I thought I’d take some time to review candidate responses for the Online Desktop question of the debate. Not just because it’s a topic I’m interested in, but because I think it speaks a lot to how the candidates view the Board, its powers and goals.

I’ve highlighted some of the candidates responses and bulleted them with headings to highlight how I felt their positioning came across. In the end I tried to supply what I felt were the responses I was looking for. Perhaps a bit late, this is in no way an endorsement of any candidate. This blog is more of a perspective view on candidates from a certain angle and you are encouraged to look at all other angles before voting. (i hope I have all the candidates and they are in no particular order)

Jeff Waugh’s Responses

Excellent:

• where the GOD developers have an idea that is unrealistic without some kind of central support, I want them to know they can lean on the Foundation.
• I don’t want our software to suffer because we don’t have the organizational cajones to support great ideas.
• Should the Foundation start hosting services?

Good:

• GOD is only one of the many interesting opportunities we have right now
• GOD raises some interesting questions for the Foundation
• What can we provide that supports our community and commercial ecosystem, without creating unnecessary competition?

I think Jeff has a clear understanding of the Board’s role in assisting GNOME developers and is looking to tackle the relevant questions. I’m not sure where he’s going with the commercial and community ecosystems but I assume it has to do with making sure that distributions aren’t fighting GNOME for the services space. I’d love to hear more about that.

Diego Escalante Urrelo Responses

(Lawyer-ish) Good:

• … the role of the Foundation with this project, I’d say that it’s important to support this and other initiatives with the resources they need -as long as it’s realistic and reasonable-.

Diego has an excellent stance on the foundation supporting projects GNOME where developers need help while he’s not specific in the application of help it leads toward what seems relevant and his -tag line- seems like he stole it from Luis. Short but on target.

Luis Villa Responses
Excellent:

• … to make sure that as we sail into uncharted waters, the rights of GNOME users and contributors are being protected.
• I think the board’s role in technical/strategic issues like OD and Services should be fairly minimal, generally.
• … the board then needs to understand that (technological) vision and help grease the way for it

Luis seems to see that GNOME developers are going to need support in the areas of infrastructure yet understands that it’s not for the board to decide which technology is best but to understand what the technology requires and how the board can support that.

Vincent Untz Responses
Good:

• (The) Board should not be making technical decisions

Vague:

• Might be able … (to provide) definition of free services

Bad:

• And maybe (blue sky dream) providing some infrastructure

I like that Vincent understands the board cannot tread in making decisions about technology or vision, however I feel he falls short by asserting that the board probably can’t provide needed infrastructure for GNOME developers.

John Palmieri Responses
Good:

• (The) Boards mission in this is to not set direction but grease the wheels

Vague:

• This could include procuring hardware for applications to run on
• … Facilitating talks between the different interested groups.

John has a clear understanding that the Board is not setting technical direction, however he’s a little vague on whether they can help when GNOME developers may require hardware infrastructure.

Og Maciel Responses
Good:

• I believe that it is not up to the Board to decide on the implementation or even which tools/languages to use

Vague:

• … but serve as a facilitator

Og clearly states that the board does not decide on implementations, however is quite vague on areas where the Board can be active. I’d like to hear more about the extent of the boards activity from his mind.

Lucas Rocha Responses

Good:

• FLOSS projects and other organizations, and making sure that hackers have the necessary infrastructure available.

Vague:

• I think the GNOME Foundation (and the Board) can help the Online Desktop initiative by bringing this topic for discussion to the Advisory Board members, promoting cooperation among companies.
• … discuss about the wider topic of free (as in freedom) web services (something that Luis is already investigating?).

I like how Lucas (reminded me I didn’t have comments for him) seems to see that FLOSS projects are requiring more and more infrastructure. I had put “shaky”, like unsure, originally for his last comment but it’s really more vague. Lucas seems to have an understanding of what’s ahead in his comments, but hopefully will look deeper into the issues as they are brought forward.

Behdad Esfahbod Responses
Good:

• I just leave that to the GOD and Gimmie developers and others who have visions and are doing the work already.

Bad:

• … I don’t think that’s relevant to the board.

Behdad has an excellent hands off attitude when it comes to technical direction, however he seems to fall short by feeling that this topic isn’t relevant to the board. Other candidates have expressed how this is a question the board is going to have to tackle head on and I’d like to have heard more. I left out any personal assertions for you and George.

Brian Cameron Responses

Excellent:

• I agree with Jeff Waugh that it is a good goal for the GNOME Foundation to provide online services which can be used with the client applications that the GNOME community is intending to provide.

Good:

• I think the GNOME Online Desktop is an important initiative, but one of many.
• I also think it would be valuable to outreach to other popular online desktop tools and develop partnerships, if possible, to make the GNOME Online Desktop a premier and supported interface for accessing their tools from the GNOME desktop.

I thought Brian had outstanding responses. Understanding that the Online Desktop is just one initiative among many others and that the board doesn’t play a role in deciding which ones should become the direction of GNOME. Also stating that the role of board is the facilitate directions the community chooses such as providing online services for the GNOME Desktop. Well done!

George Kraft Responses
Good:

• We need to make sure the Online Desktop does not fragment the GNOME community.

Bad:

• In the 90s the world wide web, the Netscape browser, and Java was the death of UNIX’s new Common Desktop Environment. Many started to say “the network is the desktop”, then some of the consortium members lost interest in CDE.

Really Bad:

• In short, I think an online desktop makes more sense for a mobile device or network computer, but not for my laptop or desktop which have a hard drive.

I’m sorry to come down on George this way but I feel your responses do not reflect those that I would want representing me on the Board. Maybe these responses are being misunderstood by me and I think you’re welcome to mail the foundation list with additional information on your position.

You’re fighting a technological direction people in the GNOME community are taking, and I don’t believe it’s the Foundations job to take stands like that. What if people were to write core GNOME applications in Ruby? Would you fight them because you don’t feel Ruby is fit for the desktop, only for the online rails world? If you don’t understand what the goals are of the Online Desktop I can understand that, however making judgments like these without understanding is reckless; especially as a Board member.

I like that you want to ensure the Online Desktop does not fragment the community, that’s a good cause for any new movement inside GNOME.

Online Desktop or Technology Change?

Though the question is about the Online Desktop, in my mind it’s really about a shift in direction and technology and really tests to see that Board members are open to those shifts. It’s the responsibility of the Board members to understand the new technologies and try to enable the people working in that direction where reasonable and prudent for the Foundation to do so.

I’m looking for candidates with responses like these:

• The Online Desktop is one of many initiatives, it’s important but the foundation needs to weigh that against it’s other priorities.
• The Online Desktop and other GNOME technologies are requiring server infrastructure more and more, the foundation can provide a way for this to happen. This is a priority.
• It is not the job of the Board Members to choose technologies or directions for GNOME but it is my responsibility to understand them and enable them where I can.
• If the Foundation is to supply a type of online services to GNOME members then an Open Services Definition would be a priority but not a road block for change

update: some weird wordpress bug seemed to mangle my earlier post, i believe it’s all fixed now.

In a slightly unscientific poll I attempted find out what users prefer, Hardware Sound Mixing or Software Sound Mixing.

Here is a GNOME user enjoying Hardware Sound Mixing, he looks happy and carefree.

Here is a GNOME user enjoying Software Sound Mixing, he also looks happy and carefree.

Thus the toss up, users seem to prefer each type just as much. One could say a lot of users (people who listen to music) don’t care where the sound gets mixed, in hardware or software, just that it does. Andrew got to the point that if you saw the pulse audio demo at the GNOME Summit you’d have been impressed, it allows us to be really reactive and dynamic with sound where we haven’t at all before; a huge step forward that’s available now.

I think Lennart’s mail while long and possibly abrasive to some people was pretty refreshing and right on at the same time. Right now we seem to be at a situation where audio sucks for most of our users all of the time because there is no mixing happing and no integration at the GNOME level. Things suck almost as much for people with hardware mixing because they have to go search the web for a solution to use their hardware mixer at the ALSA level. PA seems like it will at least bring GNOME forward for the majority of users who are likely having issues with sound, working to make hardware mixing work just as well seems like an parallel process, not a conflicting one.

And there’s no reason we couldn’t do much better than the vista sound mixer, we try. I’ve already gotten started and will need to post more mockups as I have them.

[Picture by Flickr user Fanboy30, used under a CC-BY-SA-NC license.]

Got a great comment recently about my mashing google calendar and gnome post from a while back. I had tried to use the command line in the original entry and it brought lots of issues because every distribution of linux seems to place the evolution-webcal application in a different pace.

The new solution isn’t anything grand, it’s one of those so simple you hit your head afterward.

Put the Calendar URL in the browser window

By simply copying the calendar URL into your browser address bar and prefixing it with webcal:// while removing the URLs http:// the browser will call the correct application automatically.

Here’s a form that tries to help you out »

Reading Alberto’s recent post let’s make it easier made me think of last years New York Times article A Star is Made.

The article mentions computer programming and I do think it applies well to GNOME. Some people have been practicing (contributing and coding for GNOME) more than others and many of those people were probably able to get lots of positive feedback on their early work which encouraged them to continue working more. And thus a GNOME contributor is born!

I don’t think I would have started with GNOME had there not been such a quick feedback loop for working on the project. Luckily for me, friends of mine and I had created a group at school called the Open Source Institute where we could get course credit working on lots of different Open Source projects to learn about the software creation process. Beyond our small group of GNOMErs meeting other GNOME hackers at GUADEC and the Summit was a great way to feel more connected to the people responsible for all this cool work.

Just a small suggestion to not only focus on making getting started easier but creating a good feedback loop for new people. Perhaps using communities like GNOME Love to share patches made between new contributors.

I never used to be much of a calendar person, usually I’m late to everything and never know if I’m free for a weekend or not. But since Google calendar came out I started using it pretty regularly and I think it’s actually improved my ability to be punctual and organized… well probably not, but anyway.

My Google Calendar

One major benefit I find to using this calendar versus one on my local system is that I can have GNOME import my calendar locally via the iCal and still access and use the calendar via the web interface any other time.

To set it up I went to the manage calendar page for my personal calendar. I had to right click copy the link location since they don’t give the iCal link in webcal: url format. Then open a terminal and type this, replacing $URL with the url you copied for your iCal.

/usr/libexec/evolution-webcal $URL

My Personal Calendar Manage Page

And that’s about it, the little webcal dialog will ask you silly questions about your web calendar settings and it’s imported into evolution-data-server. I don’t even use evolution anymore, but the calendar portion works just the same. I only ever view my calendar out of the calendar widget, all my updates and additions are done in Google calendar.

My Calendar Widget

I get a lot of bug mail, way too much in my mind and I’m sure there are people out there getting much more bug mail than I do. The issue isn’t so much the amount of bug mail, though that’s often a problem. The real problem I have with bug mail is how often it’s completely useless.

Who are these people?

Lets start off with some better identity. I see a lot of names on bug mail for whom I have no idea who this person is.

I’ve noticed the bugzilla web interface has started to include a points system for people who comment on a bug. I don’t really know what the points mean but that doesn’t matter, I know higher is better. Somehow I’m at 16 points right now so I guess I should pay more attention to people with at least that many points. I actually read all the bug mail I get, but it would be nice to know if I could at least feel less guilty about not responding quickly.

Showing the points of the person making the change (and my points) in the bug mail would help this out a lot. I do also see some of the people who comment have titles like “Evince developer”, this would be another great thing to show in the bug mail.

Last part on identity, it’d also be great if there was a link included in the bug mail to this persons profile. I can’t seem to find the idea of a bugzilla profile though. One page about this person, their points, where they comment, what they are a developer of, if anything.

What just happened?

Removed: NEW – Added: NEEDINFO

Who you need info from would be a great addition which most likely requires lots of changes in bugzilla. However the least that could be done is to include the previous comment and commenter. I think that would help out a lot of other types of changes as well.

Bugzilla comments often work in the form of a conversation that slowly takes place in message board style. It’s like my roommates and I trying to coordinate feeding the cat on the cork board, a slow and painful process that could be easily solved with a minute or two of face to face time.

Below the current comment, prefaced with some text saying: Last Comment on $DATE by $PERSON [$PERSON_PROFILE] and then inline the last comment. Perhaps keeping a limit on the number of characters if you have a desire to keep the emails to a small size.

Anyway, I’m not intending to gripe about GNOME bugzilla as I do think it’s one of the nicest bugzillas I’ve used. Kudos to the developers! And I realize that bugzilla wasn’t really designed for interaction through email, which is too bad because I really think it’s the better way to go. This is just some of the fixes I’d been thinking about to improve my bug email usage, haven’t looked into if they are applicable to everyone.

Colin: I guess I said “launchers” in that entry and that was a mistake. Part of what’s missing in the notifications design is how applets and “desktop service” applications work together. In other words how do we handle programatically added applets and applications like Rhythmbox and Gaim, so we’re not congesting the notification area up like what is happening currently.

What I had talked with Davyd and Mark about during GUADEC this past year was the design we had been working on for how to add applets to the panel. This design also takes into account how applications can add an applet to the panel programatically.

Panel Cursors

The first change to the panel applets system is the idea of “panel cursors”, which are places on the panel where an applet or launcher could programatically add itself to the panel. These cursor locations are described in a file that corresponds to the panel layout design, thus different panel layouts can have different cursor locations.

There might be ways of categorizing panel cursor locations for different things. So we might have an area where “Hardware Events” would normally show, and an area where “Music Players” would normally show. This can be done either through programming or just general policy. I think the policy way would probably be easiest, we could add a section to the HIG describing where applications types should appear in certain panel layouts.

The Add To Panel Experience

Immediate reactions to this design are “You can’t lock my applets down to one location, $insert_reason”. In this cursor layout the Add To Panel dialog, seen below. Has checkboxes for applets, which add them to the panel in predetermined locations. You check off an applet to add it to the panel, it appears and you are allowed to move it around to any location you would like to have it in. Thus no evil-applet-lock-in!

How applications add themselves to the panel

For an application to add itself to the panel it would have registered the cursor location that it believes it belongs in. This registration is something the source code provides when the application is installed. When the application needs to add itself to the panel, it just requests the addition and appears in the place it is supposed to appear. Again, -people could then move the applet around if they don’t like the location it appears in initially. And applets should remember where they’ve been moved to even though they aren’t on the panel anymore.

So back to Colin’s point. Launchers aren’t what we want to use for these notifications we want to use applets, however we need a better applets system.

Norstrum is my Personalized Google Word of the Day. It means: a questionable remedy.

Saw some discussion of notifications on Rodrigo’s blog and the desktop-devel-list this week. I had done some design work on desktop notification a little while back with Red Hat Desktop team so I figured I’d share what was done.

Some important things that came out of this design about notifications:

• Notifications are just like regular dialogs but they cannot steal focus.
• There is no priority in notifications, they are all low priority messages. High priority messages could be regular dialogs or take the whole screen if they have to.
• Bubbles should come out of the objects that are responsible for them, or the notification area if there is no responsible object. See I got your notification right here pal
• The major separation in notifications are “Can interact with” and “Cannot interact with”. Either it’s just a message, or it’s a choice that doesn’t really need to be made by you.

I made the point that notifications are just like dialogs because there is a tendancy to over-engineer them so they won’t be annoying. We should resist this tendancy. When designing Gtk+ it would have been ridiculously futile to try creating an API for dialogs to make sure that people wouldn’t abuse them. People will abuse them no matter what, we just need to design the applications to be smart, not try to be triple smart in the API.

So that’s all, below I have some scenarios where these types of notifications could come in handy. We used these to test out the initial design ideas.

Battery Running Low

You’re running on batteries and your computer needs to let you know it’s about to go down. A bubble up describes how much time you have until power failure. Interaction isn’t possible with this and disappears after a medium timeout. Bubble comes from the battery applet.

Krb5 Re-Auth

Because your network admin is cruel you need to re-key often. This bubble appears under a security icon that appears in the notification area, asking you if you’d like to re-auth or ignore. Interaction is possible and brings up the re-auth dialog or goes away. Disappears after a medium timeout.

VPN Login Message

After selecting a VPN connection to login to, there is a bubble up message showing that you’ve successfully connected and any useless message that your VPN provider has included. No interaction is possible, except maybe to close this. Disappears after short timeout. Bubble comes out of the networkmanager icon.

Switching network after lost connection

When NM decides to change connections after losing a wired or wireless connection. It tells you what connection it is changing to, or if there is no connection tells you that all hope is lost and you might as well destroy your laptop if you don’t have the interweb. No interaction possible, except through applet. Disappears after short timeout. Bubble comes out of the networkmanager icon.

Disk almost full

A notification that you’re about to run out of disk space with options to delete trash or clean tmp space or something like that. Interaction is possible to make space, but clears out after a medium timeout. Bubble comes out of the Trash? Perhaps a disk icons appears in the notification area? This use case sucks.

New device found

Immediately after you insert a new device your computer tells you it’s found it before it even recognizes it. No interaction is possible since we don’t know what the device is yet. Disappears after a short timeout and before device is handled by some application. Bubble comes out of a “New Device” icon that appears in the notification area.