Aug 10

A couple people have asked me recently how we do our Grid7 podcast. We have a humble talk show we do with entrepreneurs and innovators bi-weekly over on Grid7.com and we just did our 24th episode. There may be more streamlined ways of doing things but these are the steps involved from my perspective:

Capture

“Garbage in, garbage out,” as they say. The idea is to capture the highest quality raw audio to start with. If the guest is local I try to conduct the interview in person because I think the face-to-face interaction is a better dynamic. I use the internal mic on my MacBook and record directly to a track in Garageband.

If the guest is remote and I have a good internet connection, I’ll use my skype account with skypeout and capture using a great piece of software called Audio Hijack Pro. It’s nice because you can isolate the inbound and outbound audio to separate tracks and equalize the volume levels later. It generates an mp3 on your desktop and is straight forward. If I have a crappy connection I can record calls on my Treo using an app I have called CallRec. This captures the conversation as a wav file stored on the SD card. With a 2GB SD card, storage is a non-issue.

Produce

I use Garageband to refine the raw audio, add an intro/outro to the track and do the final mixdown. Provided the call quality was good, there should be no need to apply a noise filter. I have heard that Audacity is a good open source audio editor that’s available though I have not used it personally. Once I have things sounding right, I export the track to iTunes, right-click on the track -> “get info” and adjust the details in the ID3 tag. I then right-click and convert it to MP3. Once it creates the MP3, right-click-> “show in finder,” grab that file and ftp it to our server.

Publish

Last step is to publish the audio track to our site. We use WordPress as a CMS for our website and it has an open source plugin called Podpress that makes it easy to serve a podcast. Provided you have the Podpress plugin installed and activated, you author a post as you would normally do for a text entry only you click the “Add Media” button under the textarea and tell Podpress the URL of your MP3. I like to add a paragraph or two on our guest explaining his/her background and the gist of the episode and also include a headshot. If you’re writing any type of extended entry, you want to author it in a text editor and then copy/paste it into the browser (I’ve had Firefox crash after authoring a long entry in the browser and it sucks). Podpress generates the proper RSS feed and even gives you a flash-based audio player that allows the visitor to listen directly from the browser. It handles stats and can syndicate your podcast via the iTunes Store.

Promote

You’ll probably want to list your podcast in the iTunes Store (and “store” may be a misnomer – it’s just a directory so you don’t have to charge $$ to be listed). There’s plenty of other podcast directories out there- google around. I added ours to Everyzing so that the audio itself is indexed and made searchable. Running your RSS through Feedburner allows you to get stats on the listeners that subscribe via RSS. Depending on the subject matter of each episode you can then submit them to various news sites as you go. We held the #1 slot all yesterday on news.ycombinator.com from an interview I just did with the Zenter founders.

Monetize

We have not actually tried to monetize our podcast yet. It currently serves more as a vessel of exposure for us and an in-roads to make connections and meet new people. There are various options for services that provide a simple way to splice in ads dynamically. I went though and researched a bunch at one point and found Kiptronic to be the most promising (plus it sounds like they now support video blogs as well if you’re into that). AdSense is always an option for the site itself. Feedburner lets you display ads in the RSS feed itself once you cross the 500 listener threshold. I experimented with Comission Junction but saw zero dollars ever come out of it. Amazon affiliate program was equally as dismal in terms of what it generated. We’re now in the Google PPA beta so that will be interesting to see how well it works. Short of having a program so popular that you can command a dedicated monthly sponsorship, a dynamically-inserted ad via a service like Kiptronic seems like the way to go.

Anyways, of the million ways of hosting a podcast, that’s how we do things with Grid7. If you have a podcast of your own, what tools do you use?

Aug 05

Big ups to Derek Neighbors for pulling together the first ever Phoenix Dev House last night and hosting it at his place. It was a 12-hr hack session modeled after the first one conducted in SF. The idea is to get a bunch of coders together in a room and kick around ideas and then act on them, cranking out some code and getting something tangible finished by the end of the evening. At the first Super Happy Dev House they ended up writing the wiki system that was used to collaborate on future events (apparently Derek used this as well for the Phx site).

There were some great demos- Brian Shaler showed off a couple projects he’s been working on including an offline tool for digg users which he plans to unveil shortly and his crappy graphs site. This is the guy who figured out the secret of getting massive diggs for his stuff – write flash-based visualization tools for digg users. Josh Knowles showed off an app he made for the iPhone that allows him to control a keynote presentation on his mac from his iPhone – very cool. He’s also working on an app for the iPhone that’s similar to the Feelrz idea I submitted to Cambrian House. If he pulls that one off it has the potential to become absolute craze like Twitter. David Koontz showed off something he’s doing with JRuby but honestly he lost me at “J” :-( Lorin Thwaits demo’d a pet project of his that makes it easy to find very large prime numbers. I still don’t fully understand the utility of this but apparently the problem is important enough that the EFF is giving away $100,000 to the person that finds the next big one. You’ll probably see Lorin on the Discovery channel next month… James Britt was able to get some insight that helped him solve a Google Gears problem. I showed off the Sentinel application I had written a few years ago to solve the issue for people who transact business internationally of maintaining compliance with the OFAC’s SDN list. There were other demos but I had to leave early. But not before demo’ing a bomb-ass guacamole recipe (see picture).

nerdConvention2007.jpg

If you’re a nerd in Phoenix you’ll want to do a couple things if you haven’t already:

  • Sign up for the next happening on September 15th at UAT. Ideally, sign up to give a talk and not just attend.
  • Join the Elev8AZ google group and help boost visibility of tech in AZ by keeping up with announcements and promo opportunities that can get local AZ people and companies noticed
  • This week begins the voting on SXSW panels. We’ve got handful of AZ people that have submitted talk proposals so be sure to vote for your local favorites.
  • If you haven’t yet been to a Refresh Phoenix event come out this Tuesday and meet other nerds doing cool stuff in AZ.
  • And look for another Barcamp in the early December time range. You can monitor the RSS on that page to stay updated.
  • These events generate real ideas, relationships and even businesses. Thanks again to Derek for putting it together. If you’re in Phoenix and complaining that there’s nothing happening in the tech scene, there’s no excuse not to be at one of these upcoming events.

    Jul 17

    I just upgraded my version of Parallels from 2.5 to the latest version 3. VMware had upped the ante when they added snapshots to VMware fusion on the last time around but Parallels matched their bet and then raised by adding multiple snapshots. If you haven’t used either product, snapshots is basically a way of preserving the state of a VM right from the virtualization console so you can easily revert back to it and undo any changes you’ve made without having to clone the entire disk image using an external tool. It also saves space since it just stores the delta from the previous version.

    I had misunderstood thinking that Parallels just did multiple serial snapshots but it actually can do multiple branching snapshots. It’s a bit like lead climbing and placing pitons on your way up the rock face so you limit your fall. VMware gives you one route up the face and more importantly, one piton which you can continuously unhook and place right behind you. Parallels gives you unlimited routes up the face with unlimited pitons so you can choose to fall back to any previous position on the climb or jump to any point on an entirely different route.

    You can also think of it like this: remember that movie Sliding Doors with Gwyneth Paltrow (hot) where there’s dual realities occurring simultaneously? In Parallels 3 you can branch a VM indefinitely and grow multiple realities from the same VM. The coherence features and offline file explorer stuff they’ve added in this release is nice, but multiple snapshots is the killer feature now that VMware will have to match if they hope to keep up with Parallels in the desktop virtualization space. We’re partners with both companies so frankly, having this kind of competition elevate the level of play is thrilling to see.

    Jul 11

    I had two different friends last week “lose the keys” to their own web site. One situation was a disgruntled admin who thought he had leverage and decided to make demands and the other was an accident where the developer left the country for a 2yr mission in an isolated city in South America without leaving behind the server/domain credentials nor a copy of the code. Both sites were static and I was able to use the free HTTrack tool to suck down a copy and mirror them on my server. We were able to reclaim the domain for one and on the other we’re hosting on the .net alternate while we wrestle control back from the rogue ex-employee. This scenario got me thinking though and I submitted an idea to Cambrian House which is now in the running in this week’s tournament. The idea simply is this:

    Automate everything I did for my friends last week. I’m starting to think that this situation occurs more frequently than expected and not everyone has a friend with a server and the knowledge to copy a site remotely, mirror it and pursue the domain reclamation process. There is an opportunity to make a online service that allows the victim to immediately snapshot a copy of his/her site, mirror it on an alternate URL and get assistance with the process of recovering the domain in the event that it’s been lost.

    My question is: does this idea seem viable? Market big enough? If it did exist how would you go about promoting awareness of the service to the people afflicted with this problem? Not that we have any extra cycles to pursue creating this (that’s why I submitted to CH). But here’s the official entry that’s in the tournament:

    SiteLiber8: regain control of your orphaned site

    Jun 28

    I’m planning to do a few short hops this summer to neighboring user groups in the states around Arizona. If you have a developer user group on the West Coast or in the Southwest and are looking for speakers for an upcoming meeting, get in touch with me. The talks I’m tentatively planning on doing:

  • Become a project assasin: project management and revision control with Trac and Subversion
  • Low/no-cost options for effective CRM with Google Docs, Highrise, SugarCRM and vTiger
  • What are virtual appliances and what do they mean to me?
  • Virtualization for developers: an indispensable power tool
  • Using Innovation Games ® to unearth valuable customer insights to innovate on your products (interactive)
  • Software tools that power a startup: a look at the various software tools we’re using inside JumpBox (Open Source and SaaS)
  • Stone Soup Seminar: an interactive exercise in crowdsourced problem solving
  • I need to line up at least two back-to-back meetings in a city to justify a trip so if you know a neighboring user group and can make an introduction or referral, that helps. Looking forward to making the rounds and seeing what everyone is working on these days!

    Jun 27

    I love finding new productivity gems like this that shave minutes off daily mundane tasks. My partner just introduced me to the concept of fast user switching to flip back and forth between users on my Mac and I can’t believe I didn’t know about this feature until now. Here’s the specific situation and why this technique is so great:

    Situation: Right now the JumpBoxes we release require manual testing (ie. we have not yet built up a set of automated tests we can run against them). So each time we release a new round of applications, we need to manually extract them to our desktop and fire them up under VMware or Parallels and make sure they work properly.

    Problem: The downloads themselves are relatively small (~130MB) but extracted, they consume 2.8GB of disk space. This in itself is not a problem as I only test one at a time but all our laptops run the Mirra backup client which archives every bit of data in our home directories to a backup server. We needed a way to exclude the JumpBox test applications from getting picked up by the Mirra. There may be a way to do it from within the Mirra client itself but even that scenario has issues as the tests are truly sandboxed in an environment where they can’t overwrite something important. We came up with the notion of creating a separate user and running them in that context. The only trouble with that is you lose your daily environment so things you rely upon like your IM, Skype, Music, browser prefs, etc. just aren’t there plus you don’t have access to the files in your home directory.

    Solution: Fast user switching allows you to flip instantly between users and continue running whatever processes you initiated under the other account in the background. It treats it almost like you have a KVM to two different computers yet the performance hit is negligible (ie. not like running another computer, for me it was only the extra RAM required by the JumpBox).

    Available RAM before user switching:
    pre-userSwitching.png
    Available RAM after user switching running 256MB RAM JumpBox in other user’s account:
    post-userSwitching.png

    So this scenario is the best of both worlds because you can sandbox your test environment under a test user, turn on the JumpBox, get the IP and then flip back to your normal environment to do all your testing.

    This is how you enable this capability under OSX:

    Open your System Prefs and choose Accounts
    Picture-1.png

    Make sure you’ve unlocked it to make changes then choose Login Options
    Picture-2.png

    Check the option to enable it
    Picture-3.png

    Now your username will appear in the upper-right of your screen and you can easily flip back and forth

    Picture-4.png

    Thank you Apple for making this stuff work the way it should.

    preload preload preload