Infrastructure for Personal Liberation Nathaniel Borenstein Email World Conference Chairman's Address Wednesday, November 3, 1993 1. What Are Mail Enabled Applications? 2. The Promise of Liberation in Space and Time 3. Examples of Operational Mail-Enabled Technology 4. You Ain't Seen Nothin' Yet 5. Enabled Email as Infrastructure for Cooperative Work 6. An Agenda for Making it Happen - 2 - 1. What Are Mail Enabled Applications? Broad definition: Any activities that are conducted in large or small part via email There's nothing left to worry about! Some such activities work better than others today. We give a thing a name to get a handle on a problem. Practical definition: Any such activities where automation is needed Captures the essence of current needs Unstable, rolling definition (only "needed" until here) Doesn't focus our current efforts - 3 - What Are Mail Enabled Applications? (con't) Conceptual definition: RPC to human beings & vastly distributed services This is what could be done, but isn't. Good focus. RPC + Temporal disconnectedness = mail servers Remote Asynchronous Procedure Call (RAPC) RAPC + human interaction = active messaging Interactive Remote Asynchronous Procedure Call (IRAPC) Goal: shared, standardized IRAPC facilities Needed/possible capabilities are becoming clear Must reach whole Email World (Internet + X.400 + all else) Informed by research & commercial practice, but open. - 4 - 2. The Promise of Liberation in Space and Time What is freedom? A dream we never quite achieve. Mundane view: doing whatever you want ...but you always want more! Buddhist view: an end to desire ...but explain that to your kids! Evolutionary view: doing whatever you used to want ...steady progress towards an imaginary ideal Email may be the next step! - 5 - Work: An Obstacle to Freedom? Freedom is traditionally constrained by the need to eat. But some of us love our work, would or do choose it freely. Primitive people worked hard, had relative control over hours, location, activities. ("Looks like a good day to pick berries.") Industrial revolution: Work as spatial & temporal prison. Computer revolution: "house arrest". Email revolution: "beach arrest" if we do it right. How do we do it right? - 6 - 3. Examples of Operational Mail-Enabled Technology Early Systems: RITA (SRI): Earliest (1976) literature citation? (Never implemented) R2D2 (Vittal): Insecure, low-level UI Imail (Hogg): Not distributed, low-level UI Mail servers (esp BITNET LISTSERV): not generalized - 7 - The Andrew Project (CMU) Sponsored by IBM to build "computing environment of the future" Andrew Message System had 3 relevant experiments Interactive insets (Cookies) Ness -- insecure language for multimedia interaction Flames -- LISP-like language for automatic mail processing Uses: list service, vacation mail, fax gateway, bboard system operation, order processing, more. - 8 - D'l 610p 0.0i'D'l 0.0i 460p'D'l -610p 0.0i'D'l 0.0i -460p' - 9 - Relevant recent work ATOMICMAIL -- First with security, abstract portability ActiveMail (Shapiro talk) -- asynchronous invitations/hookups to somewhat synchronous applications E-Forms (Bolte talk) & Superforms (Bellcore) -- demonstrate value of specialized email forms language Various mailserver toolkits - 10 - Proprietary environments Often extraordinarily good in many regards NeXT Mail -- multimedia support Lotus Notes -- support for widely distributed writing, filing, etc. Magicap -- proprietary language for interactive graphical applications All sought to redefine the world. All are doomed to fail. (Trust me.) - 11 - 4. You Ain't Seen Nothin' Yet -- We've only scratched the surface of mail-enabled applications. To go further requires interoperation, for which the pieces are coming into place. -- Standard interchange format: MIME -- Generalized tools for mail severs: ServiceMail -- Standard format for interactive messages: Safe-Tcl? (Note limitations, contrast to ActiveMail & Notes) -- Nowhere are "open systems" more important. - 12 - D'l 595p 0.0i'D'l 0.0i 425p'D'l -595p 0.0i'D'l 0.0i -425p' - 13 - 5. Enabled Email as Infrastructure for Cooperative Work -- Imagine that some interactive message format is standardized. What becomes possible? -- Whole new kinds of distributed services can be built on the assumption of arbitrary user interaction as needed, with time delays. -- The Electric Eclectic: How mail-enabled applications can redefine publishing -- Mr. Wizard: A new approach the problem of Organizational Memory -- Grander vision: everyone who so desires working on their own time/terms - 14 - The Electric Eclectic: A Focus & Testbed for Open Mail-Enabled Applications (Yesterday) A free, volunteer-based, customizable multimedia metamagazine for the Internet, using MIME & Safe-Tcl Initial virtual magazines include "Email Universe" New magazine about and via email Socialization environment for Email People Forum for technological experimentation Vehicle for prototyping: Subscription customization Reader-feedback Advertising & Order fulfillment Interoperable mail-enabled applications in general - 15 - Mr. Wizard: A Tool for Distributed Organizational Memory Organizational Memory. Someone knows X, how do I find out? Traditional approach: Ask a neighbor. Netnews: Broadcast query to world Innovation: Malone & Ackerman's "Answer Garden" inhibited by software distribution & user buy-in. Solving the distribution problem generally one time (Safe- Tcl?) eliminates it for apps like this. User buy-in is no problem if mail-enabled. - 16 - Organizational Memory: An Interactive email approach User sends query to "Mr. Wizard" server Answers sent back to user with interactive email for feedback. If no good answers, interactive email queries go to experts for answers or names of further experts Simple queries answered from database Moderate queries answered by local experts Hard queries cross oceans to find experts Status: I've got a grad student looking at it. - 17 - 6. An Agenda for Making it Happen -- Forge consensus for a single interactive mail language. -- Eschew proprietary formats. -- Assume the best about other colleagues & competitors -- Expect continuous evolution - 18 - Forge consensus for a standard language -- Unify, unify, unify. We must converge on an interactive mail language. -- If we aren't smart, this could take decades. Standardize a programming language? Remember Ada? -- There is no better game in town than Safe-Tcl. Credibility aid: I have abandoned all my previous interactive mail languages (Ness, ATK, ATOMICMAIL) for one I did not invent. -- Safe-Tcl still has flaws. The community must refine & standardize. But it's the best starting point we have. Demand EXTREMELY good reasons for not using it. - 19 - Eschew proprietary formats It is undeniably easiest to make progress in functionality with proprietary formats. A hard-to-accept truth: In email, proprietary functionality is non-functional. We've all built great systems that only fragment the user community further. (I'm happy to accept some of the blame for the good things in Andrew, but it isn't all my fault.) Fight especially hard against the best proprietary formats, they will slow us down the most. The worst thing about Lotus Notes is how good it is. - 20 - Assume the best about other colleagues & competitors Assume the other guys are: smart well-intentioned generally on the right track You'll be wrong occasionally, but that will happen anyway. Avoid gratutious reinvention. Avoid gratuitious debates Get leverage even from your competition. - 21 - Expect continuous evolution Plan all your products for format evolution. That's how the world works. MIME is open-ended because life is open-ended. Safe-Tcl will evolve or be superseded. Better richtext formats and character sets will come. Digital smell is just around the corner. Expect a continuing adventure and you won't be disappointed. Plan your code accordingly. - 22 - The Electric Eclectic: How To Participate Subscribe: ee-subscribe@eitech.com Join discussion list: ee-discuss-request@eitech.com Volunteer: ee-volunteer@eitech.com Submissions: ee-submit@eitech.com (or physical mail) D'l 187p 0.0i'D'l 0.0i 61p'D'l -187p 0.0i'D'l 0.0i - 23 - -61p'