-32

There are two kinds of plasmoids: the typical desklet-things like a analog-clock, CPU-usage display, the taskbar etc. basically system-tools that do not interact with the user a lot
AND
more application like programms like: a stock quotes applet, a mini-browser, a twitter client,... all these could easily as well be programs and would suffer no disadvantages. I would argue they would better be programs and not plasmoids as they have a lot of disadvantages being plasmoids (see below)! What's more: if they were programs, they could be used in Gnome too, plasmoids do not work on Gnome as far as I know.

Think about it: some plasmoids are rather small programms, applications, but on a low organisational level. So plasmoids are basically expected to do the same essential things as normal applications:

1. have them continuously running
2. start and stop and restart them, if you like easily by some kind of programm-starter
3. minimize them without removing them completely from the desk and without stopping them
4. keep data once entered or be able to save the data
5. save their config for later restart from where you stopped
and other things that are possible
6. start and stop them depending on a time/date schedule
7.

Today, let's say my stock quotes plasmoid

1. is continuously running (although I do not want it during closed hours)
2. not possible to start/stop it, because I fear to loose the stocks I entered
3. minimize to somewhere in the background is not possible
4. I do not know if data is kept if I stop the plasmoid
5. same with the config (e.g. the form of the plasmoid on the desk)

The general question that arises is: does it makes sense i.e. does it bring any advantage to re-invent the application-program-principle a second time and call it 'plasmoids', or should any application-like plasmoid not better be a classic application that you can start stop and do all the things mentioned above (and more).

The 'Pastbin' plasmoid again is such a thing. Very practical, but why can it not be a classic little program? Where's the advantage of a plasmoid? Or is it just, you do it as a plasmoid because you can do it? Does that make sense?

Do not misunderstand me: I like plasmoids - in the right situation and place. But I do not want to see programming skills used to reinvent what is already there. We do not need a filemanager plasmoid (we have Dolphin and even Conqueror) or a browser plasmoid (we have, well, you know them) or a openoffice writer plasmoid or or or. These programs already exist and work fine.

So: We must distinguish what should be a plasmoid and why and what should be not. The existence of two parallel application 'worlds' within KDE4, normal applications and plasmoids is not a good thing at all as it limits the range a programming effort can be used by the users. Not to say that there are more parallel worlds besides QT4 like: KDE3, GTK, Java, .NET, which are all more or less compatible to each other but not perfectly.

Strategically thinking you must admit: this multiple split of the (programming-)worlds is one of the biggtest problems of the linux-world, as people not only have the choice but the burden to choose. Do not make this problem even bigger by developing apps that only run on the plasma engine and nowhere else!

The point of Plasmoids isn't to replace full application functionality. It's to provide limited functionality quickly and conveniently. I use Choqok quite a bit, but I still like the microblogging plasmoid in my panel.

Similarly, the Now Playing plasmoid isn't to completely replace Amarok's window - it's just quicker/more convenient to click the pause/stop button when it's in your panel or on the desktop, instead of having a full window open (why do you need the playlist and wikipedia information when all you're trying to do is press Stop?).

totally agreed.

However, there are plasmoids or efforts to program plasmoids that get dangerously close to real applications. In another thread here in Brainstorm, a guy proposes save/load/config functionality and I just want to point out where that leads. When plasmoids reach the state of functionality of complete applications this will be a problem. And the direction is clearly leading that way.

So my point is: we do not need another application plattform. Plasmoids should be what they are, and not develop into something like another applciation-API. We already have enough of them, that development effort is dangerously split up and not focused. I think already the split between GTK and Qt is a real drawback. We should not add others.

incredion wrote:
However, there are plasmoids or efforts to program plasmoids that get dangerously close to real applications.

Actually, most if not all plasmoids I've seen that could fall into that category are not shipped by KDE itself but can be found on kde-look.org - your stock-quotes examples are such plasmoids. Hm...that's the only example I can think of on top of my head. Then again, the plasmoids doesn't have to put all of their data in the plasma* files. The magic-folder one, for example, put quite a bit in one rc-file of its own.

As for the pastebin example...I'd say that the main advantage of having it as a plasmoid rather than a "standard" program is that it can easily be put wherever you want it to be. A "standard" program is either on the desktop or in the systray.

A bit longwinded, but I think this rethinking you think necessary seems to me to be necessary mostly for (some) third-party plasmoids and not the "official" ones. And "forcing" third-party devs to not write application-like plasmoids would probably not be doable without some _very_ strict guidelines of what's permissible and what is not.

Plasma Widgets can be launched like programs through the Plama Applet Viewer application.

Bascially just a different widget host other then the Plasma Desktop shell.

I don't get the point behind saying "what should be a plasmoid and what should not"... Plasma is Free Software, and that means you can use it for whatever purpose you want! If one wants to write full fledged applications as plasmoids... why not let them do it? It is their choice and their freedom fo using Plasma.

So there is no such thing, and absolutely shouldn't be anything like someone telling what "should" be a plasmoid or not!

If KDE would take this direction of telling what you can do with Plasma and what not... what would it differentiate it from becoming another Microsoft or Apple? Is it really what we want?

I hope not!