Re: RE: NN fragment -- bad news?

[Lofton Henderson <lofton@rockynet.com>]

In any case, it seems that the previous assessment of "no problem" with NN 
4.73 was wrong.

So the real question:  what is the way forward?  It seems to me that Dave 
maybe ought to reconvene the task team and start with a telcon to sort out 
what to do about this?  Netscape's lack of concern about the problem is 
more serious now (because 4.73 is affected, as well as the somewhat shakey 
new 6.0 release), and I think we need to attack as energetically it as we 
did IE.

Other thoughts?

-Lofton.

At 09:19 AM 2/6/01 +0100, Dieter Weidenbrueck wrote:
>All,
>
>I agree with Greg here. However, we still need to be able to insert a simple
>HREF pointing to an object inside a WebCGM file into HTML. To load a file
>explicitely as described by Greg you would always need a javascript.
>
>Dieter
>
>----- Original Message -----
>From: "Lofton Henderson" <lofton@rockynet.com>
>To: <cgmopen-members@lists.oasis-open.org>
>Cc: <gcg@micrografx.com>
>Sent: Monday, February 05, 2001 9:00 PM
>Subject: Fwd: RE: NN fragment -- bad news?
>
>
> > CGM Open Members,
> >
> > Greg's reply reinforces my evidence that NN 4.73 (as well as 6.0) indeed
> > has the fragment handling bug.
> >
> > His suggestions are interesting as well.  Comments?
> >
> > Regards,
> > Lofton.
> >
> > >From: Greg Gallant <gcg@micrografx.com>
> > >To: "'Lofton Henderson'" <lofton@rockynet.com>
> > >Subject: RE: NN fragment -- bad news?
> > >Date: Fri, 2 Feb 2001 16:10:03 -0600
> > >X-Mailer: Internet Mail Service (5.5.2650.21)
> > >X-RCPT-TO: <lofton@rockynet.com>
> > >
> > >[...] In our initial testing with NN, we found
> > >irregularities with the fragment.  I'm told these irregularities persist
> > >through version 6.  [...]  We have not developed a workaround for NN.
> > >
> > >One thing you all might consider is standardizing upon direct interfaces
> > >that WebCGM plugins support in a standardized way -- this would be via
> > >ActiveX for IE and via LiveConnect for Netscape.  If you have standard
>ways
> > >to "talk" to all plugins you remove yourself from being at the mercy of
>the
> > >browser developers who are always likely to break behavior in the next
> > >version.  Further, loading a page directly into a plugin is far more
> > >efficient then having the browser get a page, determine the mime type
>from
> > >the HTTP header, then hand the URI over to the plugin for the page to be
> > >retrieved again.  All those socket opens and closes are expensive.
> > >
> > >-gcg
> >