Novacut and DMedia roadmap

Registered by Bilal Akhtar on 2011-03-04

Novacut and DMedia development plans in the O cycle.

DMedia may be integrated to the Unity dash, this is to be decided in Unity sessions, and upstream work for that will be decided here.

Dmedia: http://launchpad.net/dmedia
Novacut: http://launchpad.net/novacut

Blueprint information

Status:
Not started
Approver:
None
Priority:
Undefined
Drafter:
Jason Gerard DeRose
Direction:
Needs approval
Assignee:
Jason Gerard DeRose
Definition:
Discussion
Series goal:
Proposed for oneiric
Implementation:
Unknown
Milestone target:
None

Related branches

Sprints

Whiteboard

[action items from session]

 * We're going to schedule a session for Novacut player accessibility, TBA
* Martin offered to help port PySkein to Python2
* Stuart will help Jason grok U1 SSO, plan how best to use for Novacut accounts
* desktopcouch enhancements session scheduled for Tue May 10 at 10:00 UTC (12:00 Budapest):
    https://blueprints.launchpad.net/ubuntu/+spec/desktop-o-desktopcouch-enhancements
* It was suggested we look at Qt possibly over GTK (we're quite independent from toolkit at this point, pretty much just use WebKit)

[jderose]

This seems like a good BOF session, during which we can elaborate on any of this, get feedback, answer questions, hang out. Thanks for setting this up, Bilal!

Here's my roadmap brain dump:

* From a technology standpoint, the most important next step is finalizing the dmedia hashing protocol:

  http://bazaar.launchpad.net/~dmedia/filestore/trunk/view/head:/filestore.py#L22

* To transition to the final Skein-based hashing protocol, we either need to port PySkein to Python2, or port dmedia to Python3, the former being more likely at this point (because of dmedia dependencies). I have very little cPython experience, so some help here would be greatly appreciated:

  http://packages.python.org/pyskein/

* Also need to get PySkein sponsored into Debian/Ubuntu:

  http://revu.ubuntuwire.com/p/pyskein

* All the same, we do want to move dmedia and Novacut to Python3 as soon as possible. GObject Introspection means this isn't too far off, but I'm a bit doubtful it could happen in Oneiric (depends on PyGI upstream). Hopefully this can happen for 12.04.

* Once we transition to the final hashing protocol, we should be able to quickly finalize the dmedia version zero schema. This represents a long term commitment to the user's data, after which we will consider dmedia ready for production, and recommendable for general use. We may make backward incompatible changes to the schema, but will only do so with well tested migration functionality that can losslessly migrate from old to new schema (sort of like what bzr did leading up to the 2a format, hopefully with fewer stops along the way). The schema is in very good shape at this point, and I'm ready to pull the trigger as soon as the hashing protocol is finalized:

  http://bazaar.launchpad.net/~dmedia/dmedia/trunk/view/head:/dmedia/schema.py#L22

* There are some key improvements we really want to see in desktopcouch for Oneiric, and that will be our biggest focus at UDS:

  https://wiki.ubuntu.com/DesktopCouchWishList

* We're also keen on improving Unity's productivity are large screens (say 2560x1600):

  https://blueprints.launchpad.net/ubuntu/+spec/desktop-dx-o-unity-bigscreen-productivity

* We hope to launch the player sometime this summer. We're hoping for July at this point, but as always, it can be difficult to estimate (and I'm perhaps especially bad at it).

* Player will launch on Ubuntu first, other distros will likely follow shortly after.

* Bilal will take care of downstream distro packaging, both in Debian and Ubuntu for both dmedia and novacut.

* Launching on Ubuntu is perfect for us because it's both a way to open the valve slowly (so we're not opening to the whole world on day one), and way to limit ourselves to the audience we have reason to believe will be by far the most generous (based on Humble Indie Bundles).

* Our goal is to be able to launch with no money, specifically for the Novacut revenue cut to keep up with our bandwidth costs even in the first month. This is an excellent creative constraint because it forces us to think very carefully about how we can control bandwidth costs. Initially, streaming is out of the question. Downloading is much better because people tend to watch stuff more than once, and dmedia can transfer files directly to peers via several different mechanisms. Hopefully we aren't *forced* to launch with no money, but either way, being *able* to will put us way ahead of the game.

* Novacut Inc is now officially seeking angel investors. We're gonna be picky, and will not compromise on our commitment to artists, free culture, and free software. We're seeking investors who understand that just means we aren't willing to compromise on winning (big grin). If you're interested, let's chat. Find us at UDS, or email us at <email address hidden>

* After launching player on Ubuntu/freedesktop, we're heading straight for TVs, tablets, and phones. Hi, Linaro!

* We intend to do revenue sharing with distros and hardware partners that ship the player by default. Hello, Linaro commercial sponsors! It will be a small percentage (as the vast majority will go directly to the artists), but a big pie. We think this will be money well spent in terms of establishing Novacut as a ubiquitous and high quality platform. It gives our partners an incentive for ensuring that the artists are well presented, that an excellent user experience is delivered.

* Sometime this summer, serious work will start on the video editor. I know people have (understandably) gotten impatient about this, started to wonder when/if we're going to start delivering on this promised editor. But dmedia is the foundation that was needed first. You can't have distributed editing without moving the video and audio files around in a sane way. Novacut 0.1 will be a true, dyed-in-the-wool distributed video editor (the world's first, assuming no one else beats us to it). We'd rather start with a very basic editor that's fully distributed than have a more fully featured editor that basically isn't distributed (hehe).

* Tara has done a fantastic job networking with the breakout, disruptive artists in this space. We've gotten great feedback that has helped tune our initial editor priorities, given us a feel for what artists need most urgently. One thing that is highly encouraging is that every single artist we've talked to is excited most of all about Novacut being a distributed/collaborative editor. We haven't once had to explain what distributed editing is, why it's important, or convince artists they need it. Many of these artists are already doing distributed editing, despite being stuck with software that make this epically painful and not particularly effective.

(?)

Work Items