[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Devel] wls development and future



torsdag 31. juli 2003, 12:15, skrev Andreas Schuldei:

> What in webmin do we really use? any other modules? squid config?

At first glance it seems to be 4-5 services that could 
be administrated from a web user interface. It's 
probably twice as mutch. Here's the list of things that 
could be managed from a web user interface (as you 
see I dont mention web-min at all cause that's):  

* user-admin (create, retrive, update, delite, insert, find)
 w/printer quotas
 w/goups (as classes a 2000a, 2000b etc.)
 w/limiting som rights (as disconnect some applications 
  for pupils that has broken som Internet-roules at school).
* network adm (dhcpd.conf etc)
* thin client adm (lts.conf etc)
* printer adm (cups thing, quotas should be integrated with 
   system for user adm)
* services adm (turn of and on services on the schools network)
* surveillance (of some of the primary services, system status)
* disk management (lwm etc)
* backup-system

> Is that the [web-min] interface of choice? We use cups with its
> web-interface, dont we? So if we dont really rely on webmin, we
> might just ditch the beast and use the user-interface stand
> alone.

As you and others discovered the code in Web-min is monolitic in 
nature, and limits the requirement from Petter who wants 
leveled code that is plugable. Then we can integrate the 
¨bottom layer¨ with LDAP, /etc/passwd, FEIDE, or file-import/export. 

> And then, there is the remote chance that our initiative inspires
> the webmin upstream and he starts redoing his code. or that he
> accepts our module even if it does not fit in the monolitc pile
> of modules.

It seems that we have to choose between a well arranged code 
with a security framevork that people easily could mantain -- or 
more spagetti-code :-). The hope is that the Web-min people 
se that you have som nice practic, and could catch up, or we 
have to maintain this module ourselves. It took a couple of 
month, maybe half a year before the Debian developers looked 
at what "our" people did with the installation system. If we got
the patience, I'm optimistic about the fact that Andreas could 
make a better framwork for web-min, and sombody should take 
the risk by leading the way. 

> yes, i think the code is much more reusable this way, so
> building in parallel or on top of this should be rather painless.

Isn't it posible to link in your work in web-min without using their 
messy framwork (if it is any)?

mvh Knut