9 reasons why Sun should provide JWS installers for extension APIs
The JOGL API is complex, but there is a lot of development
and deployment of JOGL based projects, and part of the
reason is the ease with which JOGL based projects can
be deployed, simply by refering to your own projects
classes in your JNLP, and the JOGL JNLP extension.
Even better, it is possible to deploy unsigned
'sand boxed' JOGL based applications, while the
JOGL code itself, is singed by 'Sun Microsystems',
rather than 'Some Random Developer With A
Self-Signed Certificate'
API's such JAI, JavaMail & JavaComm should be provided
as web-start based component-desc extensions. This
makes a lot of sense for Sun, developers, deployers
and end users.
For Developers
1) Seeing the ease of installation of other projects
based around web-start, encourages developers that
the deployment of their treasured project will be easy.
2) It saves each developer from having to work out the
intricacies of installation (and possible configuration)
of the extension, since this can be achieved once, in
the extension installer (if required).
For Deployers
3) Less files for their server to host or distribute.
For End Users
4) A more reliable installation (than the average
developer can manage) with a more standardised and
recognisable install/uninstall/management mechanism
(through the Java Control Panel, and in Windows at
least, the program 'add/remove' area of the settings.)
5) Greater variety of software. This comes indirectly
from the benefits to developers and deployers - ease
of devlopement and deployment leads to mroe projects.
6) Lower total download size. If the end user is
interested in advanced image manipulation, and installs
10 image editors - three of which are Java based,
they will probably download the JAI packages three
times. With JWS launch from a common site, the end
user should only need to donwload the JAI once.
For Sun
7) Promotion (via use) of Sun API's, & Java.
8) Ease of monitoring the 'most installed' extension
packages, and probably a more accurate indication
of the actual need for/level of use of each API,
better than most other statistics.
9) Ease of maintenance and update of the underlying
API - upload the new Jar's - everybody upgrades.
Do people here support this idea?
(Would this get your vote as a 'request for
enhancement' in the bug database?)
Are there other good reasons I missed?
