Last week, i had drinks with Ian Rogers and Kareem Mayan and we were talking about shifts in the development of technology. Although all of us have made these arguments before in different forms, we hit upon a set of metaphors that i feel the need to highlight.
Complete with references to engineering, technology development was originally seen as a type of formalized production. You design, build and ship products. And then they’re out in the wild, removed from the production cycle until you make Version 2. Of course, it didn’t take long for people to realize that when they shipped flaws, they didn’t need to do a recall. Instead, they could just ship free updates in the form of Version 1.1.
As the world went web-a-rific, companies held onto the ship-final-products mentality in its stodgy archaic form. Until the forever-in-beta hit. I, for one, love the persistent beta. It signals that the system is continuously updating, never fully baked and meant to be organic. This is the way that it should be.
Web development is fundamentally different than packaged software. Because it is the web, there’s no vast distance between producers and consumers. Distribution channels cross space and time (much to the chagrin of most old skool industries). Particularly when it comes to social software, producers can live inside their creations, directly interact with those using the system, and evolve the system alongside the practices that are emerging. In fact, not only can they, they’re stupid to do anything else.
The same revolution has happened in writing. Sure, we still ship books but what does it mean to have the author have direct interaction with the reader like they do in blogging? It’s almost as though someone revived the author from the dead [1]. And maybe turned hir into a kind of peculiar looking Frankenstein who realizes that things aren’t quite right in interpretation-land but can’t make them right no matter what. Regardless, with the author able to directly connect to the reader, one must wonder how the process changes. For example, how is the audience imagined when its presence is persistent?
I’m reminded of a book by Stewart Brand - How Building Learn. In it, Brand talks about how buildings evolve over time based on their use and the aging that takes place. A building is not just the end-result of the designer, but co-constructed by the designer, nature, and the inhabitant over time. When i started thinking about technology as architecture, i realized the significance of that book. We cannot think about technologies as finalized products, but as evolving architectures. This should affect the design process at the getgo, but it also highlights the differences between physical and digital architectures. What would it mean if 92 million people were living in the house simultaneously with different expectations for what colors the walls should be painted? What would it mean if the architect was living inside the house and fighting with the family about the intention of the mantel?
The networked nature of web technologies brings the architect into the living room of the house, but the question still remains: what is the responsibility of a live-in architect? Coming in as an authority on the house does no good - in that way, the architect should still be dead. But should the architect just be a glorified fixer-upper/plumber/electrician? Should the architect support the aging of the house to allow it to become eccentric? Should the architect build new additions for the curious tenants? What should the architect be doing? One might think that the architect should just leave the place alone… but is this how digital sites evolve? Do they just need plumbers and electricians? Perhaps the architect is not just an architect but also an urban planner… It is not just the house that is of concern, but the entire city. How the city evolves depends on a whole variety of forces that are constantly in flux. Negotiating this large-scale system is daunting - the house seems so much more manageable. But 92 million people never lived in a single house together.
[1] Note to Barthes scholars: i’m being snippy here. I realize that the author’s authority should still be contested, that multiple interpretations are still valid, and that the author is still a product of social forces. I also realize that even as i’m writing this blogpost, its reading will be out of my control, but the reality is that i’ll still - as author - get all huffy and puffy and try to be understood. Damnit.
1. Ben on July 19, 2006 1:38 PM writes...
Form follows function.
A good architect will continue to apply this principle even to later customizations. If the function adapts over time, the form shall follow.
In fact, a good architect will anticipate future functional needs and build them in early. This especially applies to city planners who anticipate future infrastructural needs that may be more expensive in the future - such as highways for increased traffic.
There are many much more information centric rules to apply, such as keeping the network stupid and allowing people at the ends to have the freedom to change things, but many principles of the past will continue to apply.
Permalink to Comment2. Anonymous on July 22, 2006 6:35 PM writes...
Form follows function is one of the biggest mis-statements about architecture of all time. It is an almost useful and hollow phrase. Whoever said it first was really on an ego trip. It is too arogant of designer's to presume they can define the function. All you can really do as a designer, is to provide a kind of solution, which will not prevent the user from making use of it, and being able to change use over the course of time. If you build in too much function, you are not designing for the 100 year usage, but rather the next decade at the most. It is hard though, the architects of the internet like Kahn weren't thinking in terms of information transmittal, but just of transporting the new fangled things called bits. David Isenberg has covered this topic quite well I think. The expression that David uses, is to make the network dumb. In other words, you do not allow your own short term definitions of 'function' to dominate your perception of problem and solution space. But rather you define as little as you can. That is the real skill of architects, at the end of the day, the real masters.
Brian O' Hanlon.
Permalink to Comment3. Mary Bryson on July 25, 2006 12:22 PM writes...
I have been scouring the metaverse for an account of instructional design that feels useful for web 2.0 - your "persistent beta". There isn't anything helpful in "education world". The closest I have come is Tom Moran talking about "everyday adaptive design".
http://www.cityofsound.com/blog/2002/08/tom_moran_on_ev.html
Instructional designers still have a lot of clout. It's definitely time to pop that bubble.
Mary
Permalink to Comment4. Josh on August 5, 2006 1:20 PM writes...
"As the world went web-a-rific, companies held onto the ship-final-products mentality in its stodgy archaic form. Until the forever-in-beta hit. I, for one, love the persistent beta. Web development is fundamentally different than packaged software."
From a programming perspective this sounds like two things: #1, total horse shit. #2, a great way to justify bugs (not only are they mis-perceived "features" like they were in desktop apps, but now they're organic!!). So #2 is kinda cool, but really this sounds like more web 2.0 hype.
Programming on the desktop is already a nightmare, do you think doing it with OO javascript and a series of packaged "modules" with which you will build your site is going to be LESS of a nightmare?
Permalink to Comment5. Anonymous on August 9, 2006 3:33 PM writes...
Software engineers have a habit of claiming to have the perfect solution. Then having implemented something overly complex, they walk away from it, and it becomes someone elses problem. Of course, as the open source story has demonstrated, sometimes the spectre of 'intellectual property' is what is to blame. There is no equivalent in the world of matter.
At google video, is an interesting lecture given by richard m. stallman, to the german university at ulm.
http://video.google.com/videoplay?docid=669200964006594520&q=stallman
Where stallman, describes a spectrum, with pharma medicine engineering at one end, and software engineering at the other. He answers the question, why is patent infringement such a danger with software. Stallman, describes elequoently the subtle differences better engineering of matter and ideas.
What is really interesting I think, is when you combine prostetic devices, limbs, things embedded in skin etc, and then expect all of it to work. We saw it already on web 1.0, with the dreaded 404 errors. What has been documented in web 1.0, is mainly it's spectacular crash. But that aside totally - the more I find out about web 1.0, the more I learn it was a half baked attempt at executing the original specs of TBL.
Thackara does consumate justice to the problems associated with 'smartness' in his chapter of that name in 'In the Bubble'. You throw in crowd sourcing, singularity and pro-active computing, and you realise how imbalanced the current situation really is.
Google video has a video interview with jaron lanier, where he nails the problem too. As Thackara managed in his 'smartness' chapter. Lanier uses the microsoft auto word completion tool, as a benign example, but also a precedent. The exact same point is repeated by John Thackara - that basically, we are expected by the software, to become the computer, and the computer is trying to become the human. From a design point of view, I guess, that strategy is fundamentally doomed.
It almost goes right back to those old struggles between the Artificial Intelligence point of view, and the Intelligence augmentation POV. If you ask me, the web, was a poorly implemented version of IA. Lanier is critical today, that Minsky, Kurzweil and their band of AI followers are winning.
As Lanier says, that goes right back to Minsky, Kurzweil etc, etc. Even Rob Cailliau, the first 'project manager' for the web, in his book, about the history of the web - noted how frequently software designers screw up the implementation of ideas. I remember, Cailliau describing the early experience of Ted Nelson working for a newspaper in new york, where they used glue for sticking together the sheet layouts. And how poorly, this artistic process, was translated by software engineers.
We have other evident too, that the future does not look good, like Bill Joy's famous contribution to the debate at wired mag.
Permalink to CommentBrian O' Hanlon.
6. Anonymous on August 12, 2006 1:37 PM writes...
"What would it mean if the architect was living inside the house and fighting with the family about the intention of the mantel?"
Now, that I think about it - Neil Gershenfeld's book, When Things Start to Think - asks the opposite question to yours. I remember some line, which may have been from Gershenfeld's book, about the a man who was suspicious, his toaster was conspiring against him.
Yourself and Gershenfeld, do both, seem to be looking at that same problem from alternate directions. Like in Kevin Kelly's book too, there are two main points. Nature become mechanical and Machines becoming natural. Of course, Gershenfeld has ran down this road too, of late. There is a very good paper by Gershenfeld at Edge.org, where he thinks about computing using natural substances.
George Dyson, has an article there too, about the Turing machine, and goes back to von Neumann etc. Very, very enjoyable reading. Don't forget though, someone like Bill Joy, who recently only made the move from technologist, to venture capitalism, in the field of biological engineering.
Bob cringely at nerd tv, has a good interview with Joy, on the subject.
Brian O' Hanlon.
Permalink to Comment7. Anonymous on August 12, 2006 2:18 PM writes...
Wrote this back in June, posted it up at CG Architect web site in PDF. I was reading Nicholas Carr's posts about utility computing at the time, and trying to discover what is this thing called Moore's Law, which people always refer to. But like many laws, such as Metcalfe's, Reed's, Brook's Law,... we tend to forget what they actually meant, in their original format. I don't think many people now, remember living prior to Moore's Law. That is the problem.
http://www.cgarchitect.com/vb/18646-article-computing.html#post129895
Nicholas Carr, wrote up a post about developing on the new utility platforms here:
http://www.roughtype.com/archives/2006/07/between_two_wor.php
Brian O' Hanlon.
Permalink to Comment8. Anonymous on August 21, 2006 12:57 PM writes...
Nice link too, to an architect offering his opinions about web space. The title alone, a space for half formed thoughts, suggesting something.
B.
Permalink to Comment9. fille8x on May 6, 2007 11:16 AM writes...
je veux les fimls de " urban planning "
Permalink to Comment