[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-core-devel
Subject:    Re: KOffice development -- Standardization concerns
From:       Lotzi Boloni <boloni () cs ! purdue ! edu>
Date:       1999-08-02 19:08:07
[Download RAW message or body]


  Reggie

  I will try to look into this a bit after I go back to my quiet lifestyle
at the university. 

  Still, I think that one of the problems why everybody is a bit afraid
about going into that is the sometimes wild turns the development is
taking. 

  I am afraid that I will start a flamewar here but there are some
decisions which might have their technical reasons, but may be bad for the
system in the long term:

 -going miniSTL instead of the standard one
 -going KScript instead of anything else (Python or whatever)
 -making kded instead of contributing a standard CORBA trader to mico
 -even thinking about making a KDE specific CORBA mapping 

  I know that there are valid technical reasons behind all these, but this
means that we are restricting the number of people working on these
projects! A lot of people wants to learn Python, a lot of people want to
learn standard CORBA (it gives you money to live while you are working on
KDE) and so on. 

  Also, whenever we are making the choice to implement our own "better"
implementation, we are making the horrible commitment that we are going to
support it till the end of times! 
  There are a lot of people outside taking care of CORBA, taking care of
STL, and even if they are slow, bloated etc, there are going to get
better. But we don't have that many KDE developers that we can commit to
all these projects. 

   I would rather have a looonnnggg list of external standards and
languages needed to run KDE (provided that we are carefull to choose ones
which are indefinitely supported in the future like CORBA, LDAP, ODBC)
than making our own implementation - slightly better suited, but having
only one or two developers really knowing what's going on and without
strong incentives for others to learn it.  I would really like KDE relying
on CORBA naming, trading and events. There are a lot of people working
on/with these things. (I am paid by AT&T to do it.) 

> I really think that KOffice can be THE Killerapplication in the free software
> development if we concentrate on application development and stabilize the
> applications and do not change the framework all the time. If we don't get on
> the correct path soon, we will never get a stable and useable KOffice and one
> day the project might fail. So it's really important that we get more people on
> board which are interested in application development!

   Agreed. This is why I am arguing for the standards to be not the
internal KDE standards but the industry standards. 

   Please flame me only moderately. :-)
      
           Lotzi

[prev in list] [next in list] [prev in thread] [next in thread]