Qt | Podcasting + conferencing + Twitter

A handful of people with ties to the maemo.org community have been kicking around the idea of a new podcast.  I’m not going to go too deeply into the proposed format at this time but rather will present the technical wishes discussed so far and solicit input from the readers on how to address them.

Interested?  Read on!

Basic Premise

While I don’t want to get into deep details of the format yet, the essential theme is a small group located anywhere in the world coming together over the internet to discuss various technology topics.  Most sessions would be recorded and released as podcasts later, but there would be occasional live broadcasts with listener participation.  Of course, Maemo and MeeGo would be key subjects.

The style would be fairly loose, provocative and hopefully highly entertaining.  The panelists would represent a diverse collection of viewpoints.  At any given time the plan is to have three to four co-hosts on any given ‘cast, with the possibility of guests.  The principals could even number more than four with a rotating and/or “as you can” mode of participation.

Technical Needs

As the four of us batted this idea around recently over Twitter, some good ideas emerged but they led to some walls.  We found it difficult to identify one application that would serve all of our desires.  Since that’s the driver behind this appeal, I’ll cut to the chase.  Those needs/wants are:

  • Internet conferencing.  Video is not important, but audio is a must.  Anyone with an internet connection (and a microphone) should be able to co-host.
  • Session recording.  Given the expected audience, OGG audio format support should be there, as well as WMA.
  • Moderator support.  The session initiator would have ultimate control and could assign moderators on-the-fly, retaking control if need be.  There could also be a “preach” mode, where the moderator allows any of the co-hosts to speak uninterrupted for a time (probably using a talk button).  When one speaker relenquished “preach” mode, another could take it until the moderator released it.
  • Twitter integration.  For live ‘casts, the audience should be able to tweet in questions or comments.  Tweets would be oriented around the moderator at the time.  It would also be nice to see where tweets are coming from, regionally.
  • Support for OpenID logins when running live events.  Listener login should be as easy as possible.
  • Call-in capability.  In addition to tweets, the app should support the ability of listeners to call in and contribute audibly.
  • Audio delay.  It would be nice to buffer live ‘casts so that if any unwanted call got through screening, it could be skipped.  This would not apply to physical events of course.
  • Multi-platform.  It would be great if this was peer-to-peer and ran on any viable platform, desktop or mobile.
  • Free.  As in beer AND speech.

What we’ve looked at

Naturally we began by looking at and for candidate applications that met the requirements.  I started with web-based chat and conferencing applications.

Peer-to-peer chat clients like Xfire and Ventrilo contain some of the necessary functionality yet are not specifically designed for this purpose but rather for gaming.  Gizmo comes closer but still falls short.  The main thing missing is support for listener participation.

Those solutions that supported most of our desires, like Webex and LiveMeeting, are commercial and have a cost attached.  Given that the founders have no funds for this little venture, that’s a showstopper.  So naturally we checked out the open source arena.  After some digging I stumbled upon Google’s openmeetings and was surprised it had not turned up in my initial Google searches (ironic, eh?).  I haven’t started testing it yet but it certainly looks promising.  One of the most useful aspects is its extensibility; theoretically we can add what’s missing.

Blog Talk Radio looks useful as an outlet once podcasts are recorded, or even for live events.  I just have not yet explored it deeply enough to know how far it will go to supporting what we’re hoping to do… but that’s in the plans.

Where Qt comes in

Even if openmeetings and/or Blog Talk Radio can get us started, I’m still interested in creating a new application that better fits the bill.  My current knowledge makes VB.Net the best solution with which to develop, and I’ve started on a proof-of-concept application I call Radiator.  It’s registered at Twitter and to that end I started with the Twitter functionality.  I have a simple client working that has incorporated an unusual and potentially controversial feature: the ability to post more than 140 character content in serialized tweets.

But allowing co-hosts to write novels on Twitter is neither the goal nor focus.  The application needs to include all of the other features listed above, and I intend to add them as I can to my VB.Net app.

However, I’d rather code this up in Qt with the idea of clients on any supported platform.  A web runtime (WRT) approach makes the most sense unless I’m mistaken.  So I’m appealing to Qt developers out there to consider joining in.  This could even be a Google Summer of Code project I think.

Next?

Ultimately I’d like to see podcasting be more supported by the open source ecosystem.  By that I mean that entities like the Linux Foundation could provide or sponsor the necessary tools and infrastructure so we would be on equal footing with commercial equivalents.

 What do you all think?  Don’t be shy– comment!

8 responses to “Qt | Podcasting + conferencing + Twitter

  1. Ouch, don’t make serialized tweets – 1) it screws up the UX for other clients, and 2) you thought getting multiple SMS’es were annoying a few years back, right? 🙂

  2. A very interesting initiative.

    Definitely worth being pursued in GSoC…

    Keep up the good work 😉
    MeeGo on 2600Hertz.com

  3. It was brought up by Alexey Zakhlestin on the MeeGo-community email list that audio captured from a group session might be poor quality. The suggestion was for each participant to record their part individually and mix them afterward. That may well have to be a consideration. I will try to test using available solutions and see just how bad it might be.

  4. I think that the lack of support for the internal cam from the N900 is the big issue. First, is still not a good quality, second, is not copabile with Flash integration.

    This makes websites like tinychat, stikcam, ustream and talkshoe hard to work with.

    I would love to see they solve the camera issue.

  5. Update: after a brief bit of interest, this one seems to have cooled off. I’ll still play with the ideas as I can, but I really hope I’ve managed to inspire some really clever coders out there…

Leave a comment