure given me a light into this I didn't know existed.  So I hope it is 
helping others to know/grow as well.

Regards
Geoff C.
----- Original Message -----
From: "Scott Huey" <shuey@xxxxxxxxx>
To: <jawsscripts@xxxxxxxxxxxxx>
Sent: Tuesday, December 03, 2013 4:12 AM
Subject: [jawsscripts] Re: Object/element identification in html land? was The 
problematic highlight bar non-reading of an HTML-based "dropDown," that isn't 
identified as an editCombo box?


> I've been building some scripting for an AJAX based web application, 
> which includes specific behavioral modification based on conditionals 
> of HTML element type as you're describing.  In the context of my 
> scripting, I was co-opting the arrow key navigation to behave uniquely when 
> the "focused"
> section of the page was built with specific components which did not 
> behave "naturally" (i.e. the bane of my existence, the "ajax__tab_xp 
> ajax__tab_container ajax__tab_default" interactive tab container DIV 
> object). It took me about a couple days to sort it all out, but in the 
> end I was able to build out a nested conditional statement which could 
> accurately check for details of the object in question. The steps 
> required are radically different depending on which cursor is active. 
> In the context of the web app I'm working with, JAWS and Invisible 
> Cursors return wildly inaccurate data, so I was able to constrain my 
> cases to a binary choice between either the PC Cursor ("IF 
> IsFormsModeActiv
> e() THEN") and the Virtual Cursor ("IF isVirtualPCCursor() THEN").
>
> If the active cursor is the PC cursor (i.e. Forms Mode is active), the 
> process is a relatively straightforward one using pure Document Object 
> Model (DOM) level interactions. The DOM contains a method for 
> instantly returning the current object with focus, 
> "document.ActiveElement", which can then be queried directly for its 
> specific details. JAWS code to retrieve the information would look as follows:
>
> /* --- Start Code Segment --- */
> Var object oDocument, object oFoc, string sElementID let oDocument = 
> IEGetCurrentDocument() let sElementID = oDocument.activeElement.ID let 
> oFoc = oDocument.getElementByID(sElementID)
> ; From here, I could now manipulate the object directly further in the 
> scripting, including directly extract data about its class, tag, and 
> style.
> /* --- End Code Segment --- */
>
> If the VirtualPC cursor is active and the user is traversing the OSM, 
> I think we're stuck using the GetElementDescription function and 
> extracting the relevant substring from its return. Conceptually, I'm 
> absolutely certain that GetElementDescription itself is using more 
> targeted sub-functions to find the individual parts it returns (id, 
> class, etc., which it then concatenates together), but I haven't found 
> any documentation indicating what they might be so I've relied on 
> delimiting the return using the newline ("\n") character as the 
> delimiter. The results are fairly reliable, under the assumption that 
> elements of interest will ALWAYS have the attribute I'm checking (i.e. 
> when I check for IDs, I already know in the specific web tool I'm 
> working on that everything with text content drawn into the OSM has a 
> programmatically assigned ID value in the DOM). For this, the code 
> looked something like
> below:
>
> /* --- Start Code Segment --- */
> Var string sDelimited, int iIndex, object oFoc, object oDocument Let 
> oDocument = IEGetCurrentDocument() let sDelimited = 
> GetElementDescription (1, FALSE) ; Returned only information regarding 
> the specific element, as Ancestor data is here irrelevant.
> let iIndex = StringSegmentIndex (sDelimited, "\n", "id=", False) 
> ;Delimit the string by the NewLine character, return the index of the 
> one for the ID via the tag "id="
> let sDelimited = StringSegment (sDelimited, "\n", iIndex) ;Pare down 
> the string to just the single segment let sDelimited = 
> StringReplaceSubstrings (sDelimited, "id=", "") ;At this point, 
> sDelimited is now purely the ID let oFoc  = 
> oDocument.getElementByID(sDelimited)
> ; From here, I could now manipulate the object directly further in the 
> scripting, including directly extract data about its class, tag, and 
> style.
> /* --- End Code Segment --- */
>
> A basic example adapted from my existing code, where I would be trying 
> to extract the "current element's" Class for the purposes of a 
> validation check, altogether looks like so. For reference, oDocument 
> was set with IEGetCurrentDocument much higher in the function.
>
> /* --- Start Code Segment --- */
> IF FormsModeActive() Then
> Let sClass = oDocument.activeElement.className ELIF 
> IsVirtualPCCursor() Then Let sDelimited = GetElementDescription(1, 
> FALSE) ; Returned only information regarding the specific element, as 
> Ancestor data is here irrelevant.
> Let iIndex = StringSegmentIndex(sDelimited, "\n", "class=", False) Let 
> sClass = StringSegment(sDelimited, "\n", iIndex) Let sClass = 
> StringReplaceSubstrings(sClass, "class=","") Else ;Logic in this 
> script assumed the JAWS and Invisible Cursors represent an invalid 
> state (as they return anomalous results), so this branch simply did 
> cleanup and then called the original implementation of the overridden 
> script EndIF
> /* --- End Code Segment --- */
>
> From that point, I would use sClass as part of an extended ELIF block, 
> essentially working as a SELECT CASE construct, to branch the 
> behaviors depending on the specific type of element on which the User 
> had (virtual) focus. I think it fairly robust, in that my 
> implementation has survived stability and reliability testing on the web-tool 
> I've designed it for.
> That said, it's a scripting solution for a web tool we host in house. 
> So while it's not one we wrote ourselves (if it was, I would have made 
> it conform to WCAG so scripting wouldn't be necessary at all), it's on 
> a server which won't be updated without our consent. I've admittedly 
> designed around a specific edition of a designer tool, so YMMV of 
> course but I also wouldn't recommend taking this sort of approach with 
> any hosted tool that could be redesigned without warning. Just about 
> any attribute you could reference for a Rich Internet Application is 
> arbitrary in its specific string. The only things I would say really 
> end up being set in stone (to the point that you could define them 
> statically in a JAWS script) as far as a Web App is concerned are the 
> flat Tag Names, and of course ARIA markups. Theoretically if there are 
> ARIA markups in your web application, then scripting shouldn't be needed.
>
>
>
>
>
> -----Original Message-----
> From: jawsscripts-bounce@xxxxxxxxxxxxx 
> [mailto:jawsscripts-bounce@xxxxxxxxxxxxx] On Behalf Of Geoff Chapman
> Sent: Sunday, December 01, 2013 4:42 PM
> To: jawsscripts@xxxxxxxxxxxxx
> Subject: [jawsscripts] Object/element identification in html land? was 
> The problematic highlight bar non-reading of an HTML-based "dropDown," 
> that isn't identified as an editCombo box?
>
> Another perhaps even more basic question related to my issue though 
> that occurs.
>
> In Html land, where it seems we're not talking about identifying 
> certain so-called windows, at the detection of which we might wish to 
> alter certain default behaviour, but rather, elements? or Objects?"
> How would one go about even setting up conditional statements to check 
> if we were focused on a certain object or element, like these 
> particular edit/list-box combo controls, to even introduce a solution? 
> Once we've come up with one that works anyway?
>
> E.g. in the element information, part of which I'll rePaste below, We 
> read things like Class, and ID, and Tag DIV, etc.
> So is it possible in Script land, to somehow code up conditional 
> statements that might look for these identifiers, and to thus insert 
> behaviour change when located?
> How might one even begin to go about doing this?
> Is such methodology in the least reliable/robust in a browser based 
> situation such as this?
>
> Or, if too onerous to answer directly, does anyone know just where I 
> might even begin to go for study of such a situation where this might 
> be being done somewhere already?
>
> Here's a little code from the original element information on this 
> particular control:
>
>
>
>>> Element Information:
>
> Tag INPUT has 7 parameters:
>
> autocomplete=off
>
> class= x-form-text x-form-field  x-form-focus
>
> id=par-1C05555973514DF39BE96CF44694A3B2-d
>
> size$
>
> style
> tabindex!002
>
> type=text
>
> MSAA Role*
>
>
> Tag DIV has 3 parameters:
>
> class=x-form-field-wrap x-form-field-trigger-wrap  
> x-trigger-wrap-focus
>
> id=ext-gen437
>
> style=
>
> Tag DIV has 3 parameters:
>
> class=x-form-element
>
> id=x-form-el-par-1C05555973514DF39BE96CF44694A3B2-d
>
> style=
>
> Tag DIV has 3 parameters:
>
> class=x-form-item
>
> id=ext-gen434
>
> tabindex=-1
>
> Tag DIV has 3 parameters:
>
> class= x-request-column x-request-column-out x-form-label-top x-column
>
> id=ext-comp-1504
>
> style=
>
> ----- Original Message -----
> From: "Jim Snowbarger" <Snowman@xxxxxxxxxxxxxxxx>
> To: <jawsscripts@xxxxxxxxxxxxx>
> Sent: Saturday, November 30, 2013 2:53 PM
> Subject: [jawsscripts] Re: The problematic highlight bar non-reading 
> of an HTML-based "dropDown," that isn't identified as an editCombo box?
>
>
>> Does anybody know what window jaws will search for a custom highlight?
>> That
>> isn't a applicaition-wide search, is it?
>> Will it not search if a normal caret can be found?
>> If not, how could one make it ignore the caret, and use highlight 
>> searching only.
>>
>> ----- Original Message -----
>> From: "Geoff Chapman" <gch@xxxxxxxxxxxxxxxx>
>> To: <jawsscripts@xxxxxxxxxxxxx>
>> Sent: Friday, November 29, 2013 9:08 PM
>> Subject: [jawsscripts] Re: The problematic highlight bar non-reading 
>> of an HTML-based "dropDown," that isn't identified as an editCombo 
>> box?
>>
>>
>> Hi everyone.
>>
>> NO joy with any of the suggestions tried so far, except for Arder's 
>> below.
>> haven't gotten to try that one yet, so keep any other suggestions on 
>> this coming.
>>
>> Jackie when I get a remote login to this system, I'll be able to 
>> trial a few more things like HSC's GetColorAtPoint functionality, as 
>> opposed to the
>> Jaws+TopRow5 Get TextColor Foreground/Bakground, the dude I'm working
>> with
>> was certainly surprised when what he thought was a blue highlight 
>> bar, got defined as Grey13 on lavender.
>> Although I thought lavender was a type of blue, but ... hey, what 
>> would I know.
>>
>> Jim tested with the Jaws cursor tucked firmly out of the way, and 
>> invisible cursor on highlighted text just with focus etc, still 
>> thinks it's grey13 on Lavender, whereas the non-highlighted text in 
>> the list is identified as
>> Grey13 on white. so we tried setting the custom highlight as anything 
>> on Lavender, since it seemed to us that it was maybe the background 
>> that was doing the highlighting, but, still no joy. jaws doesn't 
>> track it as highlight is moved up and down.
>> Sigh.
>> ----- Original Message -----
>> From: "Artur Räpp" <rtr@xxxxxxxxx>
>> To: <jawsscripts@xxxxxxxxxxxxx>
>> Sent: Friday, November 29, 2013 5:36 PM
>> Subject: [jawsscripts] Re: The problematic highlight bar non-reading 
>> of an HTML-based "dropDown," that isn't identified as an editCombo 
>> box?
>>
>>
>>> Hello,
>>>
>>> On a webpage I found following solution with a "edit-combo"
>>>
>>> I moved to edit field (edit part of the edit-combo), turned off the 
>>> virtual cursor, typed some letter, pressed once down arrow, pressed 
>>> tab once, landed on the listbox part and then moved down with down arrow.
>>> After I found what I needed, I turned virtual cursor again on.
>>>
>>> On another such HTML editcombo I didn't found any solution and 
>>> copied the choices into a file and copi-pasted from the file the 
>>> right choice into edit field. With the full text system chose the 
>>> right choice. The coices were static. To get the choices, I used 
>>> ins+ctrl+shft+f1 keyboard command (Advanced Element Information).
>>>
>>> my 0.01$.
>>>
>>> Artur
>>>
>>> 29.11.2013 2:00, Geoff Chapman kirjutas:
>>>> Hi Scripters.
>>>> Ok, in relation to my query the other day, about a particular 
>>>> so-called "drop-down,"type element on a particular client's 
>>>> workPlace site that Jaws will not handle correctly, I now have more 
>>>> information about this, and would very much appreciate anyone's 
>>>> views on where I might even start to solve this problem, as it's a 
>>>> dealBreaker for this poarticular employed person, re being able to 
>>>> do his job on the new system the company is rolling out in February.
>>>> It's a bit of a lengthy query, but I'd appreciate those who're 
>>>> able, to stick with me whilst I try and elaborate what, I, think is 
>>>> so far going on here.
>>>>
>>>> Ok So, first, this is an html-based system, and everything else 
>>>> here seems to be beautifully accessible, even the fields are all 
>>>> labelled properly.
>>>>
>>>> Second, I've turned off AutoForms mode, as I find this mode very 
>>>> problematic and just annoying to navigate, after years of using the 
>>>> manual forms mode modality.
>>>> Also trying to eliminate as much of the so-called automatic smarts, 
>>>> as possible, to eliminate unknown and possibly erroneous behaviours etc.
>>>>
>>>> However there's this one particular totally problematic 
>>>> control-type/element, which consists of initially what Jaws 
>>>> identifies as a straight ahead, edit, box.  However, this 
>>>> particular element, operates in practise, actually, more like what 
>>>> I understand to be often defined by jaws as an "EditCombo", 
>>>> control.
>>>> Whereby, as you start typing into the edit part of the field, a 
>>>> "drop-down" list appears,matching the entries it finds related to 
>>>> the letters in the text query entered so far.
>>>>
>>>> (Additionally, the field visually contains a little downpointing 
>>>> graphical triangle, which sighted users may click on to drop down 
>>>> the unfiltered list, and visually view it that way, if preferred, 
>>>> with no text query filtering applied.) But which is, not, activated 
>>>> at all by pressing Alt+DownArrow, as one would expect if Jaws 
>>>> handled the element as a proper EditCombo.
>>>>
>>>> Now. The major problem is, that even though pressing downArrow, 
>>>> does start to navigate the entries in this field, placing what it 
>>>> identifies as a Grey13 on Lavender highlight highlight color over 
>>>> each selected entry, Jaws does not read this highlight,, because 
>>>> it's locked into the fact that there's presumably still an iBeam 
>>>> edit cursor sitting in the edit box the whole time. So it just 
>>>> keeps reReading, "Edit," plus whatever the current text query value 
>>>> has been typed in, and is thus populating that edit field, upon 
>>>> each arrow press.
>>>>
>>>> now another perhaps relevant aspect to understand about this 
>>>> particular controlType, as compared to some others I've seen, which 
>>>> although not identifying as EditCombos either, appear to behave in 
>>>> this very way, but which jaws does quite happily read, Is thatt in 
>>>> our version, the control doesn't immediately populate the edit 
>>>> field, with the entire selection one navigates to with the arrow 
>>>> keys.
>>>> Whereas, the ones that jaws does happily appear to read properly, 
>>>> do seem to do this.
>>>> Therefore, in those fields where jaws does identify what one has 
>>>> arrowed to, it's probably not doing so by reading/tracking the 
>>>> highllight bar, but simply by reading what's appearing at the edit 
>>>> cursor.
>>>>
>>>> My first thought was to try and reClass the window, but, this 
>>>> didn't have any good effect, since of course as ai realized later, 
>>>> what I want to do here, is reClass, as far as Jaws is concerned, 
>>>> these individual elements, which are probably not the same as whole 
>>>> windows anyway.
>>>> When I looked at the class of the window in focus, it just showed 
>>>> some basic high-level seemingly irrelevant class  such as Class = 
>>>> MozillaContentWindowClass etc.
>>>>
>>>> So no joy there.
>>>>
>>>>
>>>>
>>>> My second thought, was to try and identify the custom highlight 
>>>> color, and assign this as a custom highlight in jaws, hoping to 
>>>> attract it's attention to the fact that it should be checking for, 
>>>> and tracking/reading, anything in that highlight color.
>>>>
>>>> However, I have no joy doing that either.
>>>>
>>>> putting the jaws cursor on it, it reckons the highlight is Grey13 
>>>> on Lavender, but, yep, this didn't work either.
>>>>
>>>>
>>>>
>>>> So has anyone got any bright ideas as to what I might try next?
>>>> Is Custom Highlight tracking even working in jaws15 at present anywayz?
>>>> Some way I can tell jaws how to properly handle this type of element?
>>>> ignoring the edit box cursor, and tracking the highlight bar instead?
>>>> Below I have included the rather comvoluted, but possibly helpful 
>>>> to someone who may know, element information from the 
>>>> jaws+Shift+f1, whilst over this particular field.
>>>>
>>>> Thanks much for any tips!
>>>> See below for the tech stuff about the elemtn.
>>>>
>>>>
>>>>
>>>> Element Information:
>>>>
>>>> Tag INPUT has 7 parameters:
>>>>
>>>> autocomplete=off
>>>>
>>>> class= x-form-text x-form-field  x-form-focus
>>>>
>>>> id=par-1C05555973514DF39BE96CF44694A3B2-d
>>>>
>>>> size$
>>>>
>>>> style
>>>> tabindex!002
>>>>
>>>> type=text
>>>>
>>>> MSAA Role*
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-form-field-wrap x-form-field-trigger-wrap  
>>>> x-trigger-wrap-focus
>>>>
>>>> id=ext-gen437
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-form-element
>>>>
>>>> id=x-form-el-par-1C05555973514DF39BE96CF44694A3B2-d
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-form-item
>>>>
>>>> id=ext-gen434
>>>>
>>>> tabindex=-1
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-request-column x-request-column-out x-form-label-top 
>>>> x-column
>>>>
>>>> id=ext-comp-1504
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-column-inner
>>>>
>>>> id=ext-gen423
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-panel-body x-panel-body-noheader x-panel-body-noborder 
>>>> x-column-layout-ct
>>>>
>>>> id=ext-gen386
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 2 parameters:
>>>>
>>>> class=x-panel-bwrap
>>>>
>>>> id=ext-gen385
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-panel request-view-category x-panel-noborder
>>>>
>>>> id=ext-comp-1503
>>>>
>>>> style=
>>>>
>>>> Tag FORM has 5 parameters:
>>>>
>>>> class=x-panel-body ss-svc-form-body x-panel-body-noheader 
>>>> x-panel-body-noborder x-form
>>>>
>>>> enctype=application/x-www-form-urlencoded
>>>>
>>>> id=ext-gen184
>>>>
>>>> method=post
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 2 parameters:
>>>>
>>>> class=x-panel-bwrap
>>>>
>>>> id=ext-gen335
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-panel ss-svc-form x-panel-noborder x-form-label-left 
>>>> x-box-item
>>>>
>>>> id=ext-comp-1270
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-box-inner
>>>>
>>>> id=ext-gen334
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-box-layout-ct
>>>>
>>>> id=ext-comp-1272
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-panel-body x-panel-body-noheader
>>>>
>>>> id=ext-gen318
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 2 parameters:
>>>>
>>>> class=x-panel-bwrap
>>>>
>>>> id=ext-gen316
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-panel x-frs-sc-param-form ss-svc-form
>>>>
>>>> id=ext-comp-1268
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-panel-body x-panel-body-noheader x-panel-body-noborder
>>>>
>>>> id=ext-gen315
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 2 parameters:
>>>>
>>>> class=x-panel-bwrap
>>>>
>>>> id=ext-gen314
>>>>
>>>> Tag DIV has 2 parameters:
>>>>
>>>> class= x-panel tab-panel parameters-tpl x-flex-portal-child 
>>>> x-panel-noborder
>>>>
>>>> id=ext-comp-1283
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-flex-portal-child-frame
>>>>
>>>> id=ext-comp-1284
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-tab-panel-body x-tab-panel-body-top
>>>>
>>>> id=ext-gen224
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 2 parameters:
>>>>
>>>> class=x-tab-panel-bwrap
>>>>
>>>> id=ext-gen223
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-tab-panel x-frs-flex-portlet
>>>>
>>>> id=ext-comp-1239
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-panel-body x-panel-body-noheader x-panel-body-noborder
>>>>
>>>> id=ext-gen208
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 2 parameters:
>>>>
>>>> class=x-panel-bwrap
>>>>
>>>> id=ext-gen207
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-panel x-frs-flex-portal x-panel-noborder x-border-panel
>>>>
>>>> id=ext-comp-1223
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-panel-body x-panel-body-noheader x-panel-body-noborder 
>>>> x-border-layout-ct
>>>>
>>>> id=ext-gen206
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 2 parameters:
>>>>
>>>> class=x-panel-bwrap
>>>>
>>>> id=ext-gen205
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-panel x-panel-noborder
>>>>
>>>> id=ext-comp-1227
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-panel-body x-frs-flex-panel-body x-panel-body-noheader
>>>>
>>>> id=ext-gen187
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 2 parameters:
>>>>
>>>> class=x-panel-bwrap
>>>>
>>>> id=ext-gen185
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-panel x-frs-formview
>>>>
>>>> id=ext-comp-1222
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-panel-body x-panel-body-noheader x-panel-body-noborder
>>>>
>>>> id=ext-gen17
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 2 parameters:
>>>>
>>>> class=x-panel-bwrap
>>>>
>>>> id=ext-gen16
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-panel x-frs-content x-border-panel
>>>>
>>>> id=ext-comp-1002
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-panel-body x-panel-body-noborder x-border-layout-ct
>>>>
>>>> id=ext-gen12
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 2 parameters:
>>>>
>>>> class=x-panel-mc
>>>>
>>>> id=ext-gen15
>>>>
>>>> Tag DIV has 1 parameters:
>>>>
>>>> class=x-panel-mr
>>>>
>>>> Tag DIV has 1 parameters:
>>>>
>>>> class=x-panel-ml
>>>>
>>>> Tag DIV has 2 parameters:
>>>>
>>>> class=x-panel-bwrap
>>>>
>>>> id=ext-gen11
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-panel x-panel-noborder
>>>>
>>>> id=WorkspaceFrame
>>>>
>>>> style=
>>>>
>>>> Tag BODY has 3 parameters:
>>>>
>>>> href=https://kfchelp-stg.saasit.com/WorkspaceLoader.aspx?Id=Service
>>>> Req&Profile=ObjectWorkspace&LayoutName=SDA-YAI&TabId=ext-comp-1101
>>>>
>>>> id=main_body
>>>>
>>>> rssúlse
>>>>
>>>> MSAA Role
>>>>
>>>> Tag IFRAME has 8 parameters:
>>>>
>>>> class=x-managed-iframe
>>>>
>>>> frameborder=0
>>>>
>>>> id=ext-gen112
>>>>
>>>> name=ext-gen112
>>>>
>>>> scrolling=auto
>>>>
>>>> src
>>>> style
>>>> title=Object Workspace
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-panel-body x-panel-body-noheader x-panel-body-noborder
>>>>
>>>> id=ext-gen111
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 2 parameters:
>>>>
>>>> class=x-panel-bwrap
>>>>
>>>> id=ext-gen110
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-panel x-panel-noborder
>>>>
>>>> id=panelext-comp-1101
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-border-panel
>>>>
>>>> id=app-frame-body
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-box-item x-border-layout-ct
>>>>
>>>> id=app-content-wrapper
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class=x-box-inner
>>>>
>>>> id=ext-gen10
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-box-layout-ct
>>>>
>>>> id=ext-comp-1003
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-app-content x-border-panel
>>>>
>>>> id=AppBody
>>>>
>>>> style=
>>>>
>>>> Tag DIV has 3 parameters:
>>>>
>>>> class= x-app-frame x-border-layout-ct
>>>>
>>>> id=BorderWrapper
>>>>
>>>> style=
>>>>
>>>> Tag BODY has 3 parameters:
>>>>
>>>> href=https://kfchelp-stg.saasit.com/Default.aspx
>>>>
>>>> id=main_body
>>>>
>>>> rssúlse
>>>>
>>>> MSAA Role
>>>>
>>>> Press ESC to close this message.
>>>>
>>>>
>>>> *****
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Also, today in IE, (haven't tried yet in other browsers) I ran 
>>>> across a Suburb/postcode field, at the below URL, which I wanted 
>>>> you to just eyeball as it
>>>> *appeared* to work similarly to how I imagine our dropdowns on your 
>>>> site are working? which this one, jaws reads perfectly fine? but am 
>>>> unsure of course of the backEnd element/Javascript running ... etc.
>>>>
>>>>
>>>>
>>>> It's at the following URL:
>>>>
>>>> https://www.oxfam.org.au/my/donate/typhoon-haiyan-yolanda?utm_campa
>>>> ign=Typhoon_2nd_ask&utm_medium=Email&utm_source=Typhoon_2nd_ask
>>>>
>>>>
>>>> Leaving virtual cursor enabled for the moment, After tabbing to the 
>>>> correct field, Hit enter to turn on forms mode, then type something 
>>>> that will bring up multiple entries in the list, then hit down 
>>>> arrow. Try it in your version of IE?
>>>>
>>>> For me here, in IE8, jaws reads each highlighted entry just fine as 
>>>> one arrows up and down in this list,
>>>>
>>>> without having to assign any custom highlight colors at all.
>>>>
>>>>
>>>> For interest, though I know it won't mean much to you, here's the 
>>>> backend stuff re the element that Jaws reports when doing a 
>>>> Control+Insert+F1 on that field in IE 8, without forms mode 
>>>> enabled.
>>>>
>>>> (note that this Insert+Shift+F1 keystroke doesn't work when either 
>>>> forms mode is enabled or virtual cursor is off, which are 
>>>> effectively the same thing
>>>> anyway:)
>>>>
>>>>
>>>> Element Information:
>>>>
>>>>
>>>> Tag INPUT has 11 parameters:
>>>> aria-autocomplete=list
>>>> aria-haspopup=true
>>>> autocomplete=off
>>>> class=ui-autocomplete-input
>>>> id=suburb_ac
>>>> jquery172036613851847330175
>>>> name=suburb_ac
>>>> role=textbox
>>>> tabindex!
>>>> type=text
>>>> value=EPPING,  NSW,  2121
>>>> MSAA Role*
>>>>
>>>>
>>>> Tag DIV has 1 parameters:
>>>> class=input
>>>>
>>>>
>>>> Tag DIV has 1 parameters:
>>>> class=clearfix
>>>>
>>>>
>>>> Tag DIV has 1 parameters:
>>>> id=autoggle-au
>>>>
>>>>
>>>> Tag DIV has 1 parameters:
>>>> class=form-row
>>>>
>>>>
>>>> Tag FIELDSET has 0 parameters:
>>>>
>>>>
>>>> Tag FORM has 8 parameters:
>>>> _submit_attached=True
>>>> accept-charset=utf-8
>>>> action=https://www.oxfam.org.au/my/donate/typhoon-haiyan-yolanda
>>>> class=myform form-stacked
>>>> enctype=application/x-www-form-urlencoded
>>>> id=donation_form
>>>> jquery172036613851847330175(
>>>> method=post
>>>>
>>>>
>>>> Tag DIV has 1 parameters:
>>>> class=span11 main-content
>>>>
>>>>
>>>> Tag DIV has 1 parameters:
>>>> class=row
>>>>
>>>>
>>>> Tag DIV has 1 parameters:
>>>> id=bootstrap
>>>>
>>>>
>>>> Tag DIV has 1 parameters:
>>>> class=main-holder
>>>>
>>>>
>>>> Tag DIV has 1 parameters:
>>>> id=main
>>>>
>>>>
>>>> Tag DIV has 1 parameters:
>>>> id=wrapper
>>>>
>>>>
>>>> Tag BODY has 2 parameters:
>>>> href=https://www.oxfam.org.au/my/donate/typhoon-haiyan-yolanda?utm_
>>>> campaign=Typhoon_2nd_ask&utm_medium=Email&utm_source=Typhoon_2nd_as
>>>> k
>>>> rssúlse
>>>> MSAA Role
>>>>
>>>>
>>>>
>>>>
>>>> Press ESC to close this message.
>>>>
>>>> __________ï¿½
>>>>
>>>> View the list's information and change your settings at 
>>>> //www.freelists.org/list/jawsscripts
>>>>
>>>>
>>> __________ï¿½
>>>
>>> View the list's information and change your settings at 
>>> //www.freelists.org/list/jawsscripts
>>>
>>>
>>
>> __________ï¿½
>>
>> View the list's information and change your settings at 
>> //www.freelists.org/list/jawsscripts
>>
>>
>>
>> __________ï¿½
>>
>> View the list's information and change your settings at 
>> //www.freelists.org/list/jawsscripts
>>
>>
>
> __________ï¿½
>
> View the list's information and change your settings at 
> //www.freelists.org/list/jawsscripts
>
>
> __________ï¿½
>
> View the list's information and change your settings at 
> //www.freelists.org/list/jawsscripts
>

__________ï¿½

View the list's information and change your settings at 
//www.freelists.org/list/jawsscripts


__________ï¿½

View the list's information and change your settings at 
//www.freelists.org/list/jawsscripts


__________ï¿½

View the list's information and change your settings at 
//www.freelists.org/list/jawsscripts
References:
[jawsscripts] Re: Object/element identification in html land? was The problematic highlight bar non-reading of an HTML-based "dropDown," that isn't identified as an editCombo box?

Other related posts:
