Il 28/12/2009 06:53, Kalle Korhonen ha scritto: > Hey Alessandro, I have to say I agree with pretty much everything > you've said. I notice the tone from the committers is a bit "fatherly" > - i.e. "just learn the framework and do it yourself, it's easy". Don't > worry about it, sometimes it's just difficult to see it from newbies' > perspective for those who are experienced with a particular > technology.
Do not worry about this topic, Kalle. I understand very well their position. If I had a good knowledge of Java and Tapestry, I would just do what they suggest. I see that it is the "right way". I'm just trying to let them (and you) know how an absolute beginner feels when it meet Java and Tapestry for the first time. I see it is hard to explain my position, even using examples (Drupal) and metaphores. Nevertheless, I do think the Java/Tapestry community have to face a few of the challenges I'm describing here. See below for the reasons. > I think your comparison to Drupal is spot on. Sure it's > uglier to customize Drupal if it doesn't have a module available for > it but it does heck of a lot of things out of the box, much more than > Tapestry or any little add-on module to it at the moment. Tapestry > architecture is great for delivering drop-in add-on modules, but what > you see is just the state of things as they are right now. Tapestry5 > is still a new web framework and there are relatively few people using > it. Yes, you are right. Tapestry and Tynamo are still new and we have to give them the time to grow up. I do hope to can be of some help in the future. As you said, Tapestry is very good at delivering a reliable drop-in module technolgy and this feature could/should be used to develop an even more developer-friendly tool, like Tynamo. Arguably, this is the right way to evolve it and I see you are already doing so. > I fully agree that security is critical - sure it's easy to > protect your pages "the right way" with a few annotations, but real > world projects need integration to existing authentication sources, > smart remember me -implementations, url protection, adaptive > permissions etc. You have got the point. Actually, I would be happy to implement a annotation-based security system if I just know how to do it (BTW: any docu around?). Unfortunately, this security system will also have to be compatible with a few other tools (in our case, at least with a legacy OpenLDAP-based directory). Maybe I'm wrong but I still have the feeling that using a widely diffused tool, like ACEGI, could be a simpler and better way to get this job done. > Regarding Tynamo, in time we'll offer proper Apache > Shiro integration (currently there are a few good "starter" Shiro > integration libraries available) with most of these features but we > are not there yet. But in any case, being an Apache Shiro committer > myself I should be in a pretty good position to deliver it. Good news to hear! :-) > One of the main issue with Java web frameworks has always been that > while people generally acknowledge the advantages of Java, starting > costs are so much higher than hacking up something with a few lines of > PHP and its existing frameworks already deployed by ISPs that Java is > often considered only later when (and if ever!) complexity, > scalability and maintainability become more of a concern. That is exactly what happend to us. We decided to pass to Java (and to a few Java-based technologies) only when Python, PHP and other tools begun to show their limits. It was much more a matter of scalability/manageability than a matter of language expressiveness. You can do with Python, Ruby or PHP everything that Java (or even Groovy and Scala) can do /but/ you cannot have anything like the smooth development process allowed by a few "modules" of Java (Hibernate, ACEGI, etc) and by its process management tools (Maven, Hudson, etc.). We knew (and still feel on our skin) that the Java world is a complex one. You have to study and dominate the basic language (that is the easy part of the job, coming for Python), a few programming paradigms like IoC and AOP, a few tools (Maven, etc), and finally a few frameworks (Tapestry, the ubiquitous Spring and Hibernate, etc.). It is a lot of stuff and a lot of work. This was the reason why we looked for a good framework to be used as a starting point. Of course, the framework has to supply us with a good set of tools and has to act as a set of reliable guidelines, as well. > However, > Java-based frameworks and tools have made steady progress in recent > years and lowered the starting costs to reduce the gap. Tapestry can > be and has been the forerunner in Java Web frameworks, now we just > need to show the world all that is possible - and especially *how* to > do it. That is exactly the way things are! Java do supply the developer with an impressive set of impressive tools but the entry ticket is still very high (not only for web developers, unfortunately. Just ask to Swing and SWT news users...). Frameworks like Tapestry and Tynamo are what is needed to lower this ticket (and to make the dev process even more smoot and more manageable). Unfortunately, it is not enough to have a good web framework like Tapestry. The new Java user needs a wide range of tools (resulting in a full-stack framework, eventually) and a reliable set of guidelines, as well. Anyway, I see that you and the other guys who are working on Tynamo have already understand all of my "complains" so I'm confident we will have a good full-stack framework (or whatever you wants to call it) in the near future. As I said before, it is just a matter of time. In the meantime, I will keep on studing Tapestry and its underling technologies. CU -- Alessandro Bottoni Website: http://www.alessandrobottoni.it/ Government big enough to supply everything...is big enough to take everything you have. The course of history shows that as a government grows, liberty decreases. -- Thomas Jefferson
--------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org For additional commands, e-mail: users-h...@tapestry.apache.org