Screenflow: Unpredictability of Bad UI and a Wasted Afternoon

frustrationScreenflow is one of those essential tools for people like me who fancy Macs and need to create screencasts.     Matthew McCullough introduced me to this tool years ago, and I haven’t looked back since recording the equivalent of many man-months of video for various clients.   It works well, when it works that it, but it also suffers from some instabilities.   Previous versions used to just decide to randomly exit and crash, and I’ve had my fair share of “Unable to open video” errors.

If you do any video work at all, and if you record your own soundtracks, you’ll understand how long it takes to record a good video and to make sure that everything is aligned to make that video work.  When you do screencasts and voiceovers you run through a checklist:

  • Is my voice rested enough to do a good voice track?
  • Have I had too much caffeine and am I going to sound like Jim Cramer?
  • Are there people in the office who are going to look in your office window and see just how strange you look when you are doing a screencast?  (I move quite a bit, it’s a strange process.)
  • Is your crazy expensive microphone plugged into the Tube Amp and the iMic just right?
  • are the levels set correctly?  Also, what does that even mean? you are not an audio engineer.
  • Have you unplugged your fridge and everything else that can add electronic noise to the audio line?
  • Have you made sure that your screen recording isn’t going to be interrupted by someone’s curse laden Skype message box?
  • Will the guy from the back of the office knock on your door to ask you another question about Apache httpd in the middle of a voiceover recording session?
  • and, most important for me, is the conference room next door free of that loud-talking narcissistic idiot who likes to make too much noise for your recording sessions?

Recording 10 minutes of audio?  Maybe it will take you 15 minutes, but it’ll more likely take you an hour (or two).  I’m never surprised that it takes 20 takes to get it right.  Now, the last thing in the world you need to have happen after the moon and stars have aligned is have Screenflow die on you as it decides it doesn’t want to be a piece of software.  I use this tool frequently, and I’m constantly cursing it because it just disappears.  Maybe, if I’m lucky, it shows me the rotating rainbow wheel of “I’m about to just crash on you”, but, more often than not, the thing just, blip, and disappears – no record of the previous project or, worse yet, a corrupted recording.

It’s enough to make me stop what I’m doing and write a blog rant about the tool… Although, it isn’t like I’m moving to anything else, Screenflow is the best tool for the job (even though it has caused me to scream as loudly as possible twice in the last two weeks…. and if you know me, that’s pretty loud – I’m a singer.)

GitHub, Deploy Hooks Could Use Some Love

GitHub offers integration with a billion zillion services, and they just present them all in one big list.  Here’s the problem. Who knows what all these services do?  Notifo, Pushalot, Grove, DeployHQ, Depending, Jaconda, Irker, Planbox, Pipplydoodah, Circleci, JimmyBox, and one hundred more… I’d love to know what these services do, I would, but I feel like this list of options is absurd.

Here’s the combined screenshot of the deploy hook screen, if it were an artwork I would name it, “Do we really need another damn issue tracker?”  After this massive screenshot, some suggestions for GitHub…


You see how that doesn’t scale well?  It’s the phonebook effect and it gives every open source project or company in the space an incentive to name themselves something that starts with “A”.

  1. Sorting alphabetically by default, can we have other options? – Imagine if Google Play gave you a list of applications starting at A?
  2. Provide some categories for deploy hooks – I’d like to see all the issue trackers or all the notification services.   This would help make sense of the list.
  3. Give me some statistics about usage – Look at how the Drupal project displays statistics about adoption with graphs over time, do that.
  4. Provide users with the ability to comment on deploy hooks – If I’m contemplating using one of these hooks, it would help to know if “the crowd” had an issue with it.
  5. Make an editorial choice and feature some hooks over others – Why not?  Maybe you could charge these companies money to say good things about them?

If you made this page more of a showcase of the deploy hooks that people are actually using, if you exposed statistics about adoption, and if you made this page friendlier for novices you’d get even more signups per day than you are getting now.

Big Data Evaporates into Nothing. News at 10.

big-data-headlineOk, “Big Data” is going the way of “Information Superhighway” and “Web 2.0”.  How do I know?  I know this because I’ve seen a number of posts by people both attacking and defending it as a “thing” that could be considered real.  There’s a piece in the NYT attacking it, a bunch of people are now talking about “Data Skepticism.”  Don’t mistake the meme for reality.  Big Data is just a name invented to refer to nothing, and the marketing department has ruined it.

Here’s some short fictional dialogue that illustrates the way in which Big Data evaporates during the budget process at a large corporation:

CIO: When you say Big Data you mean Hadoop, right?  I was looking at the budget and I don’t see Big Data anywhere, but I do see this huge line item for Hadoop.  Jack wants a promotion, and he tells me everything you say BTW, he said you called me a “greyhair that doesn’t get Big Data”… I get it: Big Data is Hadoop.

Manager: Big Data isn’t just Hadoop, it’s a whole concept. It’s bigger than just a single technology.  Big Data is a new approach to data at scale, and I’m sorry about what Jack told you I thought making fun of the boss was essential mid-level management team building.  I apologize, I said that during the offsite meeting I thought he’d keep that to himself.

CIO: Forget about it, Jack’s not getting the promotion because I think he’s an idiot.   Back to this Big Data thing… ok, so it’s analytics, but we’ve had analytics from quite some time.   Is this just another name for Business Intelligence.

Manager: Dear God. Don’t call it that! You’ll scare away all the kids we just hired.  We need them. They are the only ones who understand Map/Reduce.  Big Data can also cover analytics, yes, but it is a different thing.

CIO: Wait, so the idea that businesses collect data that needs to be processed into reports that then drive the decision making process, that’s a new thing?  I thought we were…

Manager: …It isn’t that it is new it’s different.  Big Data is a different way of thinking about old problems in a new light.

CIO: Ok, so we’re talking about a analytics at a different scale, right?

Manager: Right, and different tools. Big data is more than just working with data at scale it is about asking different questions. It is about realizing the the value is the data and making sure you are taking advantage of it.

CIO: Paul, if you keep on talking like that, maybe Jack will get the promotion. You realize that you just said nothing to me. You said it very eloquently, but if I wanted bullshit I’d go read eWeek. Back to the budget… so Big Data is stuff like Hadoop and Map/Reduce?

Manager: No, stop it. That’s too limited. Big Data is transformational, it covers not only offline reporting but any systems that need to operate at internet scale: Google is Big Data, Facebook is Big Data, but the government is also starting to adopt Big Data.

CIO: Internet scale… our technology budget is a lot less than Google or Facebook, and we’re not indexing the web.   On another note, this Big Data thing sounds like a religion. Are there churches?

Manager: Why aren’t you taking this seriously?  There’s a lot of money being spent on Big Data.  I’m not the only one saying this stuff.

CIO: There’s a lot of money being spent on religion as well and there are many people that believe the world sits atop an infinite stack of turtles, that doesn’t mean I should take it seriously.  I’m asking because you are proposing that we spend a lot of money on it.

Manager: You know what, Big Data is Big Data you’ll know it when you see it.  Haven’t you ever worked with a Data Scientist.

CIO: Hold on, a Data Scientist, you mean a Statistician? Years ago, we hired people that knew SAS – that software has been around since the 1970s. Did you put money in the budget for a Data Scientist?   And, if so, do I need to install Bunsen burners and fume hoods in that office?

Manager: No. Maybe you should come with me to this conference. A Data Scientist is someone who knows how to work with Big Data.

CIO: You know what? This is fascinating, maybe we should schedule a meeting to talk more about this, but I’ve got to go talk to our analytics guy about the report he just sent me. We’re having a problem with the budget because we spent all this money on machines to support this Hadoop initiative, and it doesn’t look like we’ve found something to use them for yet.

Manager: Right, I wanted to talk to you about that.  Our Hadoop initiative isn’t paying off just yet because we need more data to analyze.

CIO: Thanks Paul. A quick tip from upper-management for future reference.  When I end a conversation telling you that “we should schedule a meeting to talk more”, that’s just a creative way of saying, “You are full of it, and I’m walking away now.”

GigaOm’s SQL on Hadoop Report…

GigaOM released a report about “SQL-on-Hadoop” which talks about some of the trends I discussed at Strata last month.  They have a completely different perspective on the problem, I think one that is mostly informed by vendors in the topic area, namely Cloudera and the work being done on Impala. They do mention a few Apache projects, and some of those “risky startups”.

Was it sponsored by Cloudera?   Not sure, but if you do an exact search for the phrase “MapReduce does not gracefully handle many concurrent requests”, you’ll get a link to the PDF on the Cloudera site.   Clearly, they liked what they read enough to make a paid report freely available on the website.   Seems to suggest some relationship, no?   …there is nothing wrong with this, a very common practice for content marketing is to approach an analyst firm and sponsor content (and then ask them to call the competition “risky startups”).   Was it sponsored?  I don’t know.

Also, what I love about analyst reports are statements that take no risk and make no predictions.  Here’s an example:

“Their technology bets are risky but, if obviously superior, could allow them to stage a market coup.”

That’s as bold as saying, “Some companies are proposing innovative solutions which, if successful, could change everything.”  How does this stuff get approved by an editor?

That being said, maybe they are right? Those startups are risky: Drawn-to-Scale ist kaput.