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

List:       kde-devel
Subject:    Re: Developing applications for KDE4
From:       Ian Wadham <ianw2 () optusnet ! com ! au>
Date:       2007-01-30 12:57:13
Message-ID: 200701302357.14350.ianw2 () optusnet ! com ! au
[Download RAW message or body]

On Tue, 30 Jan 2007 05:51 pm, Aaron J. Seigo wrote:
> On Monday 29 January 2007 23:23, Ian Wadham wrote:

> > of what appears to be "mushroom management" in this project.
> sorry, i'm not familiar with this term. can you explain further?
>
By analogy with growing mushrooms ... keep the troops in the
dark, tip a bag of manure over them occasionally and if one of
them pops up his head, cut it off.  Any semblance of this in
KDE4 is, I hope, coincidental and accidental.

> you know, i've been telling people until quite recently (e.g. the last
> month or two) for quite some time that if you are working on applications
> to continue doing so with kde3. kde4 is only now getting to a point where
> periphery application development makes sense IMHO.
>
Silly me ... I should never have believed what they were saying in the
KDE Games list 6 months ago about Qt4, QGraphicsView and KDE4
and about only doing bug fixes in KDE3.  KDE Games started porting
to KDE4 more than 6 months ago and we are now improving existing
games and welcoming new games.  I can claim no credit for this.

Mauricio Piacentini of Sao Paulo converted KGoldrunner to native
KDE4/Qt4 and it was first cab off the rank.  He and I then spent 3 months
or so battling over a performance issue in QGraphicsView (not fast
enough for 20 frames/sec animation), which was a large part of the
7 months I am upset about.  Some of the rest was learning curve ...
SVN, CMake, Qt4 libs (especially QGraphicsView) and KDE4 libs ...
and, of course, I do not spend *all* my time working on KDE.

> in each case you're missing two points, and in some cases three:
>  - we don't direct centrally who can work on what. welcome to open source.
>
Even if you did, it would still be like trying to herd cats, in my experience
as a software developer.  There is such a thing as leadership, however,
and if you are a respected leader (as opposed to a director), you can
inspire, influence and guide people's choices.  Even if you have no
charisma at all, you can post a "wanted" list somewhere prominent.

> adobe acrobat: slow, bloated, not embeddable, only does pdf's, enforces
> drm, people have an interest on working on kpdf.
>
Know what you mean, but it works and the reader is (cost) free.  Recently
was faced with downloading 20 Mbytes on dialup for my son's new Dell
WinXP laptop.  After calculating how long that would take, I tried copying
an old 5 Mbyte version from his old Win98 PC and it worked!  Now there's
something you couldn't get in a KDE app.

> OOo: slow (koffice has blemishes on this account too), bloated, not
> embeddable, limited in scope compared to koffice. happened long after
> koffice, which people have an interest in working on.
>
Four years ago I composed 120 pages of notes for a science course I was
presenting, using OpenOffice, including diagrams, etc.  It was a delight to
use and entirely satisfactory as to speed, usability, etc.

> GNU Cash: i suppose you mean kmymoney. you do realize that that is a "3rd
> party app", right? i hope you understand we can't go around colonizing
> every project that uses kde libs and commanding them to do something else.
>
No, sorry, how would I know?  It's in the same menu area as KOffice apps
and starts with a K.

> Xephem: don't know enough about it to comment.
>
XEphem is an X-Windows astronomy application or "desktop planetarium",
also available in MS Windows I believe.  I used it in the same science
course, to great acclaim, despite its clunky UI.  It had the underlying 
functionality and UI configurability I needed, whereas KStars did not,
though it has caught up a lot in the last three years.

> > Qt has list-processing classes ... but wait! ... Qt has re-defined those
> > classes several times in the last few years, necessitating repeated
> > changes in every application that uses them, including mine.  It's
> erm. they changed with Qt 4.0. so i count that as "once". how do you
> count "several times"?
>
KGoldrunner has used QList, but a different QList, in Qt1, then it had to
use QPtrList and now it's QList again.  That's not "several", I admit, but
it's two changes too many for such a fundamental programming tool,
the idea of which has been around for at least thirty or forty years now.

> yes, so all those hours i put in porting literally millions of lines of
> code for every change i made in kdelibs was what? and my efforts pale in
> comparison to those of people like montel, faure and others. we've done as
> best we can to do the work for you; deprecated calls still exist as do
> usages of kde3support ... which is yet more work we've taken upon ourselves
> to ease transition.
>
So why isn't that a strong incentive for you to stabilise the libraries, or
at least change the implementations without changing the interfaces, other
than by adding new functionality?

> as for the "punishable by death" statement, what is "punishable by death"
> is an app that isn't maintained for which there are quality alternatives.
>
Well, in KDE Games, those that do not find a maintainer and get ported
to KDE4 are threatened with extinction.  Similarly if they do not upgrade
to scalable graphics with more modern artwork.  I'd hate to see a really
good and enjoyable game go under for technical reasons, but so far
it has not happened (touch wood).

> > the point, I do not feel that I will ever progress further and offer all
> > that I believe I can offer to the KDE community, such as working on
> > more serious applications.
> is this dramatism or do you really feel this way? (serious question)
>
No.  I really feel this way.  My alternatives are to develop in some
more stable library environment, if such a thing exists, or develop
in Qt only (which is unstable enough) or give up and go back to
woodwork or gardening.

> > On the other, if all the talent in KDE and Trolltech could put its heads
> > together to come up with some *stable* core libraries, I am sure there
> > would be a quantum leap in Open Source application development,
> > to say nothing of greater acceptance for Linux and Open Source in
> > the marketplace.  We might even have apps that could run on several
> > successive versions of the KDE Desktop.
>
> the "if only you would then ..."
>
> s,you,we,g
>
> > Is that too much to ask?
>
> what do you think?
>
Eh!  Your turn to explain ... :-)

> i'm all ears for rational and constructive criticism, but the current fad
> of negativity is more than i can bear, and i suspect i'm not the only one.
> it's dragging me down.
>
I thought I was being constructive in my last point.  I have a few other
constructive suggestions, but I do not think you are ready to receive
them yet.

> so .. frustrated? great! express that in terms of solutions.
> feeling like hi-jacking a thread on kde-devel to do little else than poison
> the spirits of others who actually happen to be making nice progress? take
> a walk in the fresh air and come back when the feeling has passed.
>
Please calm down.  I am not hi-jacking anything.  I have a right to be
here and I am not trying to poison anybody's spirits.  My aim is to lift
applications writers' thoughts above endless "maintenance" work and
to make library writers more aware of the effect of changes they make.
Have you heard of "change management" or "impact analysis"?

Peace, Ian W.
 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<

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