Skip to main content

Why don't JDNC work closely with Netbeans Platform development team.

13 replies [Last post]
anilpatel
Offline
Joined: 2003-06-30

Hi,
I see that Netbeans platform provides lot of things that JDNC swingx developers can use (murge with) for providing richer toolkit.

Like,
Netbeans form editor generates xml file with .form ext. This file is used for generating java code.
similary jdnc uses xml file for generating User interface.

Other tools like tree table and lot more are simply redundent between these two projects.

I think JDNC should use the existing resources of Netbeans Platform and present it as Enterprise application framework.

I am just trying to achieve state when we can write more application that Comman man can use, Instead of coming up with new frameworks everyday.

In .NET world, One framework is used by millions of developers. In Java world, Developers have to first strugle thru the process of evaluating Millions of frameworks, and finally end of writing their own, By the time developer is done writing a framework, Project runs out of money and time.

Please see if JDNC team can review the Netbeans Platform (again if they have already done it) and have good reasons for doing samething again.

Please hear voice of "Not so Smart" developer.

Anil

Reply viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
Marek Mosiewicz

Sure, Java is much more richer, but not in case of database gui.
I hope JDNC will change this.

I hope that You can agree that flexibility is good thing. So good framework should be flexible.
Hibernate is great example of good framework becouse they collect user experience.
They care of solutions for developers problem so they put more features (make it more flexible) all the time.

Having two technologies not good thing You will sometimes find into stupid problems. For example Your customer want report from data which he has in searched in edit grid (resonable requirement). So now you have cross point.
There is usually also some duplicated code to deal.
And in fact hated .Net DataSet shows its flexibility, becouse it can be exposed as XML.
So there You can use both DataSet->grid or DataSet->XML->xsl:fo = (pdf, rtf, awt, ps) without problem.
It makes our Matrix a little better place.
No technology is perfect and in real world You will have probably to struggle a lot of different problems.
But better frameworks makes life easier.
For me Delphi as database gui (or .Net DataSets/Forms which base on Delphi) is something really nice there.

Marek Mosiewicz
http://www.jotel.com.pl

-----Original Message-----
From: jdnc-interest@javadesktop.org
[mailto:jdnc-interest@javadesktop.org]
Sent: Monday, December 06, 2004 10:29 PM
To: jdnc@jdnc.dev.java.net
Subject: Re: RE: RE: Why don't JDNC work closely with Netbeans Platform
development team

I am not agree with you completely.
In most cases Java is much richer that Delphi (i am not talking about GUI).
Yes you have to make a choice and after you (or your application) have to live with it.
"Welcome to the real world" (c) Matrix

I am agree with you that it is very hard to make such a choice, just a few guidelines available or not at all. most of the stuff is in "do it yourself" stage.

Yes, if you will use Hibernate for reporting - you will struggle, try Hibernate in Action book.
But if you go this way (JDBC -> xml) + xsl:fo = (pdf, rtf, awt, ps) it will be much better and less pain :)
---
[Message sent by forum member 'Ivan' (Ivan Latysh)]

http://www.javadesktop.org/forums/thread.jspa?messageID=42303&#42303

---------------------------------------------------------------------
To unsubscribe, e-mail: jdnc-unsubscribe@jdnc.dev.java.net
For additional commands, e-mail: jdnc-help@jdnc.dev.java.net

---------------------------------------------------------------------
To unsubscribe, e-mail: jdnc-unsubscribe@jdnc.dev.java.net
For additional commands, e-mail: jdnc-help@jdnc.dev.java.net

anilpatel
Offline
Joined: 2003-06-30

The problem is, It takes lot of time and money in Analysis and decition of Which framework to pick?

Most of the frameworks have more developers then users. After more then 10 years we are still developing frameworks to meet basic requiremens.

I think you need a strong user community driving the development instead of developers controlling the future of open source frameworks.

If a company (Sun in this case) will have projects that can be murged, How are we ever going to see bright future....

I am developer, Project Manager and also provide funding for development of Product. I know the Pain. Its not about doing cool thing, Its about getting Job done.

Please do not waste resources. Else we will never be able to Provide a good Product that can compete in Market.

Regards

Anil Patel

rameshgupta
Offline
Joined: 2004-06-04

> Hi,
> I see that Netbeans platform provides lot of things
> that JDNC swingx developers can use (murge with) for
> providing richer toolkit.
>
> Like,
> Netbeans form editor generates xml file with .form
> ext. This file is used for generating java code.
> similary jdnc uses xml file for generating User
> interface.
>
> Other tools like tree table and lot more are simply
> redundent between these two projects.
>
> I think JDNC should use the existing resources of
> Netbeans Platform and present it as Enterprise
> application framework.
>
> I am just trying to achieve state when we can write
> more application that Comman man can use, Instead of
> coming up with new frameworks everyday.
>
> In .NET world, One framework is used by millions of
> developers. In Java world, Developers have to first
> strugle thru the process of evaluating Millions of
> frameworks, and finally end of writing their own, By
> the time developer is done writing a framework,
> Project runs out of money and time.
>
> Please see if JDNC team can review the Netbeans
> Platform (again if they have already done it) and
> have good reasons for doing samething again.
>
> Please hear voice of "Not so Smart" developer.
>
> Anil

Anil,

The JDNC team does work closely with the NetBeans team, but you have to understand that their charter is different from the charter of JDNC. While JDNC focuses on the API layer, NetBeans and Java Studio Creator (formerly Rave) focus on the tools layer.

To ensure that the JDNC APIs are easily toolable, and for NetBeans to be able to produce tools that make using JDNC easier, the two teams do work closely together. However, this does not mean that the code produced by the two teams are (or have to be) completely miscible.

For one, NetBeans lives at a higher level than JDNC. While it is expected that NetBeans should be able to call into the JDNC toolkit, the reverse is not true.

More importantly, JDNC components have to be completely reusable -- within NetBeans, within Studio, within Borland's tools, within my application -- everywhere. That cannot be said of NetBeans components. I once discussed extracting NetBeans' XML Editor with Steve Wilson, and found that it was non-trivial to lift the component cleanly and transplant it within JDNC. While it is true that NetBeans has some components similar to what the JDNC team is also trying to deliver, the amount of time and effort that would be necessary to take that code and recast it so that it meets the tests that libraries such as Swing and JDNC must meet would be prohibitive (IMO).

Ramesh

anilpatel
Offline
Joined: 2003-06-30

Ramesh,
Thanks for the response. I am agree with you and withdraw my concerns.

I hope more people will join this community as users and add value to it.

Regards
Anil Patel

charles.armstrong
Offline
Joined: 2006-02-17

> In .NET world, One framework is used by millions of
> developers. In Java world, Developers have to first
> strugle thru the process of evaluating Millions of
> frameworks

Different frameworks have different strengths and weaknesses, and are suited to different applications.

The solution to the problem you raise is for every framework to be clear about what it is useful for. The problem arises when every framework claims to be general purpose.

A diversity of Frameworks with a clear purpose is a positive - if your problem is well defined then picking
the right one shouldn't be difficult.

Charles.

Marek Mosiewicz

I like Java, so when I hear something like this I have to disagree.
That is simply not true. Such mass of framework is disaster for Java.
It increases development costs increadibly. It prevents from
creating good tools (every framework has Ecplipse plugin, but lot of them is crap)
You can say that each framework has good sides and bad sides. So choose good for You.
But usually requirements CHANGE when project starts and what You thougth will be good
can be no longer.
Framework should be simply good, no good in some areas.
It was clearly visible with CMP EJB. It was ridiculus that people(me too) even tried to use it
Such obvious limitiations like no dynamic queries. It simply had to fail. But it was framework just
good for some purposes :).
Now we have another Java data myth - POJO. Usually it will be good for You. When You do not want to have
mutable schema (that mean schema will not change without modification/recompliation of code).
If You do not need mutable schema use POJO if You need it use RowSet API - that what You will hear.
But it is so stupid. Becouse I do not like to make such decision in advance.
You can find a lot of places where You would like to use mutable schema, but choosed POJO earlier.
But there is no combined solution there like Typed DataSets in .Net. They have both and do not have
to make such decicions. They can use pure data set when binding to grid, and use POJO like solution
when write business logic (but do not have to).
Sure there is always solution market. For sure there will be new solutions in future and .Net platform will be
obsolate (like all java frameworks too). Someone will find solution to make things much more easier.
But what they did for now, they gathered all good solutions together (Java for language, Delphi for database solutions etc)
and made quite good framework.
In Java I have no such comfort :( (fortunetly have platfrom independance, free tools and some more)

Marek Mosiewicz
http://www.jotel.com.pl

-----Original Message-----
From: jdnc-interest@javadesktop.org
[mailto:jdnc-interest@javadesktop.org]
Sent: Saturday, December 04, 2004 9:03 PM
To: jdnc@jdnc.dev.java.net
Subject: Re: Why don't JDNC work closely with Netbeans Platform
development team.

> In .NET world, One framework is used by millions of
> developers. In Java world, Developers have to first
> strugle thru the process of evaluating Millions of
> frameworks

Different frameworks have different strengths and weaknesses, and are suited to different applications.

The solution to the problem you raise is for every framework to be clear about what it is useful for. The problem arises when every framework claims to be general purpose.

A diversity of Frameworks with a clear purpose is a positive - if your problem is well defined then picking
the right one shouldn't be difficult.

Charles.
---
[Message sent by forum member 'charles.armstrong' (Charles Armstrong)]

http://www.javadesktop.org/forums/thread.jspa?messageID=42051&#42051

---------------------------------------------------------------------
To unsubscribe, e-mail: jdnc-unsubscribe@jdnc.dev.java.net
For additional commands, e-mail: jdnc-help@jdnc.dev.java.net

---------------------------------------------------------------------
To unsubscribe, e-mail: jdnc-unsubscribe@jdnc.dev.java.net
For additional commands, e-mail: jdnc-help@jdnc.dev.java.net

charles.armstrong
Offline
Joined: 2006-02-17

Marek,

Deciding how you are going to model your domain objects e.g. whether or not as POJOs **MUST ALWAYS ALWAYS ALWAYS ALWAYS** be the first design decision you make.

Every other decision flows from that.

In my view not doing this is responsible for more project failures than anything else.

As a minor point POJO + Hibernate (or another O-R tool) allows as much or more schema flexbility than Rowsets.

The book 'Applying UML and Patterns' by Craig Larman, takes you through all these issues, starting with your 'Conceptual Model'. It is the textbook for the IBM Object Orientated and design exam.

rgds,
Charles.

Marek Mosiewicz

That is true I HAVE TO make that decision, becouse I have no choice.
But when I create any application where user for example need some reports I'm in trouble :)
Sure I can use POJO for object domains and rowset for reporting, but usually I will end up
with doubling some logic.

But design I would like to have is dual model:
1.It is good to have not type safe objects for data bindings, so You
can add column do database and show it in form without recompilation, but simple CONFIGURATION
(I heard that 90% of systems do not need this, but obviuosly I have bad luck becouse I found
that in 70% of my projects it would be good thing :) )
2.And You can use type safe view of object in business logic (when You want to use new column in business logic
You have to recompile code OR use some scripting which uses not type safe view of object)
With this You can write really advanced things where administrator can change what You allow him to change)
Such soluton was available in Delphi 10 years ago and I thing it is good (do You agree ?)
AS i mentioned erlier flexible data objects is also MUST HAVE feature in any reporting.

I have not said that Hibernate (which I really like) do not allow schema flexibility but said that it not allow mutable schema - it means it requires recompilation after database schema change. From Hibernate 3.0 You can use Map as
Your data object (which is like RowSet not type safe) or can write own EntityPersister (entity persisters were available in
earlier versions of Hibernate but each table had to have own class which nust not be proxy class)
With Hibernate 3.0 you can also try solution which will create proxy class with properties for all fields in database on fly and it should work. Anyway imagine our reporting wher You want to display result for some user defined query.
It is little strange form my to play with cglib to create proxy on fly to hold data from query. Then use reflection to
display this data. But this is for me little strange solution.

So back to framework choice. I do not like to have lot of frameworks. I want flexible framework.

Marek Mosiewicz
http://www.jotel.com.pl

-----Original Message-----
From: jdnc-interest@javadesktop.org
[mailto:jdnc-interest@javadesktop.org]
Sent: Monday, December 06, 2004 4:44 PM
To: jdnc@jdnc.dev.java.net
Subject: Re: RE: Why don't JDNC work closely with Netbeans Platform
development team

Marek,

Deciding how you are going to model your domain objects e.g. whether or not as POJOs **MUST ALWAYS ALWAYS ALWAYS ALWAYS** be the first design decision you make.

Every other decision flows from that.

In my view not doing this is responsible for more project failures than anything else.

As a minor point POJO + Hibernate (or another O-R tool) allows as much or more schema flexbility than Rowsets.

The book 'Applying UML and Patterns' by Craig Larman, takes you through all these issues, starting with your 'Conceptual Model'. It is the textbook for the IBM Object Orientated and design exam.

rgds,
Charles.
---
[Message sent by forum member 'charles.armstrong' (Charles Armstrong)]

http://www.javadesktop.org/forums/thread.jspa?messageID=42248&#42248

---------------------------------------------------------------------
To unsubscribe, e-mail: jdnc-unsubscribe@jdnc.dev.java.net
For additional commands, e-mail: jdnc-help@jdnc.dev.java.net

---------------------------------------------------------------------
To unsubscribe, e-mail: jdnc-unsubscribe@jdnc.dev.java.net
For additional commands, e-mail: jdnc-help@jdnc.dev.java.net

Anonymous

I am not agree with you completely.
In most cases Java is much richer that Delphi (i am not talking about GUI).
Yes you have to make a choice and after you (or your application) have to live with it.
"Welcome to the real world" (c) Matrix

I am agree with you that it is very hard to make such a choice, just a few guidelines available or not at all. most of the stuff is in "do it yourself" stage.

Yes, if you will use Hibernate for reporting - you will struggle, try Hibernate in Action book.
But if you go this way (JDBC -> xml) + xsl:fo = (pdf, rtf, awt, ps) it will be much better and less pain :)

charles.armstrong
Offline
Joined: 2006-02-17

Marek,

> So back to framework choice. I do not like to have lot of
> frameworks. I want flexible framework.

I believe this is impossible. Some design designs have right and wrong answers. However most design decisions in a Framework are tradeoffs that benefit some applications and inhibit others.

I agree with you, some Frameworks are terrible.

However I believe that in the long term natural selection will operate and Java will end up with a few general purpose Frameworks, and 10-20 specialized ones that enable specific categories of applications.

.net will end up with a whole suite of me-too applications , and its the applications enabled by the specialist, imaginative, innovative Frameworks which will win the day.

Time will tell....

Charles.

rbair
Offline
Joined: 2003-07-08

Hey Anil,

We certainly don't want to be reinventing any wheels :). We do keep in contact with the Netbeans team, and are trying to make sure we come up with optimal solutions.

Thanks!
Richard

anilpatel
Offline
Joined: 2003-06-30

May be we should evaluate "what is No so good about"

1) org.openide.nodes.Node That we need org.jdesktop.swing.data.DefaultDataModel

2) org.openide.nodes.Node.Property has almost everything that can be done with org.jdesktop.swing.data.MetaData

3) Netbeans DialogDescriptor and DialogDisplayer are rich components for Dialog delivery.

4) Form module in netbeans rich component to describ a rich user interface.

5) Netbeans PropertyEditor provides rich framework for editing a Database Entity.

If people on this forum think we can start saperate thread for each "seems to be comman compoent/feature" and discuss what to do with it.

Dear "Core developers and Project Lead on this Project"

Please do not just ignore this thread. Let us review what we are doing.

Regards
Anil

netsql
Offline
Joined: 2004-03-07

The reson I do not use the more popular SWT in favor of JDNC is because it's locked into Eclipse IDE.

I think a lot of people feel that JDNC is indepdendent, I hope it stays as such.

If netbeans grants IP rights, then some cointer may chose to repackage so Swing extensions are best they can be.
(Especialy ekit, etc.)

.V