Minimal viable Android app

[brussee]

• Search results
• Click to stream
• Record video
• Extract thumbnail
• Create My Channel
• Add to My Channel

Performance experiments:

• Startup time
• Cpu usage during idle, tunnels, download, streaming (HD)
• Search results response time
• REST API benchmark
• Multichain record creation cost
• Code coverage
• CProfiler 10 min. run
• Publish content 1 to 10 devices

[synctext]

• Coming 3 weeks: finish 3 thesis chapters...
• Then minimal viable product: VLC playback.
• Launch Beta
• Measure & experiment

[synctext]

Progress with app, now VLC integration is complete. Search results are fast and responsive. Plus: channel subscribe via phone pairing (NFC) is operational and tested.

Shown above are the sub-second search results from Tribler network (if you have a premium phone).

Great progress is the bundeling of VLC. Not as a hard to maintain library, but simply a bundled .APK within the overall Tribler.apk

[synctext]

Key experiment: content dissemination experiment with a several phones.

Graph: time on X-axes, Y-axes number of phones which received the content. Timepoint 0: file added at source.

• 1 phone has channel.
• 10 phones get subscribed to the same channel
• continuously add file to source phone (10+ swarms in seconds)
• see them appear at the various screens with a refresh

Todo next week:

• describe experiment above with content dissemination. 1 photo of the phones "experimental setup".
• Graph with fake data.
• Startup time of Tribler also briefly Youtube (same hardware clean slate). Include log snippits in thesis (code quote style). Real data + graph + text of experiment: what you did and results. Repeat 10-ish time or so for accuracy; on 2 phone models. Graph bar chart with certainty intervals (1 bar Nexus 5, 1 bar Nexus 6).
  Split: Java GUI, Core started (known variance).
• describe multichain experiment !

[synctext]

Please focus on .tex for existing Multichain experiment results.

[synctext]

Thesis ref: example from Nijmegen on how to frame and sell privacy

[informusica]

Reference on social media and the news: http://www.digitalnewsreport.org/

[informusica]

• placeholders for each new shadow chapter
• 10 times power on of devices => logs
• multichain graphs => 3 graphs
• migrate only those results that are concretely there
• make new design chapter in terms of functionality. No bullets but text!
• for chapters with unstructured content, reshuffle content WITHOUT REWRITING OR ADDING CONTENT.
• if all this is done, refer back to your draft scratch pad and migrate non-covered content to appropriate chapters

[brussee]

• placeholders for each new shadow chapter
• 10 times power on of devices => logs
• multichain graphs => 3 graphs
• migrate only those results that are concretely there
• make new design chapter in terms of functionality. No bullets but text
• for chapters with unstructured content, reshuffle content WITHOUT REWRITING OR ADDING CONTENT.
• if all this is done, refer back to your draft scratch pad and migrate non-covered content to appropriate chapters

[informusica]

@synctext some now-backlog items that may ultimately be expanded, but only if time allows, and not before halfway November:

• power on / power off timings: script is available to assess more than 10 runs
• power on / power off timings: currently using time.time() rather than device time. That would need some programming (parsing) to resolve though. Only touch if time and supervisor allow. :)

[informusica]

For this weekend: do not delete anything you typed and write blindly

• new design chapter: no bullets but text. Sync with very first post in this issue thread w.r.t. functionality.
• migrate non-covered content from draft.tex to appropriate chapters. Note that content in draft.tex may implicitly indicate your thoughts on relevant highlights of the work. Use that ultimately as a filter.
• migrate non-bullets from old thesis
• results: put figures in the *.tex. Explain rationale of test, used metrics, setup, results, again not in bullets but in text.
• fill in further details in chapters but not in bullets.
• for chapters with unstructured content, reshuffle content WITHOUT REWRITING OR ADDING CONTENT.

[synctext]

Thesis .PDF download link: https://github.com/brussee/thesis/raw/master/thesis.pdf

[synctext]

Final experiment:

• Finished video recording, time is 0
• Create torrent latency
• Add torrent to channel
• Arrived at phone 1 to 4

[synctext]

http://www.npr.org/sections/alltechconsidered/2016/11/14/500214959/internet-freedom-wanes-as-governments-target-messaging-social-apps

[synctext]

Title brainstorm: "attack-resilient media using phone-to-phone networking"

To Add: The implementation consists of 12,125 lines of code. Table X shows the 25 largest source file contributions for this thesis work. ... X lines (55 %) consists of unit testing code, Y lines (15%) is made up of Android specific libraries, and the remainder is..

279  android/TriblerApp/app/build.gradle  
166  android/TriblerApp/app/src/main/AndroidManifest.xml
141  android/TriblerApp/app/src/main/java/org/kamranzafar/jtar/Octal.java

Possible additional experiment: measure transfer time between various phones of 25MB clips or show power consumption during phone-to-phone exchange (max 1 day experiment).

[synctext]

Latest results of weekend work by hard working student. Scenario is manual transfer between phones + fast spreading if there is some networking:

Here you can see how fast a video spreads between a bunch of Internet-connected phones. This is just our initial run, without any optimizations. Likely potential is that it can reach a worldwide audience of millions within 1 minute, without any servers or video support infrastructure. Purely a phone-to-phone overlay network.

[synctext]

Official thesis .pdf file

[qstokkink]

This thesis has long since been completed.

[ProphetZarquon]

It looks like the Android version hasn't been updated in five years. Is an Android app no longer in active development? Android has more devices active than Windows 10, & recent Android versions constitute a significant percentage of active devices overall. Considering the sizable resources & connectivity available on many Android devices, I would expect development of an Android version to be a high priority for any software project hoping to improve global connectivity & freedom of communication.

Is there some Tribler package for Android that I'm not seeing, or has that branch been abandoned for some reason? If so, why?

[synctext]

True, Android app would reach a massive audience.

Sadly we are a non-profit project without the resources needed to do a good app currently. Our focus is on the desktop and build a thriving community there first.

[qstokkink]

Is there some Tribler package for Android that I'm not seeing,

Yes, we have expanded upon this PR and are still working on several apps. However these are for specialized use cases such as just the passport logic or just the market logic. Now that we have trimmed the fat off Tribler a bit, we can start looking at running the full Tribler stack on Android again.

[synctext]

All activity now moved here #4827.