The aKademy 2005 KDE Developers Conference finished yesterday with a second day of talks to prepare for KDE 4. Topics of the day included integration with other programming platforms, marketing KDE and accessibility. In their keynote, David Carson and Deepika Chauhan from Nokia described the challenges involved with porting KHTML to the series 60 platform. Another highlight was Novells desktop migration study. After four days of conference talks the KDE developers are now into a 5 day hacking marathon which will feature not only 24 hour non-stop coding but more spontaneous BoF sessions and two days of usability sessions.
The day was opened with comedy double act Aaron Seigo and Waldo Bastian on Marketing for Geeks (Slides, Transcript). "We suck" Waldo said, "I was talking to my mother the other day and we make all this great software but she doesn't know about it". The duo went on to give ideas on how to get our software better marketed including grassroot marketing like blogging and the value of cute furry animals.
Continuing the theme of presenting KDE to non-technical users better was a talk from Rainer describing how to make multimedia presentations. By using VNC recording and voiceovers we can describe our software much better than with plain text. Look out for more dynamic previews of KDE in future feature guides and documentation.
Meanwhile Mirko Böhm took a "Walk on the Dark Side" reporting on strengths and weaknesses of Java and .NET. Richard Dale gave one of the most involving talks on The State of KDE Language Bindings (Transcript). He has worked particularly hard on the KDE Ruby bindings, Ruby he says is like going back to his pop-11 roots from 30 years ago because it is so much more interactive than programming in C++.
Nokia's talk was conducted by two engineers from the Series 60 webbrowser team. It covered their experiences of porting KDE's web rendering engine KHTML to the Series 60 mobile phone platform. After a brief history of mobile phone usage and lot of technical details they outlined the challenges of porting KHTML/Webcore to a very constrained operating system. The KHTML developers seized the opportunity to get in touch with the Nokia representatives. They were keen to emphasise that they did not want to fork KHTML. They feel able to contribute back a lot of their changes in the future.
Domas Mituzas, developer for the Wikimedia Foundation and contributor to Wikipedia described details on the worlds largest and free encyclopedia project, its contributors and the technology behind it all. He then described the importance of metadata and proper ways to index data in order to make them available to third parties like KDE, e.g. via WebServices.
The Windows Desktop Migration Study (Transcript) had a few surprising examples of how users interact with KDE. They gave their testers 5 basic desktop tasks and videoed their actions. One task that caused a lot of problems was adding a new user, most of their subjects found Switch User in the K-Menu and ended up trying to create a new user through the login manager.
The KDE OpenOffice Integration talk from Jan Holesovsky (Transcript) described how they got OpenOffice's Native Widget Framework to use KDE widgets, finding and creating 1500 icons and the integration of OpenOffice with KAddressBook.
The conference closed with a presentation on the new Oxygen icon theme which looks set to become the new icon theme in KDE 4. They are keeping the exact look under wraps for now to keep it fresh for KDE 4 but the ideas include a more serious colour palette and well defined concepts for different icon types.
The kdelibs and kdebase restructuring BoF came up with a new structure for these core KDE modules in KDE 4. This will give ISVs more flexibility to use KDE while being easier to port to other platforms and have smaller memory footprint for non-GUI applications.
The coding marathon now goes on until Sunday. KDE's resident usability experts have announced their usability tracks which will take place on Thursday and Friday.
Comments
Why would you want that? That makes the editor dialog redundant And how many times do you change a person's name? Not that often I guess.
Probably because Mac OSX has it. ;)
We're making KDE, not a clone of Mac OS X. Just because Mac OS X has some feature doesn't mean that KDE should have it too.
You missed the small smiley.
Yeah I know. In a few minutes I'm gonna pack my stuff and go the residents to have some sleep :)
well, if it works without making it more complex, why not make the editor dialogue redundant??? it is a complex piece, imho. things like name, adress and picture should be much easier to change. at least I tried to change these without opening edit, the first time. i thought it should work...
btw i dunno how mac OS X does it (don't have it) so that has nothing to do with it. but if they allow easier changing of this information, kudos to them. generally, one only changes name, adress, im, picture, birthdate and categories. if you could change these in a easy way, why not allow it?
"They found that good icons are nice but words are what really helps people."
Another thing I've known for a while. Make toolbars have text under the icons, where possible, by default. Don't just give them a load of incomprehensible icons. That would help users no end.
well, it is easy to set up, and it'll be available in all KDE applications. search in the control panel....
"well, it is easy to set up, and it'll be available in all KDE applications. search in the control panel...."
No, no, no, no, no, you're side-stepping it in the usual KDE fashion. It has to be on by default if it makes sense, and it certainly makes sense. No one should have to go looking in the control panel to turn something on like that.
Is it so hard to right-click on a toolbar?
And it doesn't make sense everytime. On 1024x768 display and lower it uses a lot of screen space. And once you've learned what an icon means (which you can do also by moving the mouse over it) screen space is much valuable than a text which is redundant for you (as you already know what the icon means).
"Is it so hard to right-click on a toolbar?"
Have you read the transcript of the desktop presentation or seen the video? The point is, you shouldn't need to to it for something like that. Icons by are a good visual aid, but by themselves they simply don't help. Text is what's necessary, and that's what he was talking about.
"And it doesn't make sense everytime. On 1024x768 display and lower it uses a lot of screen space."
That's not a user's problem. It should scale adequately.
"And once you've learned what an icon means (which you can do also by moving the mouse over it) screen space is much valuable than a text which is redundant for you"
No it isn't. What we're talking about here is the time it takes you to think about what icon it is that you need, even though you might 'know' what it is. You have to be able to use the whole desktop more unconsciously than that. Those few milliseconds, or seconds, that you spend dithering about what icon you need makes all the difference.
> No one should have to go looking in the control panel to turn something on
> like that.
Fully agree. Besides, the real problem isn't even that one has to go to the control center, the real problem is that users do not know that this one precise setting would help them, so they don't do it. Not only are they non-developers, they also have no knowledge of what would make their desktop more efficient. Users don't (and at this point I'm fairly sure that I mean all user categories) - they don't run usability labs just to find out what would be better for them.
This means that providing 'sane defaults' is also a way of providing valuable information, and not providing good defaults whilst expecting the users to reconfigure their dektop means a lot more than expecting them to "right click there etc." It means expecting them to also know _what_ to do, not only how to use the kcontrol or the toolbar editor in order to do it.
* Only authenticated users with the relevant permissions should be able to add users. Adding a feature that will allow new users to be created from KDM or GDM is insane.
I disagree. On my HOME systems I want my guests to be able to create a personal account on my machine from KDM. At work this feature makes no sense. At home it would be nice to give them an automatic account. Right now I have a guest account, but the password is a target of a lot of ssh attacks (second only to root). (I mitigate this by setting the shell to nologin in /etc/passwd, and having kdm allow passwordless logins, but this isn't a good idea)
Some features only make sense for some people.
I am reading all the blogs about akademy and this year seems really impressive. Lot of things seem to be going on, and we have a big number of hackers together for more than one week. A lot of things will certainly come out of this.
Who could think 5 or 10 years ago, that free software would evolve to a point where hackers can meet on an international basis for hacking free software, with corporate sponsors ?
Looking at the linked webpage for Oxygen, I got a bit worried:
"1 - for hardware like printers, usb sticks, monitors and so on : Realistic design with a frontal high view when possible.
2- for media things like CDs, floppies, memory card I would like the same perspective of mimetypes, with the object lying on the desktop."
Isn't this a plain copy of the ideas of the Mac OSX icon set? :-\
"Isn't this a plain copy of the ideas of the Mac OSX icon set? :-\"
Maybe OS X got it right? Suppsoe that Windows had icons that resembled things that are related to the task (like, an icon of printer for printing). Would KDE be "copying Windows" if KDE had icons that resembled things that are related to the task (like, having an icon of printer for printing)? Would KDE have to do things in different way, even though the end-result would be harder to use, just so KDE wouldn't be "copying Windows"?
If some other OS does something the way it should be done, there's no harm in doing it the same way in KDE.
I was referring to the more stylistical aspects: "realistic" and "frontal high view" of hardware devices, "object lying on the desktop" for mimetypes ('paper sheet' icons).
Remember when BeOS came up with icons with isometric perspective and many free icon sets (including KDE, in 1.x/2.x era) copied it?
Then it was the translucent mania, with "bubble"/"aqua" icons everywhere (Crystal), copying the overall looks of Mac OS X.
Looks like it's the same thing, only copying now the style of the Mac's photo-realistic icons (frontal high view of devices, etc.)
There's nothing inherently wrong about that, but I'd really like to see KDE have an 'original' look and I thought this time would be it. I don't mean to bash the excellent work the KDE people is putting into it (I thank you all) -- this is just meant to be constructive criticism.
So, if any of the Appeal guys is listening: please don't make icons that look taken straight out of the Mac. Actually, that would be a good test: put some new-KDE icons in a Mac desktop and vice-versa; if they _don't_ look out-of-place, then we have a problem... Icons are a key component in an environment's visual identity; KDE icons need a look of their own.
I get a bit worried about the apple style 'keep it under wraps and make a big splash in kde4' approach to the Oxygen icon set. The icon examples on the web site look beautiful, but surely one should follow the open source 'release early, release often' mantra in this respect too? If you are going off course with the icon set in terms of useability etc. it would surely be good to know early on in the process. I am sure it will still be a surprise to Joe User when Qt4 comes out as they are unlikely to be following the development process anyway, and those that do care about these things could at least contribute suggests during the process....
Just my 2x. The oxygen icons that you have previewed look lovely so keep up the good work anyway...
Tim
Just in case you missed it: there will be alpha and beta releases of KDE 4 as well. There is no need to make the set popular and in widespread use long before the set is finished and system of which Oxygen is supposed to be part of is even publically available outside svn. What you see here is a concerted effort at doing better promotion of KDE as a whole.
Fair enough, point taken. Thanks for the great software all involved!
Hi
I have been using vnc2swf recording in the past to make screenie movies of QGIS (e.g. on http://community.qgis.org). The talk mentioned above describes how to add voice overs. It there a link with more details? I know vnc2swf lets you add a soundtrack in mp3, but its difficult to record the voice first and then the video. Did the speaker describe a way to record the voice after the screen video was shot and then add it to the video?
Many thanks
Tim Sutton
From home page...
Q: How to record audio?
A: Record your audio separately with an MP3 encoder like lame, and combine it later with edit_vnc2swf. Simultaneous audio/video recording is planned in future.