Friday, May 24, 2013

all dogfood diet




Yesterday I walked to my local t-mobile store and had them cut my SIM down to "micro" size. I did this so I could fit it into my Nexus 4. I wanted to put it into my Nexus because I decided that it was ready for me to start using full time. I put my Galaxy II away. 

I decided to do this because as of yesterday I could:


  • Import my contacts
  • Make and receive SMSs
  • Make and receive phone calls
  • Use the internet via a wireless connection


It still lacks data over the cellular network. We won't get that until next week. So, I can't really say that it's dogfoodable for everyone as per our original goals, but we are close!

Friday, May 17, 2013

Dogfood Update


At the end of April, we set the goal to have Ubuntu Touch be dogfoodable on the Nexus and Nexus 4 phones. By that we mean, the goal is to make it so that we can use our phones exclusively as our phones. Today I chatted with some of the engineering managers involved to see how much progress we have made towards that. I am happy to say that it looks like we are still on track for this goal. However, there do appear to be some risky parts, so I am keeping my fingers crossed.


  • You can make and receive phone calls: Done!
  • You can make and receive sms messages: Done!
  • You can browse the web on 3g data: Tony had been blocked on some technical issues, but thinks he's through them, so is in the debugging phase. He expects to have this done by end of May as per the dogfooding goal. For me, personally, this is the only missing part for me to be able to use the phone as my main phone around town. So, if Tony cracks this nut, then I will put away my old phone and start using my Ubuntu Phone exclusively.
  • You can browse the web on wifi: Done! This has actually been done for quite a while.
  • You can switch between wifi and 3g data: There are 2 parts to this work. There is low level networking code to get done, and then there is UI to enable it. That means that the Phone Foundations team and the Desktop team both have work to do. Both teams expect to get it done for May, but the work is not started yet.
  • The proximity sensore dims the screen when you lift the phone to talk on it: There are two parts to this also. Gather the sensor data and then making the phone app use the sensor data. Work has not started for this part either.
  • You can import contacts from somewhere, and you can add and edit contacts: There is some work done on this that imports from a *.csv file. I expect there will be some crude support for this in time for the May goal. It might be fun for someone to try out a more elegant implementation. Ubuntu Phone is using Evolution Data Server for the contacts store, so there may be folks out there who already have the experience to do this easily.
  • When you update your phone your user data is retained, even if updating with phablet-flash: Done! This part being done makes the contacts import less important to me because as I add contacts they won't get blown away. On the other hand, it means it is worth it to import contacts, since you won't have to re-important as you update your phone each day (while it is in development).


Thursday, May 16, 2013

Feel Like Friday (post-vUDS)


It feels like Friday! Why? I think it's because I am tired. I am tired because Virtual UDS turns out to be surprisingly intense.

Power to People

So, that is to say, the second Virtual UDS is over. After experience my second vUDS, I think vUDS is really a boost for the transparency of the Ubuntu Project for a few reasons.

  • Frequency. We can do it every 3 months instead of every 6 months. As I mentioned in the opening plenary, this is important because we don't actually plan only every 6 months anymore. Like any modern software project, we are continuously planning. The 3 month cadence for vUDS means that there will be less time between detecting a need to change plans and discussion about how to make those necessary changes. I pushed very hard to have the first vUDS quickly, because there was a lot of planning for Ubuntu Touch that was backed up and needed proper discussion. If we waited until now, a lot of the work would have started without a good opportunity for discussion.
  • Access. Folks don't have to travel to wherever UDS is. People with specific interests can rock those interests with a laser focus, without having to dedicate a whole week away from home. Let's face it, traveling for 2 weeks a year to participate in UDS is something that only a few privileged people can swing. Many many more people can join a hangout.
  • Persistence. The sessions are streamed live, but then instantly available for reviewing, along with the white board, links to blueprints, etc... Try it. Go to Summit for the UDS that just ended. Find a session. Click on the session. It's like you are there live. Discussions that used to exist only in the memories of a select few with some written traces are now persisted and available.

Personal Faves

I won't go into a run down of the results, because that job is taken. However, here are some of my personal favorite discussions at this vUDS. These are my favorites based only on personal interests of mine. These are by no means the most important decisions or discussions. Just things that interest me a lot personally.

Rolling Release

After the unfortunate kerfufle last cycle when I pushed hard to move Ubuntu to a model of LTSs with rolling releases in between, it was niceto close in on one nice outcome. Namely, Colin has a technical solution that will allow users to subscribe to essentially the tip of development. Instead of using "raring" or "saucy" in your sources lists, you'll subscribe to a new name which is symlinked to whatever is the current development release. In this way, each day you will be on the latest. Even the day after a development release becomes a stable release, because the symlink will just point to the next development release.

I ended up with a couple of action items from this session. Mostly, to come up with a name and bring it to the next Tech Board meeting for approval. I'm very much leaning to "rolling", but I am open to discussion ;) This would mean you could say "I am on Raring", or "I am on Precise", or "I am on Rolling". "I am on Rolling" means that you are on the tip of development. Fun!

Touch Image Testing

I've been very keen to get Ubuntu Touch out of "preview" mode and into our standard development processes so that they inherit all of the daily quality tools that we have in place. This means moving all the code of out PPAs and into the real archives, so that we get the benefits of all the efforts we have put into place around -proposed and archive maintenance. It also means getting smoke testing and regression testing automated on the Touch images. I loved hearing from the Phone Foundations team and the QA team about their vision for "not accepting regressions". We should have dog-foodable touch images as early as the end of this month. Then if we can keep the images fully usable with minimal regressions each day, we will go very fast towards completion.

Ubuntu Status Stracker

I am partial to this topic because the status tracker started out as a labor of love for me. The first real bit of code that I wrote after joining canonical was to render my version of burndown charts. If I am not mistaken this code is still in use. In any case, status.ubuntu.com is critical to maintaining our planning, and ensuring that the status of the project is visible to all.

Unity 8 in 13.10

While 13.10 is very very focused on Ubunty Touch for phones, we all know that the real prize is the fully converged client OS. With that in mind, I think it's important to get the code up on as many device types as possible as soon as possible. There was a rich discussion about the steps to offer Unity 8 on top of Mir as an option in 13.10. Now, keep in mind that the result will only be the Phone UI on the desktop, and the default will be the Unity that we know and love today (with Smart Scopes and other enhancements of course!). Still in all, I am betting that basing Unity 8 on QML means that it will be surprisingly functional on a desktop even though it won't have any real desktop support in terms of things like workspace switching, etc..

Monday, May 13, 2013

A little bit of reusable code Sound Button

Want a button for Ubuntu Components that plays a sound when pressed, and stops when released? Here's SoundButton . Perfect for your typical Sound Board type app.


import QtQuick 2.0
import Ubuntu.Components 0.1
import QtMultimedia 5.0

    Button
    {
        id: soundButton
        property string soundUrl: ""
        onPressedChanged:
        {
            if(pressed)
            {
                audio.play()
            }
            else
            {
                audio.stop()
            }
        }

    Audio
    {
        id: audio
        source:soundButton.soundUrl;
    }
}


Wednesday, May 8, 2013

Woof woof!




Last week I fell into a discussion with Mark, Pat, and others about the importance of being able to really use a piece of software to really know how far there is between where you are, and a shippable state. Of everything that is missing, it's hard to know what is really the most important unless you can really use it and find what you have to work around, versus what you can just do without.

Out of this conversation was born the idea that we should drive as hard as we can to making it so that we can use our phones with Ubuntu Touch as our real daily phones as soon as possible. Really eat our own dogfood, so to speak. woof!

So, we committed our teams to making it so that by end of May, the phone images will be usable as our daily phones, defined as the following:

  • You can make and receive phone calls
  • You can make and receive sms messages
  • You can browse the web on 3g data
  • You can browse the web on wifi
  • You can switch between wifi and 3g data
  • The proximity sensore dims the screen when you lift the phone to talk on it
  • You can import contacts from somewhere, and you can add and edit contacts
  • When you update your phone your user data is retained, even if updating with phablet-flash

We believe that at least some of us will be able to really dogfood if we accomplish that. Of course, there will be a lot missing. Off the bat, I can thinking of things like the ability to find and install new apps, hardware not working on certain reference hardware (camera on Nexus 7 for example?), lots of missing features in existing apps, etc... However, in my experience, progress accelerates when people are using, in addition to building, software.

Tuesday, May 7, 2013

Ugly Duckling to Beautiful Swan, or How an App Developer Benefits from Designer/Developer Collaboration


Last week I snatched an hour here and there to work on my Feedzilla app. I like Feedzilla because it has an api that is free for me to use, so it's easy to write the app. However, I'm not totally enamored with the content, it seems like it is often out of date, though I suppose I can apply a filter to limit the content to new stuff from this week, or whatever.

However, what really stopped me working on it was that my implementation was just depressingly ugly. I'd look at all the cool and beautiful things that other people were doing with their apps, and be totally unmotivated to work on TechNews. Last week, I decided to ask for some help in how to improve my app, and I was told about ListItems. For TechNews, it was like the sun coming out from behind the clouds.

Now, the thing abut Ubuntu.Components is that the project is fundamentally a design project. Yes, the components need, and have, an awesome development team that makes them "real", but the components are really about providing developers with the tools for making a well designed "Ubuntu App". This couldn't be more clear than when using ListItems.

For an example, to turn the categories list from this:
My very ugly list which was my honest best effort without design help.

to this:
My now lovely list that I got to be that way just by using the right components and inheriting all of the designers' knowledge and talents.

I just had to use Standard list items. First, I went ahead and imported the ListItem namespace:
import Ubuntu.Components.ListItems 0.1

Then this is what my delegate for each list item looks like. The "progression: true" declares that the item will navigate somewhere. The designers ensured that this means the list item adds that ">", so it is standard navigation in all apps!
    delegate: Standard
    {
        progression: true;
        text: articlesListView.model[index]["title"]

        onClicked:
        {
            articleSelected(articlesListView.model[index]["url"])
        }
     }

So my app went from ugly duckling to beautiful swan just by using the right components and getting all the benefit of the designers' abilities that I so sorely lack. Thanks SDK team!