Mailing-List: contact dev-help@directory.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Apache Directory Developers List" <dev@directory.apache.org>
Received-SPF: pass (herse.apache.org: local policy)
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
  s=s1024; d=yahoo.com;
  h=Received:X-YMail-OSG:Mime-Version:In-Reply-To:References:Content-Type:Message-Id:Content-Transfer-Encoding:From:Subject:Date:To:X-Mailer;
  b=nnj2ZckWGfK/9Q0lXDy3ZJP84I1RJEAy3aEwZEjOb5YQpgX/2Lu0cWOFSXdW8624Ij8xw3/SgX+IMwxGJgIWbWGQWWoenGZEb47aLvTsnkpkXEpoHKcJR5uTBYZkhUGBt6685neifDU9K3N1gRuL4RUVB2KcVoFw/bQBJiYJCfY=
  ;
Mime-Version: 1.0 (Apple Message framework v752.3)
In-Reply-To: <45A4175B.4020307@gmail.com>
References: <45A3E975.5000007@labeo.de>
 <3E1E6C3C-DBD5-4059-BDB9-D6F766096EC2@yahoo.com> <45A4175B.4020307@gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; delsp=yes; format=flowed
Message-Id: <8F9B7F8E-23A4-43B7-86F6-D2AE46BB183F@yahoo.com>
Content-Transfer-Encoding: quoted-printable
From: David Jencks <david_jencks@yahoo.com>
Subject: Re: Reducing the complexity/verbosity of default server.xml (and
 custom configuration files)
Date: Tue, 9 Jan 2007 18:02:42 -0500
To: "Apache Directory Developers List" <dev@directory.apache.org>


On Jan 9, 2007, at 5:29 PM, Emmanuel Lecharny wrote:

> Well, my position is a little bit different. Sure, too much XML suX=20
> (ml), but the problem is not that the configuration is huge, it's =20
> that we currently don't have tools to manage that (tools =3D GUI).

I think relying on tools to deal with unnecessary complexity is a =20
pretty bad idea.  Look at jax-rpc for j2ee :-) -- it's IMO totally =20
unusable and incomprehensible because the xml is so awful.
>
> But with LdapStudio, I think we might have an acceptable solution :
> - users who like to click will use it

having tools is definitely good
> - users who prefer vi and long text file will favor xml files :)

bad xml will drive everyone away, as-simple-as-possible xml will let =20
everyone understand what is going on.
>
> However, there are so much things in the server.xml file which =20
> could be defaulted, that it might be a good idea to spend some time =20=

> on this subject. I don't really see the point of having all the =20
> intercptors exhibited here. Maybe we can default those guys...

defaults are good....

We need some kind of xml, so I think devoting a reasonable effort to =20
making it really clear, simple, and expressive is a good idea.  I =20
haven't used it myself but the xbean stuff sounds really good from =20
what I've heard.

thanks
david jencks

>
> <snip/>
>
> David Jencks a =E9crit :
>
>> I think James Strachan and Dain Sundstrom have cooked up some =20
>> stuff  that automatically generates schemas for each bean and then =20=

>> lets you  use that customized xml for your spring configuration.  =20
>> It's  somewhere in the xbean project inside geronimo.  I'll try to =20=

>> find out  more if there's interest.  I'm pretty sure  activeMQ is =20
>> using this  for its configuration.  =46rom what I've heard its a big =20=

>> improvement  over normal spring configuration.
>>
>> thanks
>> david jencks
>>
>> On Jan 9, 2007, at 2:13 PM, Stefan Zoerner wrote:
>>
>>> Hi all!
>>>
>>> There is some criticism about configuration complexity of =20
>>> ApacheDS,  especially from people who argue after a first glance =20
>>> at it. The  verbosity is mostly driven by the Spring Framework.
>>>
>>> 2) Reduce length of partition configuration
>>>
>>> Currently, partition configuration are very long, mostly because =20
>>> of  the many many indices for internal attributes:
>>> ...
>>>         <bean  =20
>>> class=3D"org.apache.directory.server.core.partition.impl.btree.Mutable=
=20
>>> In dexConfiguration">
>>>           <property  =20
>>> name=3D"attributeId"><value>1.3.6.1.4.1.18060.0.4.1.2.1</value></ =20=

>>> property>
>>>           <property name=3D"cacheSize"><value>100</value></property>
>>>         </bean>
>>> (many others follow, in each partition the same
>>> ...
>>>
>>> If it is highly recommend to use these indices, how about having =20
>>> a  special IndexConfiguration as a default, which only has to be  =20=

>>> extended in order to adjust it to your needs?
>>
> yeah, sure, good point.
>
> Stefan, can you do a list of all the elements that could be =20
> defaulted ?
>
> Emmanuel