Screencasts: Watch Them, Love Them and What's Wrong With Videos

Wed, January 30, 2008, 02:52 PM under Random
Why is it that 50% of the people watch the screencasts and are asking for more and the other half don't (see links from Q8 and the comments)? In my view, everybody should watch screencasts (to be clear, I mean as a complementary activity to reading articles). If you haven't ever watched one, there are some links here (plus check out the UK MSDN page).

"What is so great about screencasts anyway?"
Many things:
a) Following a document step-by-step to achieve a goal is fine, but what about when you don't have the right stuff installed? It is many times that I have received emails from people saying words to the effect: "Having watched your video I am now downloading XYZ to give it a try!". Those people would not have gotten the same enthusiastic experience from reading a document. E.g. when Vista wasn't so widely available, this Sidebar Gadgets video could not have had the same effect in written form.

b) Even if you have the right stuff installed, what is it that would motivate you to invest time in following some steps to achieve a goal, i.e. how do you know it is going to be worth it? Instead you watch a short video and it is easy to decide if you want to go ahead and explore this technology further or not, e.g. how many desktop developers own a Windows Mobile phone and are aware of the dedicated managed API (not talking about the Compact Framework here)?

c) Let's face it, a picture is worth 1000 words which makes a video worth numberOfFrames*1000 words. I explained the LINQ-to-objects relationship with the language features on this blog and received good feedback. I described it in a video and I received outstanding feedback. How would you capture the effect of my LINQ video in written form?

d) It is not just about the topic being demonstrated, there are side learnings. Whenever I watch someone else do a screencast I pick up productivity/usability tips (e.g. some new ones in VS2008) about the way the presenter uses the tools or talks about certain concepts that is different to my mileage. You just don't get that in a written piece of text. Screencasts bring a demo at your desk, which usually you pay money for to see at a conference!

e) I was going to add a few more reasons here but they are weaker and I don't want to dilute the main points above. Feel free to add your own reasons why you like or hate screencasts in the comments below.

Now, let me tell you why I find videos sometimes lacking: Non-scanability. In the age of information overload, I usually scan pieces of text for the interesting points and probably abandon 80% of the web pages that I randomly visit within seconds. You cannot do that with videos today – once you've started watching it is hard to scan through without missing huge chunks of content. This is the reason that I seldom even attempt to watch videos over 30' unless I know for sure that I am interested in the topic in a big way! Table of contents with timings help slightly but everything I have seen that tried to address the problem is inadequate. My attempt to make it less painful is to keep screencasts at 15' long because then you are investing a reasonable amount of time and you can probably tell if it is worth to continue to watch after 3'. To that extent, I try to have a quick intro telling you what's coming. I also offer the additional download option which means that clicking forward/backwards to scan will be very fast as opposed to the streaming media (craze that is taking over the world) that largely relies on sequential viewing...

The larger question of how do we make long videos scanable remains. I think there is an opportunity here for some clever software: I point this piece of software to a wmv file and it instantly produces a complete transcript from the audio and it highlights every word as the video continues to play. No matter what word I click on in the transcript, the video follows suit by synchronising itself. Now I can watch and listen to the video and concurrently quickly scan for words/phrases of interest so that I can jump directly to that piece with a single click. Anyone building such software or does it even exist already? I can be a beta tester ;-)
Wednesday, January 30, 2008 3:24:00 PM (Pacific Standard Time, UTC-08:00)
I'm one of the 50% that said they don't watch the screencasts (I don't mind publicly admitting that). The reason I don't is because there's generally too much guff surrounding the substance that may be in the video. As soon as the screencaster goes off-topic, I tune out.

I think that for screencasts to work as a learning tool they have to be extremely to the point with no extraneous material (no digressions, no attempts to be funny, etc.). Almost clinical in content, I guess. That's not to say the screencaster should purposely throw a cloak over their personality. There's just a time and place.

Now if the screencasts where indexed in the way you described, there is scope for yet another piece of clever software that can detect when the presenter goes off-topic and skip that bit.
Wednesday, January 30, 2008 7:33:00 PM (Pacific Standard Time, UTC-08:00)
I think there is an opportunity here for some clever software: I point this piece of software to a wmv file and it instantly produces a complete transcript from the audio and it highlights every word as the video continues to play.

Would that not be the Expression Media Encoder? :) I think Channel 9 has some videos showing how quick it is to build a player with bookmarks in that!
Wednesday, January 30, 2008 10:48:00 PM (Pacific Standard Time, UTC-08:00)
Another great thing about a screencast compared to a blog posting: I can watch two or three screencasts while doing my 10 km run on the treadmill at home. I have not yet had any success reading neither books nor blogs while running ;-)

Wednesday, January 30, 2008 11:10:00 PM (Pacific Standard Time, UTC-08:00)
Hi Daniel,

I think the idea of a scanable video is a great one. The other issue for me, though, is time. When I actually have time to watch these things is when I'm traveling.

As anyone who's tried watching a code based video on a Zune / iPod knows, it's virtually impossible.

It would be interesting to try and produce a video with the constraints of those devices in mind.

Thursday, January 31, 2008 3:15:00 AM (Pacific Standard Time, UTC-08:00)
Hi Daniel,
I find screencasts very useful as a learning tool specially when I know little about the topic. The amount of knowledge that I gain from a 15' video is a lot more than what I learn after reading a blog for 15'.
What I like about the way you make your screencasts is that what you do is easily reproducible.
For a more in depth topics however, I'd rather read blogs.
Thursday, January 31, 2008 5:45:00 AM (Pacific Standard Time, UTC-08:00)
It depends entirely on the presenter for the screencast to be good. I do like your screencasts Daniel, especially Visual Studio 2008 screencasts :)
Thursday, January 31, 2008 2:17:00 PM (Pacific Standard Time, UTC-08:00)
What amazes me the most about this post is that someone I dont know well actually linked to my blog :).

Its the little things in life that make me happy LOL.
Thursday, January 31, 2008 2:24:23 PM (Pacific Standard Time, UTC-08:00)
Neil: I agree with the assessment of being almost clinical for 15' – I like the way you put that. I would be very interested to know (publically or privately makes no difference ;)) if you observed any of the "guff" in any of my videos. Always looking for feedback and I am sure you'd have ideas that would contribute to producing leaner meaner screencasts.

Nidonocu: Not AFAIK. Having sporadic bookmarks puts the producer in control. The requirement I described puts the consumer in control (to scan forwards/backwards as they see fit from the auto generated complete transcript). Makes sense or am I missing some super duper feature?

Kasper: I've never heard that one before – cool :). Your comment makes me think whether I should start a podcast series but that's a separate topic...

John: You are hitting a sore issue. I have asked about encoding for portable formats and the consensus was that the demand was not high enough for us to invest the time. I'll revisit this if there is more demand/proof that I can take back to the team.

Payman: Interesting, thanks for sharing. I do tend to agree which is why I do many introductory videos. Whenever I tried to do something a bit deeper it ended up beyond the 15' mark which kind of validates what you said.

Chaks: Thanks, you'll be pleased to know I have a couple more VS2008 in the drafting stages :)

Dmitry: It has nothing to do with "knowing" people :). You had a nice description and also 3 links that I would have included anyway, so you saved me the trouble by enabling me to just link to one – yours!
Friday, February 1, 2008 7:46:00 PM (Pacific Standard Time, UTC-08:00)
Not AFAIK. Having sporadic bookmarks puts the producer in control. The requirement I described puts the consumer in control (to scan forwards/backwards as they see fit from the auto generated complete transcript). Makes sense or am I missing some super duper feature?

Currently, the best solution I can think of is typing in the transcript. Placing reference markers at the start of a sentence or paragraph and then data binding a scrollbar to the 'needle' control so as you scrolled through the text the video kept roughly in sync. Unless you have software which can listen to the spoken text, compare it to the written stuff and work out the rest. :)
I do know that scrolling a streamed video is possible at least. I've dragged the needle on C9 Silverlight based videos before and had it rebuffer and resume playing from the new position. Flash it rarely seems to work on. ;)
Saturday, February 2, 2008 9:36:08 AM (Pacific Standard Time, UTC-08:00)
Nidonocu: Of course, you can advance the streaming video (e.g. with Silverlight) but it is too slow while it re-buffers which misses the whole point of scanability (which by definition is about speed).

And, yes, you can type out the script and manually place markers, but I am not looking for a manual-combo-hack-solution; I am looking for an integrated, automated solution. The disconnected items of functionality that you describe prove that some smart software could combine the whole thing into the desired experience. So when you've implemented that let me know :-D
Wednesday, February 13, 2008 1:08:00 AM (Pacific Standard Time, UTC-08:00)
Hello everyone

Here is my "manual-combo-hack" solution to this problem.

Daniel, would you like to beta test my application? :-)

Tuesday, February 26, 2008 1:13:27 PM (Pacific Standard Time, UTC-08:00)
aravind: You haven't actually linked to a specific blog post so I am not sure which one to look at. Either way, if you really have a solution for the "spec" I posted - yes please email me (see top left) so I can test drive it :-)
Comments are closed.