TechFold - Bold tech & web commentary

Why I Don’t Like Video - My Brain is a Mini-Google, Yours is Too

I finally figured out why I don’t like any serious application on online video, vlogging, etc., and why I think I’ve only watched one of Scoble’s video interviews since he went to the video side from MS. I always found it a bit recidivistic - to want to emulate TV on the net, but never took the time to figure out why - until now.

Its one part learning theory, one part reading mechanics, and one part personal preference. What prompted this thinking was the Gabe Riviera (TechMeme guy) Interview by 1938 Media. I’m sure its full of interesting, useful insights - but what has prevented me from watching it is the fact that I’m busy like crazy, and video is just unparseable.

Here’s the mini-Google theory: A text blog post I can skim and pick out the parts that are interesting to me for a more detailed read - essentially my brain functions as a mini-Google, crawling the post, getting a superficial understanding of its contents, assigning a PageRank to each paragraph and subsection, identifying relationships between them, and flagging certain parts (or the whole thing) for further review.

With video and voice, this is impossible: there’s no way to crawl at high speed, get the gestalt, and the dive deeper - its an all or nothing, real-time proposition that to me makes sponging and organizaing massive amounts of information (the joy of the internet!) slow and cumbersome. So I don’t bother.

At the end of the day, I’m guessing I might be at the extreme end of the spectrum of skimmer vs. detailed reader, but for better or for worse, that’s how I am. What about you?

Tagging and ratings might add an element of filtering, but before video is any use to me, I really want a transcription-based indexing search engine that break up a 20 minute PodTech video into Q&A subsections indexed by keyword, viewable individually. Does this exist already? I’m not sure - I seem to recall some buzz around video markup and timeline chopping sites a while back - a topic for further research.

In the meantime, the secrets of Gabe’s success will remain unavailable to me, locked in the ultra-low-bandwidth, forced agenda, passive medium of video.

1938media, blogging, google, learning, reading, sponging, techmeme, video youtube

Share This

If you enjoyed this post, make sure you subscribe to my RSS feed!

6 Suggestions for SuTree: Vertical Video Lessons

SuTree’s motto is “watch & learn” - and the site is a great place to do so. SuTree indexes video lessons from across the web: lessons are submitted to the site by members and editors, and can be rated, commented on, searched, and categorized. The site is clean, and the categories make it easy to drill down to the content you’re looking for. SuTree launched in the beginning of April

The “Submit” process for getting a video onto SuTree allows for thumbnails to be uploaded, tags to be added, and other descriptive information appended. Submissions are added into a Category > Sub-Category > Sub-Category > Sub-Category structure for very granular categorization.

While SuTree has the fundamentals down, its newness shows. In addition to missing a lot of social features (which I assume are in the pipeline), here’s 6 other suggestions for SuTree to move forward:

1. Improve your Search: Search appears to be a literal text string search. i.e.: There’s a video on SuTree called How to Pitch a Pop-Up Tent - searching for “pitch a tent,” however, returns nothing. That does the SuTree index a real disservice. Given the complexity of good search, if I were SuTree, I’d be tempted to just add a search widget and let Google deal with it.
2. Open Up the Taxonomy: Either enable users to suggest categories, or add a full tagging implementation (see next point).
3. Complete your Tagging Implementation: Submitting a link lets you add tags, but from that point on, they aren’t visible. You can’t search on them, there’s no tag cloud, tags aren’t listed for videos… they basically just disappear into the void. Tagging adds real value in a context like this, supplementing/complementing the rigid category taxonomy.
4. Streamline & Automate the Submission Process: SuTree would likely get a lot more submissions if doing so didn’t require entering a bunch of meta-data that can be grabbed programmatically. If SuTree could build a submission process that let a user point a crawler to a video page and walk away, that would be better. Given the diversity of sources on SuTree, however, such an implementation would be difficult; perhaps a teired submission process would be an answer. If the submission is from YouTube, users just need to post the YouTube link. If its from elsewhere, more data is required.
5. Clarify your Submission Practices: When you add video, it appears to go into a queue to be approved by a SuTree editor. Once approved, it goes public. I couldn’t find any verbiage on the site describing this process. Perhaps it does so later in the submission process. Either way, this would be nice to know for both consumers (that SuTree is handpicked, quality video) and submitters (don’t bother submitting crap).
6. Clean Up your Vocabulary: On the same page, SuTree will ask you to both Add and Upload video. Given that you can’t actually upload anything to SuTree, this just ends up being confusing. Pick a single word and use it consistently. “Add” or “Submit” would be good. Given that editors appear to need to approve video before it goes public, “Submit” would probably better as you can’t “add” until approved anyway.

All of that being said, I think SuTree serves a valid need, and does so competently - which is a compliment given that the site launched in early April. I’m looking forward to seeing how SuTree does once its implementation is a little more complete, and more social features in place.

More on SuTree from around the blogosphere:

1. Dumblittleman.com and Amit Agarwal put together great lists of other video-lesson providers.
2. Digital Alchemy compares SuTree to del.icio.us.
3. BizDevBlog points out that SuTree is missing some key social user-promotion elements - a “My Lessons” widget and API for starters.
4. Blonde 2.0 points out that SuTree hasn’t yet connect users in anyway - so the social element is still more or less MIA.

5min, lessons, sutree, tutorials, video youtube

Share This

If you enjoyed this post, make sure you subscribe to my RSS feed!

BlinkX, Video Search, and the Corporate Shell Game

The Financial Times broke the news, but Mashable summed it up more succinctly: $20 and a free pony if you can figure this one out.

The demerger of the consumer business is a complex transaction, in which Autonomy will first take ownership of Blinkx, a separate company founded by Autonomy’s former US chief technology officer, Suranga Chandratillake, which already uses Autonomy’s consumer search technology. In exchange Blinkx will be given exclusive rights to the technology, everywhere outside China. Then the Blinkx business will be demerged again and floated. [From the FT article]

Sounds like an complicated way for Suranga Chandratillake to take a pay day while keeping shareholders happy. Or something.

Meanwhile, try using Blinkx. Personally, I found the search results from YouTube alone a lot better: less clutter in the results, more relevance.

BlinkX for “Nissan Skyline.”
YouTube for “Nissan Skyline.”

BlinkX results are full of a bunch of remote-controlled car stuff - not what I was looking for. Plus, the BlinkX interface, with preview videos playing unsolicited, video thumbnails, and so on, is distracting to the point of making the site painful to use - however technically impressive it may be. Say what you will about the aesthetics/usability of YouTube, but its better than BlinkX.

Finally, BlinkX doesn’t seem to be indexing YouTube properly either. The result sets from each don’t match, and I have no idea how BlinkX ranks video results from one site over another anyway. When it comes right down to it, BlinkX seems counter-intuitive, dis-organized, and a confusing way to find video. This is the fate I fear for CastTV.

blinkx, casttv, mashable, mergers, video youtube

Share This

If you enjoyed this post, make sure you subscribe to my RSS feed!

CastTV - is there room between verticals and gootube?

I got into a little back and forth on Mike’s post on TechCrunch about CastTV (see the original TechCrunch CastTV post too). CastTV is a video search engine - indexing video from top video sites (YouTube, etc, etc), as well as across the Internet.

CastTV brings some unique twists to the table with their indexing system - it accumulates index data about a video across the internet, for instance - i.e.: if the video is posted in 15 different places, CastTV will treat the video as a single, indexed entity, with fifteen logical locations, as opposed to 15 different entities with different keywords, etc. According to TechCrunch, CastTV also allows for comparison shopping between download sites.

Anyway - my basic conjecture is that CastTV is another Riya - a cool technology in search of a business in a hype-heavy segment. Riya went from a buzz superstar facial recognition technology powerhouse to a marginal “visual search” tool (at Like.com) that let’s you search the internet for handbags you might like or rugs with nice patterns.

My conjecture is based on the fact that between YouTube/GVid and iTunes, most video on the web is well indexed, and keyword, tag, and genre searches, as well as social recommendations, meet most everyone’s search needs. Perhaps if the user-submitted video market becomes more fragmented over time (with an ascendant MySpace video or Photobucket), there may be a case for CastTV. Similarly, if CastTV wanted to delve into the grey market waters of torrents and P2P, it might have more appeal. But as a largely meta-search engine for the top video sites, I question why I would go to CastTV instead of the site where I know CastTV is going to return its listings from anyway.

My other thought is that the specialized searches for which CastTV will have the most appeal will be better executed by vertical sites such as SuTree.com that specialize in a particular video niche.

So - for mainstream searches, I’m seeing CastTV as redundant, and for specialized searches, I see it outgunned by dedicated vertical sites. That being said, I’m very willing to be proven wrong, and happily requested a preview invitation. The other suggestion I made in the TechCrunch comment thread was that CastTV deploy its unique indexing technology elsewhere - using the cross-internet-meta-index for keywords and phrases would at the very least provide a hype alternative to Powerset for the google-pundits.

Note: Here’s what I mean about being proven wrong. Maybe Like.com and the Riya team are finding their footing?

bittorrent, casttv, google, like, p2p, powerset, riya, techcrunch, video, visualsearch youtube

Share This

If you enjoyed this post, make sure you subscribe to my RSS feed!