This is another reason why a cookbook approach might work well. I envision the cookbook as a collection of examples vs. packaged solutions - lower expectations for what the code will do, and less to maintain.
On Feb 23, 2010, at 8:41 AM, Martijn Dashorst wrote: > With wicket we have a sf.net project called Wicket Stuff (which was > already in place before we joined Apache). This is the place where we > invite anyone from our community to publish their work: components, > integrations etc. This place was also created to host non-compatible > code (Hibernate, and other GPL/LGPL/etc libraries). > > The Wicket Stuff project is successful in that it attracted a big > number of contributors, but IMO it failed to get the necessary quality > due to lack of interest of the same contributors. Many of the projects > are unsupported and never had a proper release, let alone some > documentation. I see the Wicket Stuff project as a blessing and a > curse: a great place to find components and integrations, but also a > place where things are half finished, and that reflects on the > original product as well. > > The typical behavior we see is that someone commits a first draft of a > component they've created. They might even hook it up to the standard > infrastructure (multimodule Maven project). Then they leave. Next > someone wants to use the project, finds a bug or missing feature, > commits their solution and moves on. No releases, no finished product, > and that is IMO problematic in the eyes of folks that want to write > production ready code: they want releases, changelogs, release notes > and some evidence that someone is maintaining it. > > Be prepared to have to heavily invest in such a second leg if you want > to use it as an attractor for 'enterprisy developers'. > > Martijn > > On Tue, Feb 23, 2010 at 2:12 PM, Noel Grandin <[email protected]> wrote: >> >> The SwingX project has a nice system where people put their own >> not-ready-for-platform code into something like >> >> /trunk/users-dev/noelgrandin/src/MyWeirdWidget.java >> >> Then it becomes visible to everybody and people can comment and improve >> on each other's code. >> Quite a lot of stuff ended up making its way into the mainline like that. >> >> -- Noel >> >> On 2010-02-23 14:29, Todd Volkert wrote: >>> That's a good point Noel. I actually created >>> http://code.google.com/p/pivot-contrib/ a couple of weeks ago just to backup >>> a layout container that I was working on that was too app-specific to be in >>> the platform. I don't like the idea of putting stuff like that >>> (app-specific add-ons) in the platform, but there's no legal reason to put >>> it on Google Code either... it's almost as if we could create a separate >>> hierarchy in SVN that lived off the trunk and never got released where we >>> could put stuff like this. Then if newbies wanted widget X, and someone had >>> built it before in this playground, we could just point the newbie there, >>> and they could fork it and build it themselves. >>> >>> I know Niclas created "skunk" as a sibling to "trunk", but I'm not sure >>> that's appropriate, as I think it was meant for experimental features that >>> may be included in the trunk some day... >>> >>> In any case, what do others think? >>> -T >>> >>> On Tue, Feb 23, 2010 at 3:19 AM, Noel Grandin <[email protected]> wrote: >>> >>> >>>> Hi >>>> >>>> Pivot is still showing normal early stage adoption i.e. a trickle of >>>> interest. These things tend to ramp in bursts, so my counsel would >>>> simply be to be patient. >>>> >>>> I agree on the SWT issue - I don't think porting Pivot to SWT would >>>> improve adoption. SWT already has JFace, Nebula and various other >>>> additional widget libraries. >>>> >>>> On the other hand, Pivot is a great example of how good Swing could be >>>> if it was allowed to evolve :-) >>>> >>>> I do notice that we're getting various conversations along the lines of >>>> >>>> Newbie: "X is very easy to do with toolkit Y" >>>> Pivot-guru: "You could implement X on top of component C" >>>> Newbie: "That's too hard! Can't you just add it?" >>>> Pivot-guru: "Adding that feature doesn't really fit into our architecture" >>>> >>>> Which is reasonable, but maybe we should be implementing these features >>>> in some kind of extras package until we have a good enough idea of how >>>> to fit the features into the main codebase? >>>> >>>> -- Noel >>>> >>>> On 2010-02-23 00:14, Greg Brown wrote: >>>> >>>>> Though we have only gotten two responses on the SWT question, it seems as >>>>> >>>> though an SWT port may not be the best way to move Pivot forward. Michael >>>> made some great suggestions. What do others think? What can we do to help >>>> raise awareness of and interest in Pivot as a viable alternative to other >>>> Java-based UI technologies? >>>> >>>>> >>>>> >>>> >>>> >>> >> >> > > > > -- > Become a Wicket expert, learn from the best: http://wicketinaction.com > Apache Wicket 1.4 increases type safety for web applications > Get it now: http://www.apache.org/dyn/closer.cgi/wicket/1.4.4
