Facebook Home

So I thought I’d spend a few moments writing about why I am not too interested in a social phone, and basically just bitch and whine about why I am not too happy about having all my phone interactions to be  viewed through the lens of facebook.

There are probably 5-10 people I call on a regular basis, I don’t think i’d call more people if my phone was more social, would I post more to facebook ? Would I upload more photos ? Probably not…

Where is the value add for the consumer, I sometimes wonder if companies invent ideas to satisfy there need to do something rather than satisfy a need in the market.  My most pressing concern with facebook at this point is that the software is slow.  I can’t really go into why it is slow, because the code is not open sourced. But compared to Google+ it is molasses.

I very well don’t want something that is slow / buggy to take over my homescreen.  I dislike this concept immensely. (Given that my phone was released approx 2 years ago)

A new idea:

The problem I have with my phone is around the following list.  List is not ordered according to any priority.

1. Privacy: I don’t know what most apps are doing.  As a user you have no clue when an app is gathering data, and when it is being disseminated. I don’t think giving an app blanket permissions for geolocation access is enough, I generally want to know what data it is using and when. I want to be able to analyze this data. (Tough problem to solve)
2. Data Segregation: I would like data segregation to happen further up in the app ecosystem so data is stored in an app independent datastore.  Apps should share not just activities, but data.  Ex: After my run, I would have a copy of the geolocation information on my phone, and have it setup to share that information with Strava / mapmyrun..etc..etc.  What I don’t want is for a specific app to gather information, and then surreptitiously pass that information up to their service, and all of a sudden I dont have a copy of my own data. (Tough problem to solve)
   
3. Async: All apps should by default not block.  I think spinning wheel  = annoying, if the internet is not accessible, highlight that on the device, if there is a queue of events that need to be processed, and something is blocking then highlight that, an let the user make the appropriate choice. (Tough problem to solve)

I really dislike that there is this idea being pushed that for the user having a social phone is a good thing ?  When for me it doesn’t solve any of the problems that I actually care about.  If my friend is listening to Rhianna on his lunch break, which although interesting and confusing (why would any guy be listening to Rhianna on their lunch break), it unfortunately has a lower priority on my list.

Solving any one of the above problems and then blanketing my phone with navy blue facebook colors is more acceptable, but just giving up all your data without getting anything in return..

That is just silly.

Part III: Profiling Firefox OS.. An Entry point

Gathering profiling Data on Firefox OS can be done in several different ways.
#1) Simpler is better, sometimes a stop watch is a great medium with which to gather timing data for an application you don’t understand.

Stop Watch

2) In Application Profiling

A great way to accomplish this is to use timestamps in javascript to determine when specific things are happening, and if you are having a bottle neck insdie your application.

Ex:

var d = new Date();

console.log(“The time since epoch = ” + d.getTime());

Once you insert a few of these statements in “strategically” placed points.

Just cd gaia; make install-gaia, this will install the new version of gaia to your phone, and you can see what code path gets triggered at which points.

3)Measuring app startup Times

Turn on show app load times  by loading your settings app.

Click Device Information –> More Information –> Developer –> Show Time to Load.

This is one mechanism for measuring app start times, but this only measure time until mozbrowserfirstpaint which measure the time until first pixels are available for an app.

More details can be found in Bug: 787378

Part II: About Memory Gathering / Dissecting

I have my B2G device in hand….. how do I start gathering memory statistics for the apps I use.

Firstly git clone the b2g repo >

git clone https://github.com/mozilla-b2g/B2G

cd B2G /tools

python get_about_memory.py –help

There are several options you might want to consider with the tool, for now you can just run without any options.

Next Open Firefox Aurora or Nightly, and type about:memory.  At the bottom of the page is a button for loading the memory report from File.

Select and open the file you just saved in B2G/tools/about-memory

Part I: A how to for performance testing on Firefox OS.

Performance testing (http://en.wikipedia.org/wiki/Software_performance_testing) for B2G is being done by many dogfooders / early users as we speak.  The beginnings of this effort are to begin a framework for gathering this data, as well as empower both power users and employees to really start to make sense of all this complex data !. 

Toolset: Mind, stop watch, performance profiler, about:memory tools 

Background: 

https://developer.mozilla.org/en-US/docs/Performance/Profiling_with_the_Built-in_Profiler

http://blog.mozilla.org/nnethercote/2012/11/14/memshrink-progress-week-73-74/

http://andreasgal.com/2010/10/13/compartments/ (an oldie but a goodie)

First Steps in performance measurements, were hashed out over the B2G work week most recently in San Francisco. The most immediate things we want to gather data on are the following.

• B2g Startup Time:

Measure / Record the startup time for the overall system. We want to know how long it takes to get to the unlock screen. 

1. Use case: As a user I do not want to wait for a long period of time before my phone will start. 

Requirements:
     – The phone is off
     – Start phone
    — wait for unlock screen, with animation.

•  B2G Startup Time for each App:

1.  Use case: As a user I want to be able to open apps one after another and have the operating system be able to quickly load each app, and switch to the homescreen.

Part II of this blog post will go over How To’s of memory gathering / dissecting. 

 

Tune into Part II: About Memory Gathering / Dissecting 

If you are Launching your startup HTML5 is the way to go ( imho )

As we progress through the webapps project there is an astounding see of change coming to HTML5 Apps near you.

If you are starting a new project, I would recommend that you think about some of the ideas I am laying out below.

Graphic Design:

Apps will have different characteristics than webapps.  You generally will not just want to layer css ontop of your page, and consider it mobile enabled and done.

You will have to design a completely different set of views for your mobile app than your desktop site, and possibly even your tablet as each platform will offer different capabilities that you’ll want to take advantage off.  Your layouts will accomodate different styles, and furthermore different functionality.

Redirects: Another thing to consider is that an App doesn’t allow navigation off of the page, so you should think about that and what it means for your app.  You wouldn’t want to redirect your users off of your domain.

Apps should be targeted / focused towards doing one thing well, and bubbling that information up to the user.

Popups: These are a web browser concept, that doesn’t translate into the App experience, so expect that popups won’t work.

So if you want a kick ass App experience,  plan on having completely different views for different device sizes.  The good news is everyone will be writing the same language, and basically using the same stylesheets !

There must be a ton of user UI tips, for writing kick ass mobile /tablet  apps, if you have any add it to the conversation

Infrastructure:

WebApps probably won’t make too many changes to the internal infrastructure of data centers.  But to the infrastructure / mechanisms you use for exposing your product to the world, those are / have changed.  The rest api revolution has been replaced by the Javascript Object revolution.  Exposing your services as part of a Javascript SDK is the new kid on the block, and it’s knocking down walls.

So far I haven’t seen too much change in this area, the cloud is pretty much a given at this point, at least for a startup.

Personnel:

The time for the Android / IoS developer is over, don’t do it, don’t fragment your infrastructure, your support, your ability to ship early and often.. It’s just not worth it.  If you are a startup and anyone is telling you to go native, i would seriously get second / third opinions.

Having said that using Phonegap or something of the like is a good idea to help bridge some of the functionality gaps.

If you want to take a look at how to  load client side / server site javascript in phonegap and all have it process you should take a look at github.com/mozilla/soup

Why:

The advantage of being small is that you can move quickly.  Try doing that with an android release , iphone release, and website release.. All with differing capabilities / code bases, and then dealing with android fragmentation.  It will make you slow, and slow is bad….

Learn More.

Github Webapps MochiTest Repo

Writing your First MochiTest:

Can be a daunting task, all this mozilla code, all these strange words, mochitest, mozilla-central, jumping into the code base.. All are daunting tasks, but hopefully this tutorial will walk you through the general process.

Learn by Example:

The example that I am going to show is the soon to be performed work on mozilla central to port the webapps extension into the mozilla code base.

The exciting part about this we get to test against code that hasn’t been written.  The codebase i want to test will eventually be on mozilla central , but is currently not.  I didn’t select the option of testing against a debug build, because we might actually want to test against patches to the codebase. (which is what we are doing)

Building Firefox:

https://developer.mozilla.org/En/Developer_Guide/Build_Instructions

MochiTest Basics:

https://developer.mozilla.org/en/Mochitest

Assuming you have built mozilla central and you are able to run MochiTests, then what next.Well firstly you will want to choose a location for your tests.

Now If you are working on webapps this directory might be: dom/tests/mochitest, as it looks as a lot of tests that are dom related lie there. For each project it will be different, and you may want to seek out different opinions as to where things might go.

You will then have to create a directory “webapps”, and edit Makefile.in and add your directory below

DIRS  += \
dom-level0 \
dom-level1-core \
dom-level2-core \
dom-level2-html \
webapps \
storageevent \
$(NULL)

Next step would be to create a Makefile.in inside your directory foobar.  Below is a github commit by Geo Mealer which establishes our base repository.

Github Webapps MochiTest Repo

Running your Dummy Test Case:

navigate to the top directory in your mozilla-central codebase

TEST_PATH=dom/tests/mochitest/webapps/test_dummy.html make -C obj-ff-dbg/ mochitest-plain

Level II on your path:

Next you will be forced to consider / raise several issues with your development team.

You will want the ability to pref off certain chrome elements.

• prefs.setBoolPref(“apps.testing.mode”, true);
• prefs.setCharPref(“apps.testing.allowmgmt”, “http://mochi.test:8088″);

Determining what those prefs are that accomplish the goal is a conversation that needs to happen with the development team.!

It’s best to hold off writing until you have a general agreement, about this piece.

•  Designing your test framework.

Mochitests are a simple html / js test framework, it uses simple expressions like “is”, “is_not”, “ok”.  So the real difficulty is in designing your test framework to get the results you need.  Sometimes you will need greater access to objects other than prefs, sometimes you will need access to specific data structures to retrieve and set information.

There are tons of examples in the mochitest framework to work with:

localStorage, geolocation, general all have good examples. It really depends on what you are attempting to do, and how much effort you want to expend.

• Webapps Test Run

The test fails because the code isn’t implemented.  This codebase is a part of the same repo, but in a branch entire suite btw TDD.

 

Web Apps Sikuli / Python testing getting off of the ground

Just thought I’d spend a few minutes and blog about the Web Apps Sikuli Project.

What we have gotten done so far

• Basic set of tests working in Windows / Mac
• A working documentation set
• We have a few people contributing so far

I would like to say thank you to a few people for their support so far:

• Mohamed Dabbagah our intern from last semester, who has been pitching in whenever he can
• BYK – https://github.com/BYK – Thanks for the python tips!!

Please remember to not forget to note the Sikuli Test Day will be January 20th 2012

If you have spare cycles, and are interested in the project, check us out on github

Pick up an issue, try to solve, ask questions, contact me.. whichever suites your fancy.

 

-David Clarke

twitter: onecyrenus

email:dclarke@mozilla.com