I call bullshit. These are unprovable assertions based on purely imaginative history. The web has proven that cooperation and standards can emerge from billions of diverse voices. If anything TCP/IP probably had as much to do with where we’re at today as anything. Computers were just expensive word processors until the Web emerged. Most of the utility of modern day computing lies in the Web which Gates completely didn’t see and tried to hold back.
Bill Gates was smart and exploited every opportunity that he could. However, it’s just as easy to make the argument that the industry is still trying to recover from they way he ruthlessly exploited Microsoft’s scale. It’s fine to recognize Mr. Gates’ actual achievements, but spare us the grandiose drivel.
Mark Schmitt paints a big tent two-party picture where issue oriented constituencies are able to effectively organize and compete by virtue of Clay Shirky’s lowered transactions costs.
The use of constituencies rather than parties is telling. The American political system’s implementation of Montesqueian separation of powers provides a durable, but extremely rigid political system by pitting institutions against each other (gridlock). The US system also uses a single member district plurality / winner-takes-all voting system that promotes two-party entrenchment. This balanced, if unwieldy, system tends to constrain American political thought within a well defined sphere of ideals or political realities (one of the reasons for the dearth of great political thought from within one of the modern world’s earliest democracies). What’s not clear is if Mark’s take is a sin of ommission–the result of an American cultural tendency to assume an immutable democratic process–or a nuanced understanding of realpolitik (in a domestic political connotation).
I’m left wondering, perhaps naively, whether lower transaction costs in the political sphere have resulted in a tsunami (party and process change) which we just aren’t aware of because we are in the relatively deep, open water the system, process and cultural bias provide; or whether we’ll just see wildly spectacular, but relatively unremarkable surface waves (issue constituencies emerging within the big-tent, two-party system). Technology has so far failed to deliver easier participation in actual democratic action (voting), but is successfully connecting people and values in social networks with which they can influence rather than act. The question is how empowering are these networks and just how much change can they exact? Will they be subsumed by the existing modus operandi, or fundamentally change the system?
From RWW:
Realistic video search seems to be a long ways off from where I sit, but I’m not sure I buy how disruptive tagging actually is. Yep, I’ve read Shirky, I use del.icio.us regularly and I appreciate and argue for tagging in the applications and services I use. The problem is I actually loathe it. Most of the time I can’t be bothered to tag anything. I’m lazy. I don’t tag diddly squat in Lightroom, and find tagging in del.icio.us a chore (I appreciate it later on when I’m looking something up, but man does it suck during creation). Granted, I’m a weirdo, but I don’t think I’m alone in being a belligerent and lazy tagger.
In my view it’s going to be up to client software to automatically tag items for me. Working for a screen capture / screen recording company I often think of this in terms of the the recording client tracking what application I’m using or what URL I’m visiting and attaching metadata to my output. This “contextual use” data would be stored alongside things that normally come to mind when you think of automatically generated metadata — time, operating system, geographic location, etc.
When you combine baseline metadata with contextual information I think you end up meeting the “good enough” threshold Bernard establishes (sprinkle in some good old demographic data and things get even more interesting). And this approach is not confined to screen recording clients and the media they spit out. All content creation clients need to get “smart” and provide this extra layer of data, or at least expose it as a set of automatic default tags which can easily be cleared away or supplemented.
Perhaps I’m just not seeing the forest for the trees — the open, public social media repositories on the web and massive scale of users may mean that none of us actually have to tag all that much. Or maybe speech recognition is all that’s needed to really need to drive down the tagging pain threshold. If you’ve got insight, I’m listening.
Mark Cuban sees no-model and a bleak future for video on the web, but I’m inclined to agree with one commenter at least.
Call it old media, or old software, but one thing is clear — neither seems to really know who the hell their audience / customers are. You see this reflected back in passive ads that fail to engage (why pay attention if what I’m watching has no relevance to me) and unsightly, uncompelling software that’s difficult to use.
It’s no longer possible to assume a broad and captive audience, but that’s a good thing if you’re a content creator, advertiser or software developer that has a deep understanding of who your audience is and where they’re at.
To pull a Clay Shirky, the real problem is that the distribution of video is no longer monetizable. Control of delivery guaranteed “final cut” over content creation and total leverage over monetization. Since relatively little could be known about the viewing audience (who was watching and their level of interest) only mass market content was financed and marketed to consumers and advertisers. This is any easy system to administer for distributors and advertisers, but is extremely inefficient.
Other factors also have contributed to the old model’s decline. Television replaced radio as the only game in town for a largely rural / suburban population, but has since been challenged by the increasing number of entertainment options available to a highly mobile and increasingly urbanized population. Combine these factors with the rise of social computing and interactive media and you have a lethal cocktail.
As with most things, I see a lesson in software development here. Old software assumes a broad, apathetic, if not quite passive, audience and tries to lock them into a one-size fits all model. Old software doesn’t really understand its audience, so it tries draw an audience by offering a smorgasbord of diluted features. Old software won’t be effective in making an emotional connection with its audience (beyond anger) and isn’t relevant enough to withstand cost or vertical pressure.
…at least that’s the way I see it. :-)
What happens when you lower transaction costs to the point that making “interface” changes becomes an unremarkable effort? You get more change and less discipline. In some cases the results are positive and in others, not so much as Danah Boyd points out.
In the software industry we often pride ourselves on iteration and view the ability to quickly change an interface as inherently good. It’s nice to be reminded of the costs of “improvement”.
The realization that targeted branding is actually very close to user interface design was hammered home as I read through an old Danah Boyd post.
Danah Boyd on personalization within the context of tailored branding:
Bret Victor on context-sensitive information graphics:
One of the reasons Flash has been so important to the brand conscious on the web, is because its ability to present emotional, powerful branded aesthetics combined with rich contextualization. What Danah and Bret do extremely well is explain why this is important. Going forward I’d argue that most software–desktop, web, RIA, or otherwise–will be judged by the market on its ability to deliver the two pillars of rich experience:
- personalization
- contextualization
I surf the web, my feed reader, twitter and a lot of the other information sources. Often I find I want to use a highlighter or otherwise annotate my reading material just as I would of done back in college (mabye a little less liberally with the highlighting). Jing is great for this and I’m a huge fan of it’s tag template mechanism which allows me to create embed / image tag templates that get returned when I press the link button.
Here’s what gets returned to me by Jing after clicking the “embed / tag” button:
However, what I really want most of the time is an image (img) tag wrapped by an anchor (a) tag. The anchor tag contains the link to the article where the screenshot was taken:
I call this technique disclosure to context (maybe “establishing shot” is more apt–spit it out if you have something better). Annotated screenshots and screencasts are brilliant at providing micro-context (hey, this over here is what I find interesting), but fall flat when providing macro-context (where the hell did this come from and why can’t I see). If you put them together however, you have the ability to “stream” what you find salient within its broader context - a more cerebral form of lifestreaming. There’s all kinds of contexts where this can be used…maybe I’ve stumbled on a new microblogging format (jingstream, screenstream, dstream - desktop stream, vstream - virtual stream).
I’d love to know if anyone knows of a microformat which does this for items that don’t have a web URI (this is my image and this is the application, or context, it came from). Also any pearls on embedding this type of information within the binary file itself with XMP, or otherwise, gets you a beer or my undying gratitude the next time I run into you. ;-)
It’s amazing how much the world of software development is changing. What was once a clandestine affair now extends to early public alphas / betas. For those not quite so inclusive, a Flickr stream of UI screenshots lets the masses get involved.
Why screenshots of the UI? Like it or not, everyone considers themselves a UI expert. An application or website UI is likely to draw a considerable amount of feedback (or ire — anyone remember the great Adobe icon debacle of ‘06). In the case of the “icon debacle” by releasing previews early Adobe was able to free the actual product release window from considerable negative noise by showing the icons early and engaging in a dialog with the protestors. Adobe didn’t change the icons, but it gave consumers a chance to digest the change and develop an alternative set of icons that could be used if you were a hater.
A stream of ui screenshots seems like a trick every product marketer / PM should have in their bag. What I’d really like to see is a stream that went from start to finish. It would be great to see the complete evolution of ideas from the initial designer comps to the final build. On release day throw together a time lapse video and you’ve got ready made marketing Kool-Aid that illustrates the hard work and refinement that went into the release.
For the record, I liked the CS 3 icons then and still like ‘em now. ;-P
If you are looking for something similar to flvtool2 or flvmdi, or you are a command line junky then I strongly encourage you to check out MP4Box. MP4Box has a ton of capabilities, but I’m just going to highlight 4 fundamentals:
- Read file info (basic metadata).
- Split or cut files.
- Merge two or more files.
- Convert to iPod.
Read basic file info from an MPEG-4 file:
Split / extract / cut an MPEG-4 file:
Read the info from one of our splices and make sure we’re getting what we expect:
Concatenate / merge multiple MPEG-4 files:
Merged file information:
Convert to iPod file:
MP4Box is a GPAC project, but they only distribute the source (no binary downloads). If you are like me and not keen to try and get Make to compile the source you can grab a binary from here. One last thing–read the docs–you’ll be amazed.
I spent about 15 minutes in the Google Code twilight zone this morning. For some reason when hitting the SWFObject 2 Google Code repos (via a Google search) I wasn’t finding anything on the wiki, downloads page or any source in the repository.
An empty wiki page.
Downloads MIA.
I tried to hit the SWFFix dev blog to check for additional insights and found that if I tried to get to it via a google search (hitting the top link named ‘SWFFix Dev Blog’) it redirects to a domain parking page.
I hit the top link in the search results with a URL that matches what I want
I end up redirected someplace else.
If I type the url (http://www.swffix.org) manually I get to the blog without issue. However, there’s no mention of any Google Code issues.
…Something is rotten in the state of Denmark. I did another Google search for SWFObject 2 and this time the search results sent me the actual repository. It turns out my first search results had led me, (for some unknown reason) to a Google Code Project with a nearly identical URL.
The actual Google Code project.
The incorrect Google Code project.
Easy mistake to make, but what the hell is up with Google giving me craptacular / inconsistent search results and the results resolving to different urls than they should be?
…I’ve got my eye on you Google.