A Wicket Diary

I chickened out at the blogger meetup: I didn’t want to end up like Bileblog Hani just before my session on JavaOne, so I left when they started issuing car bombs. Sorry Matt!

All in all it was very interesting to meet the bloggers out there in person and it was pretty neat to meet Matt Raible, even though he can be quite annoying ;-). My esteemed collegue Eelco Hillenius is heading out right now to meet them and hold up the Wicket honour (I hope Matt forgot to bring his camera).

Just finished a rehearsal of my presentation and it went pretty well. Speaker coaching is really good here and they make you feel much more up to the task of addressing all you guys :-). I have no doubt it will be a fun and interesting experience.

Now my last problem (I hope) is to find my co-speaker, Miko, and go through the last stuff. Does it effect your karma to meet with your cos-speaker, just 15 minutes before the talk?

Our media campaign has been quite successful, although the hits do need to get going. We’ve been all over the international press:

• Japan
• Germany
• The Netherlands
• USA - infoworld
• Newsforge
• O’Reilly OSDir
• Australia Computerworld
• IT World
• Webhost magazine
• JDJ magazine

Or you could Google it yourself.

Howard Lewis Ship has commented on the release announcement of Wicket 1.0 at the serverside and Jonathan Locke responded to that. Here’s my addendum to those comments:

First I thank you, Howard, for your interest in Wicket. I hope to meet you and share a (couple of) beers at JavaOne.

At the end of the day, it’s great to see folks “gunning” for Tapestry, it’s a kind of validation.

Even though most people will state that we are ‘attacking’ Tapestry, the reality is different, if you ask me. Both Tapestry and Wicket are aiming at the same group: the developers currently working with Model2 frameworks. In a sense we are competing (which is only good according to many people) since we are both fishing in the same pond. We are not aiming to take over any Tapestry developer, even though I suspect there will be some that will cross over. We are hoping to lure model 2 developers and in greater quantities than Tapestry and JSF

I’m concerned that each session will get a giant serialized tree of components, something that Tapestry’s structure works exceptionally hard to avoid

With what I’ve come to understand of Tapestry is that in order to support page pooling, you have to use the dynamically generated properties, or reset your page at the end of the request into a virgin state. Putting this responsibility on the shoulders of the developer makes it rather complex and error prone to build applications using Tapestry development.

Furthermore, I read somewhere (I can’t seem to find the quote) that it is very easy to slap 4GB of memory in a server, and when you’ve filled that up with sessions, lack of processing power and bandwidth problems have become your closest friends.

For failover clustering this will be a problem, but Wicket does or will (with 1.1) allow you to control maximally what enters up in your session. I think however that the main bulk of development doesn’t have the requirement to do failover clustering, so I think the choice to enable quick and easy application development is correct. When the going gets tough and the session size becomes a problem Wicket allows you to tweak that.

I hope to meet you on JavaOne and share some insights (I’d love to hear more about Hivemind, as I want to use it in one of our projects).

Wicket is featured on the frontpage of The Serverside.com and we already receive some attention with the usual questions.