Read More: http://code.google.com/p/guzz/

"you can query databases through jsp taglib" - this is supposed to be a cool feature?? I wouln't even mention it to developers

Is it again a cool feature?!!

"you can query databases through jsp taglib" - this is supposed to be a cool feature?? I wouln't even mention it to developers

Yes, MVC is great! But sometimes, in some kind of applications, this would be very useful. Especially for some read-only features. We have used this "cool" feature in some projects, it behaved nice. This is not  just a imagination.

I know this is not just an imagination, I saw it in practice and that's why I'm scared

"you can query databases through jsp taglib" - this is supposed to be a cool feature?? I wouln't even mention it to developers

Yes, MVC is great! But sometimes, in some kind of applications, this would be very useful. Especially for some read-only features. We have used this "cool" feature in some projects, it behaved nice. This is not  just a imagination.

 

That would come under the heading of "anti pattern".

"you can query databases through jsp taglib" - this is supposed to be a cool feature?? I wouln't even mention it to developers

Yes, MVC is great! But sometimes, in some kind of applications, this would be very useful. Especially for some read-only features. We have used this "cool" feature in some projects, it behaved nice. This is not  just a imagination.

 

That would come under the heading of "anti pattern".

No, I would rather call it a new pattern. The "pattern" name is data declaration. 

We are familiar with Model + View, which separate all java logic to the model side to avoid a mess file. But you must admit that only the View pages and we the developers know exactly what the data they need.  The Model don't! The problem of MV is we have to tell the model what the view need, and the model must change when the view changed(new data is required).  But, why can't the View determine what she need herself, and tell a clever model to load it for her? 

This is a different approach. In fact, our taglib turns the MV architecture to MMV, the traditional Model + declared data model + view pages(template files), and the delcared data is declared in the view pages represents as jsp taglib or velocity directors. The declared data model is the clever model.

In our "pattern", global, user-inputed and context related data is passed to the View as usual, but "feature" related data which only the view page itself knows is declared and loaded as the declared data. The framework combines the two sources together, and walks down as the traditional MV does to generate the returned result.

So, taglib doesn't mess your java logic and view pages together, and break the popular M+V down. It just moves your "feature data java logic" written in every projects to a framework, and leaves you another easier way to just declare for it. It is a choice.

So, is it just an ugly combination of sub-features of Spring, Hibernate, iBatis SQL Maps, and Hibernate Shard?

Hibernate Shard does the horizontal splitting and secondary tables does the vertical splitting.

So why should someone ignore bullet proof solutions and use yours?

So, is it just an ugly combination of sub-features of Spring, Hibernate, iBatis SQL Maps, and Hibernate Shard?

Hibernate Shard does the horizontal splitting and secondary tables does the vertical splitting.

So why should someone ignore bullet proof solutions and use yours?

I have studied hibernate and ibatis before doing this job. They are cool and powerful, but not good enough for "large system".

Guzz's shard  is different from Hibernate's Shard. 

In guzz, you declare database groups, and choose which database group to persist your tables in a configuration file. It means you can distribute tables in several  databases when NEEDED,  without changing any code. Besides, guzz separates read/write for you, say, you can add slave machines when needed 

About splitting a big table. In guzz, you just have to implement a interface to do that. It is easy to control.

Guzz admits that an application is growing to a large one. So, the features in guzz are  designed to improve your performance with little efforts in the future. It is designed for both developing and maintaining. I don't think it is just a copy of other frameworks.

Combining features of all the frameworks available does not make a system usable ... It becomes clumsy ... all these claims will work only in demo ... in real life situations, these are unworkable, unmanageable  and un-programmable/un-configurable...

We have too many frameworks now... Each one claiming to be better than the other and at the same time 'allowing plug-ins' to all others ...

This can nevr be a general purpose framework.... may be suitable for some special needs ...

Guzz is located in China?

Redora, http://code.google.com/p/redora/ is developed in NanJing. Where exactly are you located?

Jaap