launch applet with APPLET or PLUGIN

[Andrew Thompson]

BLAH! I did it!

I was experimenting with your example and realised that
if you nest the elements the other way around, both IE
*and* Opera use the <OBJECT> element with classid.

It is *only* the Mozilla browsers which (apparently
intentionally) ignore the <OBJECT> element where classid
is specified.

Further, since Opera (correctly) ignores the inner <OBJECT>
element, it is only IE that actually loads the object twice,
and oddly, if the second applet is not hidden, IE does not
load a second applet, but instead shows a ..'text input'
type box (go figure?!?).

The second <OBJECT> element can be hidden from IE using
*either* the CSS you showed (or a slight variant of the
technique for the reordered elements), or by using IE
conditional comments.

The latter is not valid, but might be worth considering as well.

The most astounding thing about nesting it the other way
is that I managed to load this variant of the page in NN 4.8.
Since it specified a post 1.1 Java, ..the browser actually
produced an 'Install Plug-In' dialog.

I did not risk actually loading Java 1.5 in NN 4.8, but
it was an improvement over the first examples that caused
it to immediately crash. Instead, ..now it crashes when
you *leave* the page. ;-)

Since Dag's id correction variant seems equally applicable
to both techniques, I have not yet incorporated it.

Here are my latest experiments..
<http://www.physci.org/test/appletcall/index8.html>

 

[Andrew Thompson]

Since Dag's id correction variant seems equally applicable
to both techniques, I have not yet incorporated it.

That *should* have read..

Since Dag's id correction technique seems equally applicable
to both variants (of hiding an applet from IE), I have not
yet incorporated it.
....

 

[Chris Head]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I was experimenting with your example and realised that
if you nest the elements the other way around, both IE
*and* Opera use the <OBJECT> element with classid.


It is *only* the Mozilla browsers which (apparently
intentionally) ignore the <OBJECT> element where classid
is specified.

OK. Ideally all browsers would use the Mozilla solution, since MIME
types (in my mind at least) are far better than class IDs.

Further, since Opera (correctly) ignores the inner <OBJECT>
element, it is only IE that actually loads the object twice,
and oddly, if the second applet is not hidden, IE does not
load a second applet, but instead shows a ..'text input'
type box (go figure?!?).

Strange... in my tests, I got a "broken image" icon, not a text input box.

The second <OBJECT> element can be hidden from IE using
*either* the CSS you showed (or a slight variant of the
technique for the reordered elements), or by using IE
conditional comments.

The latter is not valid, but might be worth considering as well.

I just looked up what conditional comments are. I had never heard of
them until now.

The most astounding thing about nesting it the other way
is that I managed to load this variant of the page in NN 4.8.
Since it specified a post 1.1 Java, ..the browser actually
produced an 'Install Plug-In' dialog.
Wow...


I did not risk actually loading Java 1.5 in NN 4.8, but
it was an improvement over the first examples that caused
it to immediately crash. Instead, ..now it crashes when
you *leave* the page. ;-)

NOT so wow

Since Dag's id correction variant seems equally applicable
to both techniques, I have not yet incorporated it.

Here are my latest experiments..
<http://www.physci.org/test/appletcall/index8.html>

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (MingW32)

iD8DBQFDEhFU6ZGQ8LKA8nwRAmDfAJ9UCaxvZNfQSPgQTrf8n8oLYMiX2gCgpizN
GiFjJ45qg3JLEE7MNNqnzJw=
=nHzz
-----END PGP SIGNATURE-----