We moved the 0.40 prototype from Chris' on/off SVN repository to Launchpad and the move was worth it: 0.40 is making progress again and we're seeing new developers writing blueprints and uploading code!

Launchpad is a good hosting service. Easy management of all the permissions to the repositories, the bugs, translations, future plans and basic FAQs in the same place really helps focusing on the main point: code.

We already have three branches there: gdesklets0.40 (a.k.a. "trunk"), gdesklets0.35 and desklets-basic which will grow to contain a group of the best desklets and controls that can then be easily distributed as a package.

Visit the new launchpad page at launchpad.net/gdesklets

My three month employment with COSS and the summer of code is now over. I will no longer receive payment for my efforts on gDesklets :'(

But just when the gdesklets community thought they would finally get rid of me, they we're mistaken! Bwahahahah :-D ! I wouldn't dream of quitting work on gD now. There's still so much to do: getHTML control is nearly perfect, 040 core still needs a lot of love, DD awaits final touches and integration into 040, and we've just started moving management and stuff into launchpad.. not to mention the work needed to bring the new site up and working!

Overall the summer was magnificent. I'm quite happy with all the stuff I've accomplished and (especially) learned during the it. The workplan was suprisingly well estimated. Nearly all the tasks we're completed and the average workday was quite close to the basic 6-8 hours (very, very hard to estimate). To compensate for the "nearly" word in the previous sentence, numerous extra features that weren't mentioned in the workplan got implemented.

So what next? Well.. tomorrow I'll get up early and continue on the getHTML-control and I'll try to find the guy responsible for the gd.org site and suggest a few improvements.

In the future this blog will still be the central place to hear about my adventures with gDesklets. The only thing that changes is the (nearly) weekly blogging cycle and the (almost) consistent report structure.

Thanks to COSS and its supporters for making this possible!<

And btw, those that are looking for Weekly Report #12: it got lost in the coding category.. sorry about that!

[EDIT]Oh.. yeah.. forgot to say what I've done these few days. Well, I've worked on the getHTML control and written pages on the wiki. [/EDIT]

Last week I managed to...

• Release a bug fix + a few new features to DD. It can now wait for 0.40 to reach a more stable form. Then I'll continue on it so that it'll work nicely with it.
• A few plans and proposals about the GDE (Graphical Desklet Editor) and desklet packet structure. I'll post them on the wiki as soon as it gets back online (should be around the 30th).
• UrbanExplainer - a desklet (and a control) to query words from Urban Dictionary. A screenie here and the packet here.

• This week I'll continue planning the GDE and looking for possible implementation solutions. I'll also write a HTML-control to support easy HTML-fetching and analyzing through desklets (good for doing web-queries). I'll write a dict.leo.org desklet to test it.

This week I've done really long hours and accomplished...

• in releasing Desklets Desklet alpha version. It's actually closer to a beta, but not quite there yet. Of course the biggest downer is that it's designed ultimately for 040, which is not here yet... But the current DD can actually be used for desklet management :-D
• improving Desklet Control so that it now supports actual installing and starting and removing of desklets (network transparently :-o ).

I'm quite happy with this week. I managed to almost catch up with the workplan. Next week I'll continue doing long hours on the DD and start thinking about the Graphical Desklet Editor. Unfortunately the wiki is still down, so I'll have to do design locally. I would love to do the server side php/python to create a dynamic XML-file for DD, but I don't have access to the server. I'll ask if I can get one next week.

Traditional sunday report once again on monday :-)

• Completed the desklets Baskerville and RSSboy (screenie here).
• Got the GUI prototype from Ashwin. Lot has been done but A LOT requires doing. This will be my main task for this week. Long working days ahead...
• SVN is down and Christian (the admin) is in China.. I'm trying to get an alternate server to run on my laptop.

Approximately one month of work left. Looking at the workplan things look good, but I feel that some explaining and contemplating has to be done on what has been accomplished and what remains to be done.

Done

All the things due by today in the workplan have been completed. As things started progressing quite fast, I've also implemented multiple extra additions to the core (i18n-support, Behavior-classes, children properties, etc.) and written a few real desklets that require a bit of tweaking (beagle searcher "Baskerville", RSS reader "RSSboy" and audio player "Boombox").

Since Ashwin took responsibility of the GUI frontend, I've been focusing more and more on the core and on the controls that the GUI will need. We've discussed deadlines with Ashwin and try to get DD ready in August.

To Do

• Complete the Baskerville and RSSboy desklets (this week)
• Close the conversation (here and here) over the control metadata and start working on control installation through DD (loading of controls still not possible in 040 through). Probably next week.
• Finish the Desklet control so that Ashwin can use it and DD works properly.
• Write a idea page for the desklet editor and start thinking about implementation (this has a low priority. I'll do this if there isn't anything I can do to speed up the release of 0.40). Due end of month.

And during the weekend I completed an early version of the Launcher control mentioned before.

This is a half-week-report. I'm going away for the weekend and won't be connected to the internet until next week so I'll explain what I'll do/I've done now:

• The Baskerville desklet for beagle searches is quite complete. It can be seen here and a screenie is here. It still lacks support for actually launching stuff from the found results...
• ...which brings me to what I'll be up to this weekend: the Launcher Control. This will make it possible to launch applications from desklets (user specified "favourite" apps). This is closely related to my other idea of "global preferences", that are shared among desklets (see Untargetted/Core section of the Brainstorming page)

Missed the sunday deadline, but anyway here's what I've been up to:

• Did the children property for container elements
• Continued work on Baskerville - a beagle seach desklet I used to test the search functionality
• Created a control for the Desklets Desklet (new gui). It does not function properly yet, because it's aimed for 040 core and all things have not yet been agreed on. Ashwin can however use it to test the display frontend.
• Wrote Behavior classes for the core - another of those a bit boring chores that needed doing to reach 040.

... and in other news: style support was moved to version 0.41, so I won't work on it this summer (see the Roadmap on the wiki). This week I'll be focusing on the Baskerville desklet and Desklets Desklet with Ashwin.

This week I've been mostly...
* hacking to get real DOM support in the core. This idea was however rejected by the others. The end result is to provide a "children" property to all container elements. This I'll probably do next week.
* implementing proper i18n support and writing a howto in the wiki.
* planning the style support. This is something I'll start implementing next week.

This week I was also contacted by Ashwin, a biochemist from India. He was also working on a GUI for gD. We've been hacking together on it now and it looks very promising.

* Started on the GUI. Discussion and shell mockups can be seen here
* Created DOMv2 - a modified version of the internal desklet structure that supports hierachical referencing. This was my first serious dive into the core of gD and it was rather positive - sparking a new wikipage and discussion: http://gdesklets.gotdns.org/gDeskletsWiki/DOMv2
* Next week I'll continue on the shell and improve DOMv2. Then I'll think about creating i18n support and style support as IMHO these are features that should be implemented before the new GUI itself.

DOMv2 can be taken from the SVN repo or seen here

* Improved the control documentation and wrote the DTD as described in the workplan
* Fixed the RSS-control and started work on RSSboy - a simple RSS reader desklet. Currently the desklet reads one source nicely, but I'm planning on adding support for multiple sources
* Next week I'll start thinking about the new GUI for gDesklets. Instead of creating a normal GUI with GTK, the whole user interface might be a desklet that is capable of controlling other desklets. This is something we've already discussed in the wiki. My other plans include adding real tree style DOM-support to the parser as described on this wiki page.

And here is the first exception to the Sunday rule :-).

* Beagle control works, but integration into gdesklets-core is problematic (running gobject.mainloop inside another mainloop seems to be the problem)
* Seeker - a display to use the beagle control is under development
* AudioPlayer control was improved in a collaborative session with the collaborative editor "gobby". This was the first time I used a collaborative editor and it seemed to work quite nicely! Participating we're both of the main developers (Martin and Christian) and me.

* Next week I'll improve the docs as soon as Beagle control becomes stable or stablish..

And as usual the fruits of labour are available here

Came back from the trip and started working full-time again. Accomplishments:
* AudioPlayer-control works. Segfaults removed, interface improved.
* BoomBox is born: a desklet to test the functionality of the audioplayer. Screenshot here.
* Now I'll move on to writing the beagle control (might be difficult, 'cos I can't find any docs)

And from now on I'll write the weekly report each sunday

Haven't really done anything since leaving Austria and arriving in Germany. Mainly because I have no time and no constant access to a computer. Have been checking my mail though and there's been some discussion about a terrible bug on the gdesklets site. I'd be happy to fix it (quite good with php), but I don't have the "resources" right now :-(.

Another topic is JavaScript support for gDesklets. This way we could possibly display Apple's Dashboard widgets. This started the discussion. I agree with Christian and Martin: encouraging DHTML(HTML+CSS+JS) instead of clean Python+XML is not a very good idea. However the possibility of supporting JS as a scripting language is not entirely discarded.

Working from Klagenfurt, Austria.

* Wrote a bunch of pages in the gDesklets wiki (control docs, personal info, ideas, etc.).

Did some work on the controls as described in the workplan. The early alpha versions are available here.
* GST-control works, but has some known problems
* RSS-control should work nicely, but I haven't had the time to test it yet

Today I'll head out to central Europe and I'll be back on the 18th of June. During the trip I'll try to get a part of the documentation done and perhaps start on the DTD.

I started working today. I'll try to get ahead as much as possible this week, since I'll be cruising around central Europe from the 22th until 18th of June. I think I can get some work done on the road as well, but I can't promise anything.

My workplan remains the same and it can be viewed here. I think I'll manage doing few of the controls this week and that way I won't get behind in schedule.

Here you can read about my progress regarding kesäkoodi 2006 work.