gmane.comp.web.server.aida

Aida-Roles and Aida-Parties

Hi all,

I'm thinking about the usefulness of Aida-Roles and Aida-Parties...
I don't think those 2 packages should be included Aida. First, IMHO it
has nothing to do in a web framework, and it makes Aida bigger for
almost nothing. Maybe we should keep only the Person class, or move it
to Scribo?

Janko, if you use it for BiArt, maybe you could just move it from Aida
to BiArt?

What do others think?

Cheers!

Nico
--

-- 
Nicolas Petton
http://nico.bioskop.fr
            ___
          ooooooo
         OOOOOOOOO
        |Smalltalk|
         OOOOOOOOO
          ooooooo
           \   /
            [|]
--------------------------------
Ma clé PGP est disponible ici :
http://nico.bioskop.fr/pgp-key.html

(Continue reading)

headers

Janko Mivšek | 17 May 14:13

Re: Aida-Roles and Aida-Parties

Nicolas Petton wrote:

> I'm thinking about the usefulness of Aida-Roles and Aida-Parties...
> I don't think those 2 packages should be included Aida. First, IMHO it
> has nothing to do in a web framework, and it makes Aida bigger for
> almost nothing. Maybe we should keep only the Person class, or move it
> to Scribo?

Party Role framework can be very valuable when you build more complex 
security scenarios, for introducing role based access control for 
instance. So even that is currently not used, can be in near future.

Duty of any framework is to provide tools for you and not necessary for 
other parts of framework alone. And from my experience this tool is 
valuable, that's why I decided to move it directly to Aida.

Another reason is that because of lack of such framework we have now two 
objects for persons: WebUser and Person in Scribo. With party/role 
framework in Aida we can now migrate in only one: Person with a WebUser 
role!

Party role framework currently looks quite complicated but when a bit of 
description will (hopefully soon) arise you see that it is not so 
complicated as seems. But very powerful.

Janko

> 
> Janko, if you use it for BiArt, maybe you could just move it from Aida
> to BiArt?

(Continue reading)

headers

Nicolas Petton | 17 May 14:05

Re: Aida-Roles and Aida-Parties


Le samedi 17 mai 2008 à 14:13 +0200, Janko Mivšek a écrit :
> Nicolas Petton wrote:
> 
> > I'm thinking about the usefulness of Aida-Roles and Aida-Parties...
> > I don't think those 2 packages should be included Aida. First, IMHO it
> > has nothing to do in a web framework, and it makes Aida bigger for
> > almost nothing. Maybe we should keep only the Person class, or move it
> > to Scribo?
> 
> Party Role framework can be very valuable when you build more complex 
> security scenarios, for introducing role based access control for 
> instance. So even that is currently not used, can be in near future.
> 
> Duty of any framework is to provide tools for you and not necessary for 
> other parts of framework alone. And from my experience this tool is 
> valuable, that's why I decided to move it directly to Aida.
> 
> Another reason is that because of lack of such framework we have now two 
> objects for persons: WebUser and Person in Scribo. With party/role 
> framework in Aida we can now migrate in only one: Person with a WebUser 
> role!
> 
> Party role framework currently looks quite complicated but when a bit of 
> description will (hopefully soon) arise you see that it is not so 
> complicated as seems. But very powerful.

I didn't say it wasn't, but Aida is a web framework, not a Role/Party
framework. IMHO it's not its role to provide such party framework, and
it has nothing to do with web programming. Furthermore, I like to have

(Continue reading)

headers

Janko Mivšek | 17 May 14:43

Re: Aida-Roles and Aida-Parties

Nicolas Petton wrote:

>> Another reason is that because of lack of such framework we have now two 
>> objects for persons: WebUser and Person in Scribo. With party/role 
>> framework in Aida we can now migrate in only one: Person with a WebUser 
>> role!
>>
>> Party role framework currently looks quite complicated but when a bit of 
>> description will (hopefully soon) arise you see that it is not so 
>> complicated as seems. But very powerful.
> 
> I didn't say it wasn't, but Aida is a web framework, not a Role/Party
> framework. IMHO it's not its role to provide such party framework, and
> it has nothing to do with web programming. Furthermore, I like to have
> things clean and optimized, and instead of this role/party framework, I
> would rather work on a persistence framework, which is more needed in my
> opinion, and at least has something to do with a web framework.

User/Group/Security framework is also not necessary for a web framework 
but we all know how big advantage is that for Aida.

Party framework looks not so important in the first sight but when you 
start solving already mentioned WebUser/Person duality, you soon come to 
conclusion that you need to move part of party framework to Aida. But if 
you move part, why not more, at least as much as needed that party 
framework is complete. That was my line of thinking when decided to move 
it to Aida.

Also, don't underestimate all possible and not possible security needs 
by your customers. Role based security for instance is quite common

(Continue reading)