As someone who lives outside the current range of effective ADSL I am pretty picky about what media files I download. I have bemoaned the growing trend towards posting video as a means of communicating things as diverse as information on new technologies to a record of someone’s birthday party. OK, (me as case in point) we can’t all be great writers but text is a great communicator. If video is beyond my daily download budget then sound is just possible so I am regularly tempted to download a podcast or two from sites such as IT Conversations and even Google. I see something billed as being about a subject or technology of interest with good reviews so I go for it – almost always to be disappointed.
Podcasts can be great (sometimes). They are great when the material being presented is/was intended to be delivered largely by voice. Someone describing past events (say) can make a good tale of them and tales are for telling and being listened to. They can work as print but work exceptionally well as an oral presentation. Listen to Steve Wozniak doing his “stand-up” routine at Gnomedex 4 – it makes compelling and entertaining listening as does Paul Graham on Great Hackers BUT (that’s a big but) I prefer to read Paul Graham’s essays on the web. Paul Graham works hard at his essays and a good essay communicates more that just the sum of the words – I takes you on a journey and fosters thought and mental analysis as it develops it’s theme.
I had to sit on my hands for a moment here until I had decided not to site a specific podcast as an example of what is terribly wrong with most. Still you would not have to range too widely through the catalogue to come up with some dreadful examples. Picking one would be unfair on the participants and would only invite comment on the merits of that individual choice.
One common podcast theme that almost always fails miserably is three guys talking about some topic – in the worst cases over a telephone link. This is, superficially at least, a parallel to the TV “chat show”. You do not expect even the most entertaining chat show session to elicit much information but the audience can enjoy the interaction between a well briefed show host and the “personality” flogging whatever wares they currently have for sale. The podcast not only excludes the audience from most of that interaction but telephone links also mean that the participants are missing it too. Add in a generally feeble level of host preparation (or prior knowledge of the supposed podcast subject) and you are left with a subset of the words – and pretty sparse on the information front they tend to be. If people find it tough to get over their points in a text – most find it even tougher in conversation. Sure, a conversation is easier (less nerve racking) than a presentation but it is still a terribly limited form of communication for all but the most expert.
The last paragraph might give you the impression that I would advocate video as an improvement to sound only media but I fear that would not be so. Creating professional level video is even more expensive and demanding than sound and (in the mainstream) relies upon well trained professional presenters. I am aware that sites like YouTube have given a “voice” to many new talents – some of them may mature and endure but most are only really effective in short sessions where we can make allowances for production deficiencies.
Why am I worried? – I fear that podcast/video conversations may take over as the dominant form of technology communication on the web. Video is not searchable or effectively indexable and that is a major problem in it’s own right. The bigger problem is the loss of content a move to these new media would result in. I was looking at an excellent blog entry from Anthony Moore on the BCL team blog at MSDN. His subject is a new DateTime type in the pipeline for .NET 3.5. It is detailed, well argued and informative – cracking documentation. Let us hope it will not get lost in the myriad of MSDN pages. But here is the point. Provided the content is not deleted it will always be findable by search engines. The other side of the point is that the best job a podcast could have done with this subject matter would have been to make you aware of the deficiencies in the current .NET DateTime type and aware that a new type was being added to the framework. Such a podcast would soon be lost and un-findable and would not (indeed could not) include the content to be found in the text.
An MP3 download of this rant will not be available