OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

cgmo-webcgm message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Re: [cgmo-webcgm] IE ActiveX Update


Thanks to you and Ralf for this heads up.  It is ugly.  And it is looming, 
like in 1-2 weeks?!

At 04:49 PM 4/4/2006 -0400, Benoit Bezaire wrote:
>   Microsoft is about to release some troubling IE updates for ActiveX
>   controls... this is regarding the EOLAS vs Microsoft lawsuit. The
>   deployment (Windows Automatic Update) of these updates will
>   apparently start in April. Future versions of Windows and IE
>   (pre-installed on systems) will be shipping with the update
>   (specific dates unknown).

As I gather, then very soon the test suite (2.0 parts) will no longer work, 
right?  Yikes!!!

 From the telecon, I heard some talk about "delay for 30 days", 
"...June".  Can you tell more specifically about the timetable?  Will all 
Win/IE systems with Automatic Update get hit in April?

We were just starting the process to publish the test suite, i.e., make the 
2.0 bits publicly available for the first time.  And this is required for 
W3C process.

>   A system with such an updated version of IE will require user
>   interaction (via a click) before a CGM file embedded in HTML can be
>   displayed on screen. A bad user experience is about to begin for CGM
>   users unless they start changing their web pages.

Please correct me if this is wrong.  I looked at the MSDN page that Ralf 
referenced.  It seems to me that the CGM picture would display, but it 
would be inactive.  Is that right?

Or ... would our typical 2.0 file not even view, because of we use the DOM 
API to tell the control what CGM to load?

<body onload="document.getElementById('ivx1').getWebCGMDocument().src= 
'AppStructure-visibility.cgm'; OnBtnDOM();">
     <td><object id="ivx1" type="image/cgm;Version=4;ProfileId=WebCGM" 
width="480" height="360"></object>

>   The workaround suggested by Microsoft, Apple, Macromedia etc... is
>   to remove the <object> tag from HTML pages and add a script to every
>   web page which dynamically inserts an object tag (if the <object>
>   tag is dynamically inserted, it's not in violation of the EOLAS
>   patent).

Have you (Itedo) guys yet written a prototype/example Test Suite (2.0) file 
that would work (view and be active)?

>   This needs to be dealt with at multiple level...
>   - CGM vendors have to deal with this (generators of HTML + CGM)

As I read Ralf's MSDN reference, it actually appears that there are 
multiple possible solutions.  One assures that the control is interactive 
when loaded.  But the other, if the control were loaded *inactive* and the 
picture were displayed, could be made interactive by a user click.  So ... 
is it possible that different fixes might be appropriate for different 
applications?  (E.g., could a fast fix of "click to activate test" in each 
picture, after ensuring picture is loaded, be used for emergency repair of 
the Test Suite?)

>   - Users needs to update their existing web page (or block the
>   update).

The latter is impractical.

>   - How do we (cgmopen) ease the transition for the webcgm community?
>   - The test suite will need to be updated (in my opinion).

Yes.  Do you have an example of what a test fix/update might look like, for 
any one of our 2.0 DOM tests?

>   - Some early experiences with Itedo's code base shows some problems
>   with the WebCGM 2.0 'onload' param when the <object> tag is
>   dynamically inserted. This may be specific to our implementation,

Do you have a simple example of a fix that you think should work, but that 
turns out to be problematic?  (The others could play with it.)

>   but if not; the group needs to take some decisions about the
>   'onload' param.
>   It would be good if vendors/users of this mailing list could
>   collaborate about this as it affects all of us.


As a first priority, and to get discussion started, what is the shortest 
path to making the Test Suite viable again (even if there are some issues 
like the 'onload' PARAM.)


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]