Plugins mean extensibility (Re: Kate part in KDevelop-2 working)

[Andrew Sutton]

On Friday 14 December 2001 08:03 pm, Joseph Wenninger wrote:
> The big problem with all those open source projects, even those which I 
> work at, ist that the developers only code for fun, but don't really think
> about usability, GUI design, ...... And this is the point, why will loose
> in the long run and Microsoft (or other companies, which behave like them)
> will win. Even though their products are instable and have many design
> mistakes, they concentrate on only one thing and at least try to write
> usefull integrated applications. Look at our open source products. Look at
> gideon, or other IDEs, office packages, .... . They are all patchwork. Even
> they have good underlying technologies and are mostly quite stable, they
> are only patchwork, because we try to be over extensable. Look at gideon,
> there is now real GUI concept in it, at least I, and some of my friends
> (not KDE developers) see it that way. Microsoft offers one editor, which is
> fully integrated in ther IDEs, and always supports all features and even
> though many people think working with those IDE's is a pain, they use it,
> look at Borland and others, they all loose against Microsoft.

actually, almost everything microsoft writes is as patchwork as you can get. 
take internet explorer. it isn't even a program. it's a bunch of jumbled 
together COM components that glue nicely together to provide a cohesive web 
browsing package. the program itself is just a shell. the same for office and 
visual studio, email, etc...

extensibility is a trend that isn't going to go away. and i don't think the 
extensibility here is really an overgeneralization issue. what has been done 
- as far as i've seen - has been to leave holes where other people can plug 
things in. like, for example, a PICO editor component.

these "holes" are the result of unresolvable personal preferences between a 
vast array of developers. if you don't like the way part of the IDE works, 
reimplement it with something you're familiar with. its called 
"replaceability".

> All in all,I look indifferently at these editor interface quarrels, they
> are just a good example of wide spread, instead of concentrated efforts,
> and that's why Open Source, despite all other opinions will fail in the
> long run. Be it office packages, editors, IDEs, ...... .

i don't think that anybody can really deny that these devisions hurt the 
overall development progress :) but i don't think it should be seen as a bad 
thing. why shouldn't people be inventive? why should one way win out in the 
end?

the crux of the problem isn't the derivation and deviation of software 
packages, it's the reintegration of those packages under one roof. i think 
gideon goes along way to show what can be done for this. personally, i think 
its great that a single dev environment would allow me to use kate, emacs, 
vi, or whatever. btw great work with that! its a good example of comprimise 
and integration :)

> I just want to say: Stop believing in over extensibility, which is not
> really needed, concentrate on the core. Gideon is a mess: many plugins, but
> not really a usefull common GUI.

i'd have to retort: never underestimate the need for extensibility. who knows 
what crazy (but spectacularly good) ideas are around the corner? why bother 
rewriting the entire core of a software package when new ideas can just be 
plugged into the existing core... take a stand! consider the refinement of 
the gui secondary to a good architecture ;)

and keep up the good work.

andrew sutton
ansutton at sep.com