> I'm not sure you could merge GF and WLS from a
> commercial stand-point. I believe the customer groups
> are quite different.
>
> Integrating Oracle app server and WLS was much easier
> in that regards, they both had a closed-source
> license + support business model.
As far as Oracle is concerned, they have the full copyright to Glassfish (since Sun does), so in that respect, it's just like a closed source code base. It happens to have a zillion copies floating about the internet, but they can fork it internally and do with it whatever they want.
> I think it's quite a bit more difficult to reconsile
> GF + WLS, I basically see two approaches:
> * Create two offerings one that is cheap / open
> source "low-end" and the other closed source
> "high-end" and let them live separate lives
IBM in theory kind of sorta of does this with Geronimo. Bundling it up and sticking a "Websphere Light" or some other nonsensical name on it. Or maybe it was a Tomcat/JMS server bundle they were pushing. I don't recall. No matter. Because while the code may well port, I doubt any of the administration would port, or deployment problems or other nuances.
I don't know about Oracle WLS versions, but, for example, you can download a full install of Oracle and use it for development. If they extend that to WLS, and make WLS free for development, there is little need in many shops for a "WLS Lite", and promoting GF in that role wouldn't make much sense since, essentially, nothing would carry over -- administration, deployment issues, etc.
> * Merge GF + WLS, though I don't see too much that GF
> can contribute that WLS doesn't already have ( at
> least in GF 2.X , GF 3 might have something to
> contribute, but I'm not sure exactly what it would
> be). Have two WLS offerings (as today) high / low
> end.
Well, GF v3 would have all of the EJB 3.1 glue code in it if nothing else. The amusing thing is that Oracle could use and incorporate that in to WLS whether they bought Sun or not.
> Either way they need to create enough differentiators
> between the offerings to not cannibalize sales which
> makes the first approach as bad as the next I'm
> afraid.
Yea, don't see much hope there.
> The sad part is that if GF had a year or two more to
> build a community, it might have had enough traction
> to continue, now I'm not too optimistic unfortunately
> :(
While GF's future may seem dim, I'm less concerned about the community. GF has been pretty good about being a true open source project, in terms of a lot of the development has been happening in the open, a lot of stuff is on public mailing lists, etc. Naturally there's back channel chatter among the devs we haven't seen, but that's just folks being folks.
GF v2.1 is rock solid, frankly, so I think that it will stay around a long time and hopefully enough for a new community to form around it to help maintain it.
v3, on the dark side, may get hit at the last minute depending on Oracles plans for it. I think they will see it through as the RI, but they can do that by focusing on spec compliance rather than features, tuning and testing.
On the bright side, tho, I think v3 will be easier to maintain and for other developers to pick up because of its new architecture and modularity. So I can see it being easier for a community to form around.
The real loss, tho, would simply be all the good folks at Sun. There's not a small amount of folks that are working on this. It's a good sized team. Plus they have folks that do a decent job at docs and what not, which many communities fall down upon (as developers notoriously don't do documentation well).
However, all that said, there IS a bunch of artifacts in place for a community to build around, they don't have to "create" anything so much, and simply keep the existing engine running as best it can: fixing bugs, incremental documentation additions, performance tuning anything glaring, etc. It's not like it would have to start from scratch, which is a great boon I think.
I imagine there are some internal tools that the devs may use that would potentially be useful to new maintainers, but I dunno if they could be released or not.
So, like I said, I don't think GF has a bright future at Oracle, but that doesn't mean the GF doesn't have a bright future. It's not some last minute "here we're open sourcing it, good night and good luck" code dump. It's a living project.
[Message sent by forum member 'whartung' (whartung)]
http://forums.java.net/jive/thread.jspa?messageID=343155