[PLUG] Re: linux IDEs

Eric Wilhelm scratchcomputing at gmail.com
Thu Aug 2 06:14:37 UTC 2007


# from Quentin Hartman
# on Wednesday 01 August 2007 05:36 pm:

>How do we get there from here?
>
>- De-balkanize Linux by reducing the number of distros?

won't happen

>- Reducing the number of Desktop Environments? 

nope

>- Adhering more closely to LSB?

yep

>- Create a good GUI IDE that has the strengths of VS by nicely tying
> all the Linux pieces together?

Freedesktop has done some good work in being sort-of the LSB for 
desktops.  I think it is finally gaining some foothold.

Now, will it go deeper into inter-app communication?

I would love to see something *like* dcop for all applications, but 
simply getting some deeper hooks into kwin/kicker would be nice for 
e.g. being able to programmatically raise/lower a (non-kde) window.

For example, this program would probably be a lot less ugly if gvim used 
dcop.  But then, it wouldn't exist if gvim didn't have client/server 
support.

  http://scratchcomputing.com/svn/misc/trunk/code/perl/bin/eedit

Admittedly, emacs does the IPC thing a bit more elegantly.  Now, if only 
it had an editor...

Of course, vim tends to be a bit cruftier in the windowing department 
(But, its programming language has a syntax (though really, it's the 
modality that I want (and viper just doesn't work for me (I never did 
take to lisp (despite the outward appearance of this sentence.)))))

>  - Improve Glade/GTKbuilder? What about KDE?
>  - Improve Kdevelop? What about Gnome?

See smalltalk.  See smalltalk run.  Run smalltalk, run!

The trick is simply to get it out of being stuck in that stupid box.  
Eclipse seems to have a similar problem from what I've seen.  Sockets 
and daemons, anyone?

>It's a hard problem, and the only "real" solutions are sure to make
> someone cry.

It is a hard problem, but I think the real solution is to make money.

>What about web applications? If someone, like Ken, is looking to build
>web apps, what sort of "studio" do we have on offer for them? Arguably
>this is the battleground that the current generation of developers will
>be won or lost.

Maybe.  Catalyst, Jifty, and Rails still tend to stick to the console 
and assume some knowledge of SQL, sysadmin, etc.  I'm not sure that 
involving SQL is a good thing, but I'm guessing SQLite would have 
something to do with the out-of-the-box environment.

Likely, by the time anything appears, it will be totally outmoded as we 
move on to web 3.0.

Meanwhile, I just want better/cleaner glue and more access to the 
innards of various systems from userland.  If there is a "Linux way" to 
build an integrated system, I think that's probably it.

FAM+dcop = refresh konqueror when you save a file

  http://scratchcomputing.com/svn/misc/trunk/code/perl/bin/reloader2

FAM+rsync = transfer to another server when you save a file

  http://scratchcomputing.com/svn/fabsync/trunk/bin/fabsync

FAM+git = ???

And I'm still pondering what can be done with foot-pedals.

We might have touched on it a little in this thread, but aptitude and 
CPAN are really great reasons to use open source.  The ability to 
automatically resolve and satisfy complex chains of dependencies with 
zero procurement hassle means *writing less code*.  Better 
coordination, search, QA, automation, and other improvements will 
continue to add usefulness and power.

--Eric
-- 
"It works better if you plug it in!"
--Sattinger's Law
---------------------------------------------------
    http://scratchcomputing.com
---------------------------------------------------



More information about the PLUG mailing list