Peter Denning sobre Beyond Calculation

(em inglês)

O nosso modelo actual de "desktop"
está ultrapassado - é limitativo

O co-organizador da colectânea Beyon Calculation não esteve presente na Mesa Redonda virtual que organizámos com Vinton Cerf, Gordon Bell e Bob Metcalfe. No entanto, ele acedeu a responder a questões similares para publicação on line.

Peter Denning é um dos históricos da computação. Decidimos manter a versão inglesa da entrevista.

Jorge Nascimento Rodrigues com Peter Denning

James Burke, in Beyond Calculation, talks about a domino effect in the social and cultural fabric from each one of the computing innovations in the last 50 years. What was (is) for you the most unexpected social and cultural change from the emergence of the following events in the last 25 years:
Foto: Peter Denning - the computer chip
- the personal computer
- the user friendly concept
- the email
- the public Net
- the WWW

PETER DENNING - For me, the most significant change has been the entrenchment of the computer into everyday practice and common sense. When I entered MIT in 1964 to study computers, MIT had no computer science department and was extremely cautious about using the words "science" and "computers" in the same phrase.

I was part of the Multics project. At the first symposium on operating systems principles in 1967, I heard Larry Roberts and Donald Davies present the first technical papers about the architecture of the network that would two years later be launched as the ARPANET. Beginning in 1966, I was heavily involved in the design of virtual memory systems, well before virtual memory was became a standard part of every computer.

In 1981 I co-founded the CSNET, which marked the transition of networking into the US National Science Foundation.

These personal examples illustrate a fraction of the range of the deeply technical design projects in which computer scientists and engineers were involved. Yet, who among us expected that these technologies would become standard fare with PC superstores, mass market software, email addresses on business cards, http addresses on TV advertising, or Dilbert cartoons?

I certainly did not. In retrospect, it is mind-boggling that the esoteric technologies on which I worked have become household words.

From your personal predictions in the last 30 years, what was the most notorious wrong one? And why (viewing from today realities)?

DENNING - Heavens, I've made my share of bad predictions!

In 1975 I was convinced that structured programming was a passing fad because it seemed "obvious" that everyone one would want understandable and demonstrably correct programs -- but the software crisis is still with us.

In 1980 I was convinced that object-oriented computers capable of automatically managing objects at the hardware level would become the next generation of architectures -- but those machines were eventually labeled CISC (complex instruction set computers) and succumbed to RISC (reduced instructed set computers).

In 1985 I was convinced that CSNET would be a model for the Internet because it showed how a community could gather resources to operate its own network -- but today a large majority want networking to be free.

From the expectable trends in computing in the next 50 years, what is the most striking for you?

DENNING - Biotechnological engineering: the prospect of growing organic computers, embedding computers into our bodies, building replacement parts, building nanorobots that repair cells. I recently asked an audience of teachers what it would mean if the human lifespan were extended by replaceable body parts. Some yelled back, "It would take 200 years to get tenure!"

Do you think we have already entered a new wave, a fourth wave, completely different from the Toffler's third wave borned with the transistor and massified with the personal computer? And do you think your personal work (and innovations) contributed to this new wave?

DENNING - We have entered an age of networking. People are beginning to see that the silicon chip and the glass fiber are not the driving forces.

The human urge to engage in relationships, to expand the social space, can now be satisfied in astonishing new ways by the communication tools built from computers and networks. As this happens we are finding that we must cope with all sorts of breakdowns such as our inability to do serious collaborative work over the Internet.

This suggests to me that our perception that the computer processes information and the network transmits it is flawed: we need tools that enhance our ability to build new worlds together. I find the works of Fernando Flores to be especially helpful in thinking about this. (His latest book is Disclosing New Worlds, MIT Press, 1997.)

The transistor era is at the end, as wrote professor Paul Romer?

DENNING - No. Transistors are getting smaller, and much cheaper, but they are not going away. The first transistor I encountered was as a hobbyist kid the CK722, as big as a fat eraser on a thick pencil, costing $10. Today for the same price you can get a chip with a million transistors on it. All the people who are confident that Moore's law will continue to predict growth of chips and memory for the next 20 years are assuming tacitly that chips and memory are made of transistors.

Are we in a transition period from computing to communications, from the computer as a tool to internetworked machines as a new medium, as Terry Winograd pointed out in your ACM recent Conference?

DENNING - Yes indeed. But the shift of computers from crunchers to communicators has been on the minds of our visionary leaders for at least two decades. Those who saw it coming have built industries around this notion.

Do you think we can corral the Net/Web environment into a icon inserted in an operating system, something to add to other icons in my display like Word, Excel, Paint shop pro, etc?

DENNING - No. The network enables us to engage in many relationships. Computers and software programs are objects that we use to enact commitments we make in our relationships. Our current "desktop model" represents only the objects, not the relationships. It is too limiting. The network is not an object, it is the context in which our desktops and the objects on them make sense. The network icon will be replaced by conversations, projects, collaborations, and relationships.

In 50 years, will ACM with C for computing, retains the same name, or will transform computing in something else?

DENNING - Who knows? I have been a member of ACM Council since 1970. Every seven years someone has brought forward a proposal to change the name to association for computing. The constitutional amendments used to fail for lack of votes. More recently, they have failed for lack of a quorum. ACM is in the process of changing itself into an electronic (virtual) community. Who knows what its members will want to call it in 50 years, or whether it will even exist?

Do you think Moore's law will continue, or we will assist a breaktrhough with a new computing device, as a DNA one or a quantum one? What is hype, and what is reality?

DENNING - Moore's Law will hold for another 20 years or so, according to industry experts, as I said before. It is good that people are examining alternatives, even if some of them will never produce a useful machine. For breakthroughs, I'd look to nanotechnology and biomolecular engineering for computational devices made of meat.

What will be your favorite killer application in the near future? For instance speech recognition or a new kind of physical touched interface with the machine?

DENNING - Killer Apps are by definition fads. I don't like to predict fads.