Re: T5: Will it be a full stack framework?
Todd Orr wrote: [] More to the point, I believe that Howard and the Tapestry developers will be utilized as development resources most effectively if the focus remains on creating the #1 MVC framework. Thanks a million guys! Just to continue arguing in this way : I use Tapestry 5 to build an LDAP manager web application, and I really don't care to have a full stack framework for that, because this kind of framework not event know what LDAP is... As soon as you step aside from the most common use cases, these framework are useless. In fact, I choose Tapestry 5 because it does it job _very_ well, but it just do it (there was other pretenders, but certainly not SEAM, RIFE or *rails). - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: T5: Will it be a full stack framework?
Francois Armand wrote: Todd Orr wrote: [] More to the point, I believe that Howard and the Tapestry developers will be utilized as development resources most effectively if the focus remains on creating the #1 MVC framework. Thanks a million guys! Just to continue arguing in this way : I use Tapestry 5 to build an LDAP manager web application, and I really don't care to have a full stack framework for that, because this kind of framework not event know what LDAP is... As soon as you step aside from the most common use cases, these framework are useless. In fact, I choose Tapestry 5 because it does it job _very_ well, but it just do it (there was other pretenders, but certainly not SEAM, RIFE or *rails). I certainly don't think that tapestry-core should ever be a full stack, but I hope that there is eventually something like the quick-start which allows new users to generate a quick and dirty CRUD application very easily (it would be nice it this were less typing than the standard maven command line as this is off-putting to most new comers, but copy-in-paste does work wonders), no one should expect such quick-starts to be full-strength applications, but it would provide a great way to get someone new (like myself) up-to-speed and productive. The power of annotations would seem to be to enable this via IoC without dumping any additional logic into the tapestry-core code base. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: T5: Will it be a full stack framework?
Robert Sanders wrote: I certainly don't think that tapestry-core should ever be a full stack, but I hope that there is eventually something like the quick-start which allows new users to generate a quick and dirty CRUD application very easily (it would be nice it this were less typing than the standard maven command line as this is off-putting to most new comers, but copy-in-paste does work wonders), no one should expect such quick-starts to be full-strength applications, but it would provide a great way to get someone new (like myself) up-to-speed and productive. The power of annotations would seem to be to enable this via IoC without dumping any additional logic into the tapestry-core code base. Oh, ok :) My personal thought are that this kind of application will come to life in a near future... It already exists Trails (http://www.trailsframework.org) for Tapestry 4, and well... Tapestry 5 seems to be so pleasant and so powerful... Just need to be done by someone ;) - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: T5: Will it be a full stack framework?
On 6/28/07, Francois Armand [EMAIL PROTECTED] wrote: My personal thought are that this kind of application will come to life in a near future... It already exists Trails (http://www.trailsframework.org) for Tapestry 4, and well... Tapestry 5 seems to be so pleasant and so powerful... Just need to be done by someone ;) Tapestry 4 isn't far behind in powerfulness and in some areas way ahead, but certainly with a higher learning curve. Supporting Tap5 is a constant topic within Trails, but the reality is that it's a preview and will take some time to reach the maturity level of Tap4.x. In the meantime though, we are hard at work for upgrading to Tap4.1.2 in Trails. We'll evaluate Tap5 again when a final release comes closer. Kalle
Re: T5: Will it be a full stack framework?
Kalle Korhonen wrote: Tapestry 4 isn't far behind in powerfulness and in some areas way ahead, but certainly with a higher learning curve. Well, I'm just a beginner with Tapestry 5, and don't even had a look to Tapestry 4, so my notice was just about the thrill I have to use T5 :) (I'm coming from a world of Struts and even worse home-made framework... just imagine that) Supporting Tap5 is a constant topic within Trails, but the reality is that it's a preview and will take some time to reach the maturity level of Tap4.x. That's clear, and your position seems to be quite sensible. In the meantime though, we are hard at work for upgrading to Tap4.1.2 in Trails. We'll evaluate Tap5 again when a final release comes closer. That sound great, thanks for your work and for you reply ! - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: T5: Will it be a full stack framework?
On 6/28/07, Todd Orr [EMAIL PROTECTED] wrote: Transactions declared in the Web tier? Oy vey. What's next, Tapestry-SOAP??? Come on. Don't be shortsighted. I managed to add a service layer and bind it in the AppModule. I was thinking to have declarative transaction each time the web layer invoke the method in this service layer. -- Let's create a highly maintainable and efficient code YM!: thejavafreak Blog: http://www.nagasakti.or.id/roller/joshua/ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: T5: Will it be a full stack framework?
Come on. Don't be shortsighted. I managed to add a service layer and bind it in the AppModule. I was thinking to have declarative transaction each time the web layer invoke the method in this service layer. Whatever works for you. Every project has its own nuances. For me, transactions are not a concern of the Web teir. In my ideal development world, concerns are separated strictly. I feel as though if I do not separate my concerns then I'm cutting a corner. But, this is only my opinion and its based on a perspective of development that I have. When I hear about transactional demarcation in the web tier I immediately think, what if I want to swap my service implementation with a WS call? The TX has to be thought of completely differently in this respect. Without WS-TX the usual route is to go with course-grained methods that apply a TX around the backend logic, or something of that sort. In any case, the web tier has no say in the TX. I'm not saying everyone should think this way. It is the way I have come to think because my exposure to certain architectures. Consider one such that core services are exposed via SOAP and orchestrated via ESB. I have no direct interaction with the code, per se. This influences me because now when I build anything I think in highly decoupled terms - my Tap app should not know anything about the services it uses, not even whether there is a transactional state to manage. Honestly, this makes the development of a site more difficult - it's harder to build this way. But the gains I receive in the overall architecture, scalability, as well as the development lifecycle impact, are suitably large to make it worthwhile. With regards to implementing this in Tapestry, I can't say it's bad in all cases. I can say that there are more [easily] scalable solutions that have more clearly defined architectural borders. This is the reason I wouldn't use this feature. I know that having this clear separation allows me to modify my service layer without having to alter my web layer in any way. I can move development ahead on many fronts between many teams with minimal impact between the teams. This isn't everyone's goal. All that aside, this is a moot point. Tapestry core will remain focused as Howard says. I get to have my cake, as do you, and we can both eat it. On 6/28/07, Joshua Jackson [EMAIL PROTECTED] wrote: On 6/28/07, Todd Orr [EMAIL PROTECTED] wrote: Transactions declared in the Web tier? Oy vey. What's next, Tapestry-SOAP??? Come on. Don't be shortsighted. I managed to add a service layer and bind it in the AppModule. I was thinking to have declarative transaction each time the web layer invoke the method in this service layer. -- Let's create a highly maintainable and efficient code YM!: thejavafreak Blog: http://www.nagasakti.or.id/roller/joshua/ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: T5: Will it be a full stack framework?
On 6/27/07, Ivan Dubrov [EMAIL PROTECTED] wrote: There is some chaotic code which is available in the SVN repo here: http://atlas.wfrag.org/svn/trunk/atlas-security/ Ah, real nice, I was working on this as well, but I've just got time to work on it and now I don't have to. :) Can you put this together as a maven package or something? It adds support for @Secured for the pages and components. However, it does not work for listener methods since I didn't found a way to extend method with around code (which is executed at the beginning and at the end.). Only after was available (ClassTransformation#extendMethod). Maybe Howard is on this one already? -- regards, Robin - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: T5: Will it be a full stack framework?
Robin Ericsson wrote: Ah, real nice, I was working on this as well, but I've just got time to work on it and now I don't have to. :) Can you put this together as a maven package or something? Of course. I thought making something like tapestry-security module, but now I'm completely out of time to do it. -- WBR, Ivan S. Dubrov signature.asc Description: OpenPGP digital signature
Re: T5: Will it be a full stack framework?
On 6/27/07, Ivan Dubrov [EMAIL PROTECTED] wrote: Of course. I thought making something like tapestry-security module, but now I'm completely out of time to do it. Ok, I've checked out the code from you SVN so I'll poke around a bit. Anything specific that's not working? Listerner methods isn't a big deal for me at the moment :) -- regards, Robin - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: T5: Will it be a full stack framework?
Robin Ericsson wrote: Ok, I've checked out the code from you SVN so I'll poke around a bit. Anything specific that's not working? Listerner methods isn't a big deal for me at the moment :) It works basically, but the IoC configuration code is not very flexible (since it was developed for concrete case). Some efforts are needed to make this code more generic (like tapestry-acegi for T4), so it could be configured for different cases. -- WBR, Ivan S. Dubrov signature.asc Description: OpenPGP digital signature
Re: T5: Will it be a full stack framework?
Regarding the original poster's statements: I think that Tapestry is so great because it does the job of an MVC really well. I hope that it doesn't become a full stack solution. I think it would be more beneficial to the community to ensure that Tapestry's focus is the MVC. Additional functionality should be brought on through modules as some already are. Default security and transaction schemes will only become monsters of their own with devs requesting enhancements until quite a bit of Tapestry's development focus is shifted onto various sub-projects. And really, with all the fantastic frameworks out there - that are as good as they are because of focus - why carry the baggage of some interim scheme for developers not wanting to perform the minimal legwork of integration themselves? I think projects like tapestry-hibernate are nice because they eliminate common integration problems. But its job is to provide enhanced integration, not provide persistence for a Tapestry app. Personally, I don't use tapestry-hibernate because I am still one of those relics that believes having a properly tiered application is good architecture. Why should Tapestry care what persistence strategy I am using? Having said that, this module is nice for the quick app that doesn't need or desire the type of benefits that come with tiered development and greatly helps to integrate hibernate. More to the point, I believe that Howard and the Tapestry developers will be utilized as development resources most effectively if the focus remains on creating the #1 MVC framework. Thanks a million guys! Transactions declared in the Web tier? Oy vey. What's next, Tapestry-SOAP??? On 6/26/07, Joshua Jackson [EMAIL PROTECTED] wrote: Dear all, I'm quite happy currently trying out T5 now. And I really like the T5 IoC concept compared to other IoC container. - I really like how it has (or will have) built in AJAX components. - I really like how easy it is to create custom components compared to other component based web framework. - I also like the integration with hibernate. It's just too damn simple and works out of the box. - I really like the GridPager component (though it got me struggle to work it out :P). I don't know any other web framework that has built in pager component. Everything just seems to be so natural and how a web framework should have been. But I have a questions here. With the so complete features that I have just mentioned, will Tapestry 5 have the vision to be a full stack framework in the future? Some features that I think will be great to be added if T5 will be a full stack framework is: - Built in security or user authentication and authorization. I must admit that using Acegi for the security is just unnecessary because I will need Spring just to get Acegi running. And I will have 2 IoC container in my apps. I think T5 IoC already have the features (like the decorator) to make security feature made available. - Built in declarative transaction, since there's already tapestry-hibernate module. I don't know whether this is possible, but when I read the T5 IoC decorator feature, I think I can start and end transaction before and after I invoke a method. Is it possible? Well that's my opinion and some ideas I came up with. Hopefully it can come true. And thank you so much for making T5. I can see that it will be a great framework in the future. Keep up the good work guys. Hats off for you people -- Let's create a highly maintainable and efficient code YM!: thejavafreak Blog: http://www.nagasakti.or.id/roller/joshua/ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
T5: Will it be a full stack framework?
Dear all, I'm quite happy currently trying out T5 now. And I really like the T5 IoC concept compared to other IoC container. - I really like how it has (or will have) built in AJAX components. - I really like how easy it is to create custom components compared to other component based web framework. - I also like the integration with hibernate. It's just too damn simple and works out of the box. - I really like the GridPager component (though it got me struggle to work it out :P). I don't know any other web framework that has built in pager component. Everything just seems to be so natural and how a web framework should have been. But I have a questions here. With the so complete features that I have just mentioned, will Tapestry 5 have the vision to be a full stack framework in the future? Some features that I think will be great to be added if T5 will be a full stack framework is: - Built in security or user authentication and authorization. I must admit that using Acegi for the security is just unnecessary because I will need Spring just to get Acegi running. And I will have 2 IoC container in my apps. I think T5 IoC already have the features (like the decorator) to make security feature made available. - Built in declarative transaction, since there's already tapestry-hibernate module. I don't know whether this is possible, but when I read the T5 IoC decorator feature, I think I can start and end transaction before and after I invoke a method. Is it possible? Well that's my opinion and some ideas I came up with. Hopefully it can come true. And thank you so much for making T5. I can see that it will be a great framework in the future. Keep up the good work guys. Hats off for you people -- Let's create a highly maintainable and efficient code YM!: thejavafreak Blog: http://www.nagasakti.or.id/roller/joshua/ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: T5: Will it be a full stack framework?
Not really sure how Howard feels regarding the whole security question, but for my money I've seen the built-in security framework idea blow up again and again. Most of the company's I've worked for already have security frameworks in place that their applications simply have to use. Therefore trying to juggle between a framework's built-in security versus that of a proprietary internal one can be a nightmare. I don't disagree that default implementations or schemes should be made available, but this is certainly a feature no one should be forced to use or even download. To me, it would be better if this were a separate module that wrapped the existing micro-kernel and provides a wall through which all communications must pass. Otherwise, any bug fixes to the security can have serious implications on the product's release. Likewise, built-in security generally means more unmaintainable code when coupled with the container. Just my two cents. Cheers. On 6/26/07, Joshua Jackson [EMAIL PROTECTED] wrote: Dear all, I'm quite happy currently trying out T5 now. And I really like the T5 IoC concept compared to other IoC container. - I really like how it has (or will have) built in AJAX components. - I really like how easy it is to create custom components compared to other component based web framework. - I also like the integration with hibernate. It's just too damn simple and works out of the box. - I really like the GridPager component (though it got me struggle to work it out :P). I don't know any other web framework that has built in pager component. Everything just seems to be so natural and how a web framework should have been. But I have a questions here. With the so complete features that I have just mentioned, will Tapestry 5 have the vision to be a full stack framework in the future? Some features that I think will be great to be added if T5 will be a full stack framework is: - Built in security or user authentication and authorization. I must admit that using Acegi for the security is just unnecessary because I will need Spring just to get Acegi running. And I will have 2 IoC container in my apps. I think T5 IoC already have the features (like the decorator) to make security feature made available. - Built in declarative transaction, since there's already tapestry-hibernate module. I don't know whether this is possible, but when I read the T5 IoC decorator feature, I think I can start and end transaction before and after I invoke a method. Is it possible? Well that's my opinion and some ideas I came up with. Hopefully it can come true. And thank you so much for making T5. I can see that it will be a great framework in the future. Keep up the good work guys. Hats off for you people -- Let's create a highly maintainable and efficient code YM!: thejavafreak Blog: http://www.nagasakti.or.id/roller/joshua/ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- Gotta find my destiny, before it gets too late.-- Ian Curtis
Re: T5: Will it be a full stack framework?
Joshua Jackson wrote: - Built in security or user authentication and authorization. I must admit that using Acegi for the security is just unnecessary because I will need Spring just to get Acegi running. And I will have 2 IoC container in my apps. I think T5 IoC already have the features (like the decorator) to make security feature made available. In fact, you do not need Spring IoC to get Acegi running (you can configure Acegi beans using Tapestry IoC). However, Acegi requires a number of utility classes from Spring, so there will be about 1Mb or so of Spring dependencies. -- WBR, Ivan S. Dubrov signature.asc Description: OpenPGP digital signature
Re: T5: Will it be a full stack framework?
Good to know. I have Acegi kind of working using Spring IoC. Didn't think I could use Tapestry's IoC for it. If you have any sample code WRT this, it would be nice to see! Daniel On 6/26/07, Ivan Dubrov [EMAIL PROTECTED] wrote: Joshua Jackson wrote: - Built in security or user authentication and authorization. I must admit that using Acegi for the security is just unnecessary because I will need Spring just to get Acegi running. And I will have 2 IoC container in my apps. I think T5 IoC already have the features (like the decorator) to make security feature made available. In fact, you do not need Spring IoC to get Acegi running (you can configure Acegi beans using Tapestry IoC). However, Acegi requires a number of utility classes from Spring, so there will be about 1Mb or so of Spring dependencies. -- WBR, Ivan S. Dubrov - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: T5: Will it be a full stack framework?
There is some chaotic code which is available in the SVN repo here: http://atlas.wfrag.org/svn/trunk/atlas-security/ It adds support for @Secured for the pages and components. However, it does not work for listener methods since I didn't found a way to extend method with around code (which is executed at the beginning and at the end.). Only after was available (ClassTransformation#extendMethod). Daniel Jue wrote: Good to know. I have Acegi kind of working using Spring IoC. Didn't think I could use Tapestry's IoC for it. If you have any sample code WRT this, it would be nice to see! Daniel On 6/26/07, Ivan Dubrov [EMAIL PROTECTED] wrote: Joshua Jackson wrote: - Built in security or user authentication and authorization. I must admit that using Acegi for the security is just unnecessary because I will need Spring just to get Acegi running. And I will have 2 IoC container in my apps. I think T5 IoC already have the features (like the decorator) to make security feature made available. In fact, you do not need Spring IoC to get Acegi running (you can configure Acegi beans using Tapestry IoC). However, Acegi requires a number of utility classes from Spring, so there will be about 1Mb or so of Spring dependencies. -- WBR, Ivan S. Dubrov - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- WBR, Ivan S. Dubrov signature.asc Description: OpenPGP digital signature