I agree that this approach would avoid network dependency but I'm not as convinced that network access presents a problem. Most of the installers I've used lately depend on network access in one way or another. I also suspect that users will customize their server right after downloading it.
Besides a smaller download size, the plugin approach has the advantage that the CLI and admin console are already available for driving the server customization process. Should we invest additional effort in providing the user with another way to achieve effectively the same results? If so then will it be clear to users when to use the plugin installer vs. when to use this alternative mechanism? One other factor to consider is that the "one big assembly" approach would only deactivate components when they are replaced and not actually uninstall them (at least if I understand your proposal correctly). If the deactivated components were later reactivated from the admin console or from an environmental dependency then the server could enter an unusable state. Best wishes, Paul On 2/10/07, Jason Dillon <[EMAIL PROTECTED]> wrote:
How big would one assembly be if we include *everything* like jetty, tomcat, axis2, cxf, everything. Not turned on though... then just provide people with a way to switch between personalities from the command line, and make one of them as default, so that if the server starts to boot up with no personality (hehe), then it will apply the default to itself when bootstrapping? Its probably gonna make the assembly zip a wee bit larger, but we'd only have one of em... so build time would be much faster, and if people want to try out different bits they don't have to redownload all that other stuff... but also, everything we need to make a javaee server is already in the assembly zip, so don't have to worry about networking muck to get the right personality up and running. Thoughts? --jason On Feb 9, 2007, at 12:32 PM, Paul McMahan wrote: > On 2/8/07, Jason Dillon <[EMAIL PROTECTED]> wrote: >> I'm definitely *NOT* in-favor of 8 assemblies. > > Ditto. Even if there was time and manpower to test every possible > assembly then I still don't think the end user would be prepared to > make an informed choice about which one to download. > >> On Feb 8, 2007, at 6:37 AM, Matt Hogstrom wrote: >> > If there is a plugin option then I think the TCK discussion becomes >> > simpler. Anyway, for those more skilled in that art than I what >> > are the community thoughts on how to address our expanding set of >> > pluggable components? > > I think that presenting the user with lots of choices is a good thing > if geronimo can : > 1.) provide a TCK tested default assembly > 2.) help users make informed decisions about changing the defaults > 3.) make it easy to enact their decisions > 4.) allow them to change their minds later > > With that in mind, I think the ideal scenario (from a user's > perspective) would be to provide one fully tested JEE5 assembly from > the download page and then make it easy to swap out components after > installation using plugins. Components that have passed the TCK in > any assembly can be marked as such in the plugin catalog, along with > any other useful information about that component such as which JEE > spec it implements, etc. Components that are mutually exclusive like > cxf and axis2, jetty and tomcat, etc can provide metadata that will > prompt the plugin system to uninstall the component that is being > replaced. > > There are lots of details and what-ifs that would need to be worked > out before this approach can be fully realized. But if there's > consensus around it then the next release could at least take a step > in the right direction. AFAIK most if not all of the necessary > functionality and infrastructure are already in place. > > Best wishes, > Paul
