21912

Physical Interface Standard - Thesis

--> There are three components to every physical interface - the structural, the logical and the informational. The structural defines the configuration of the physical interface. The logical defines the dialog (language) between assembled components. The informational defines the contents of that dialog. A nut and a bolt have three interface components - 1) the physical male/female thread pattern, 2) the friction and mechanical qualities of the threads, 3) the temperature differences, voltage differences, etc. of the nut and the bolt. A USB interface also has three components - 1) the physical pin and connector layout, 2) the electrical signal protocol, 3) the information implied by the bits flowing through. A handshake - 1) the hands, 2) the mechanical qualities of the connected palms and fingers, 3) the pressure, temperature of the individual hands. Drinking a beer - 1)the lips and the rim of the glass, 2) gravity, 3) the beer.

How does one use this to construct a philosophy of building things?

Maybe define rules for each component.

Structural - interface components must be explicit and easy to understand.
Logical - how the interface works must be obvious. Permutations of the interface should also be obvious or easily discovered.
Informational - what the interface communicates must be highly structured, contextual and self sufficient - "transactions" have knowledge of themselves, etc. Much could be taken from the digital world here.

Needs work, but I feel like I'm moving in the right direction.

Object Sentences

I love language and have always been fascinated by its mechanics - how the assembly of words in a sequence can either convey meaning or sow confusion. Given my interest in the value of splitting structure (grammar/syntax) and content (words), language has been a potent catalyst. When I was younger, one of the questions I turned over in my head all the time was - why can't there be a language in which it was impossible to speak nonsense - meaning a language where ANY combination of words made sense somehow. It's a wacky question but points, I think, to an important strength of every successful language - it's incredible flexibility. Without the ability to write nonsense, nothing new could be done. Our idea of nonsense is subjective. The poetry of e.e. cummings would have likely been viewed as complete garbage 200 years ago. Or how about a computer program?? But the other thing that fascinates (inspires) me is that even though the interface between words is formally defined as grammar, it is also totally open, meaning those rules are really only a guide. Poetry routinely flouts the rules to great effect. If I know the words of more than one language the options for sentence construction increase, as long as the reader understands the languages used and doesn't mind dealing with no formal grammatical structure. Words, like atoms, seem capable of infinite types of combinations, but gain particular value when used with a logical grammar (e.g. Shakespeare or Hemingway). All of the above, by the way, is equally true of music, but I digress.

Lately I have been puzzling over creating what in essence is a grammar for the construction of not word sentences but "object sentences". Is there a formal way to define how objects should/could go together to create a useful thing or "sentence"? Sounds crazy because maybe it is. But consider electronics, which for the purposes of this argument I will equate to a language - meaning it is contained and its moving pieces defined. At a very high level every device, gadget and gizmo is identical - power supply, processor(s), i/o devices. Mix them up in lots of different ways and you get everything from a digital watch, to an iPod, to an IBM mainframe; like the words of a sentence. Take vehicles of any kind - the same high level analysis produces the same result - all vehicles are identical. So what? My point is that, at least in some cases, there is a starting baseline of components (words) for the construction of certain categories of product. If you could elegantly define the interfaces between these basic categories perhaps you could start to generalize and identify an interface "grammar", at least for that "language". Will work on that next.

Warmspots

Not hot, warm. Meaning, a portable router contained in a small wireless device that an individual carries around with them (could even be a cell phone). Create a mesh of users. Assume that some percentage of people are always near an actual hotspot and therefore wirelessly connected to a broadband connection. Make this mesh generally available and you now present voice carriers with a big problem in the form of VoIP. This is already happening on the vehicular level. Given the potential for disruption, I'm not surprised by reports like this or this. Seems like a great way to wipe out the strangle-hold wireless carriers want to keep on access to their infrastructure. It could also mean that, like the restriction-free WWW, real innovation in the wireless space may finally bloom.

3D-XML

XML has revolutionized how we manipulate information by effectively separating content structure from the content itself. No longer do programmers need to worry about what a particular data stream contains as long as it conforms to the XML spec. Conversely, content providers need never worry again about the mechanics of publishing their information, which alleviates a huge headache. XML is at the heart of the latest Web Services craze and promises to open up whole new categories of composite applications (see Halsey Minor's Grand Central Communications for a great example) that would have been impossible to do just a few years ago (anyone remember Asera?). Ultimately, it's a move from closed, proprietary interfaces to open, standardized ones with all the attendant benefits. But these advantages are really only available in the digital world. On the physical interface front, precious little has been done. There are standardized electronics connectors - USB, IEEE-based things (Firewire, etc.) and MIDI, standardized techniques (how to safely build a house out of wood) and standardized testing bodies ("UL-listed") but nothing even remotely close to XML exists with respect to the formal elegance it imposes on the discrete components of, say, a web site or a financial transaction. In the 3D world there is no formal, standardized roadmap for the construction of any physical interface. Shouldn't there be a high level way to define how, for example, I should connect my computer to my car? Or connect my sink to my refrigerator. Or my computer to my sink(?) Is there a model one could set up that neatly defines the separation of the interface structure from the interface presentation? No, or at least not to my knowledge. It's an interesting issue that may, in fact, have no answer. Digital content is really just ones and zeros, so at a fundamental level, it's all the same fabric. The physical world is far from that simple so it may be that getting to a common interface definition can't happen (at least not unless you're comfortable talking about atoms). Need to think more about this.