From carson@siggraph.org Wed May 7 01:55:00 1997 Received: from ratatosk.DK.net (root@ratatosk.DK.net [193.88.44.22]) by dkuug.dk (8.6.12/8.6.12) with ESMTP id BAA25421 for ; Wed, 7 May 1997 01:54:58 +0200 Received: from siggraph.cgrg.ohio-state.edu (siggraph.cgrg.ohio-state.edu [128.146.18.100]) by ratatosk.DK.net (8.6.12/8.6.12) with ESMTP id BAA03565 for ; Wed, 7 May 1997 01:54:56 +0200 Received: from study.huntleigh.com (carson@siggraph.org) by siggraph.cgrg.ohio-state.edu (8.8.5/941010.52) with SMTP id TAA00498 for ; Tue, 6 May 1997 19:53:18 -0400 (EDT) Message-Id: <3.0.32.19970506174902.00775cf4@siggraph.cgrg.ohio-state.edu> X-Sender: carson@siggraph.cgrg.ohio-state.edu X-Mailer: Windows Eudora Pro Version 3.0 (32) Date: Tue, 06 May 1997 17:55:07 -0600 To: SC24@dkuug.dk From: Steve Carson Subject: Re-engineering Technical Directions Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable To help SC24 National Bodies formulate their positions on the proposed Technical Directions for JTC1, I have prepared some text with supporting rationale as a starting point for further discussion. I have been helped in this by private comments received from several NBs. Rationale ----------- First, I propose that any response (SC24 or your own input to your JTC1 level NB organization) be organized around answering the four questions proposed by the Reengineering ad-hoc. Second, I believe that all of us find serious flaws in the list of proposed "Technical Directions". I suggest that we try some gentle persuasion to make the list better, but we have to realize that Reengineering in JTC1 is being accomplished by individuals who are non-technical managers of standards work rather than by technologists who understand the subject matter they are dealing with to an appropriate depth. Third, I suggest that we point out that the entire present scope of SC24 is left out of the Technical Directions and we: a) make up a new Technical Direction that includes our scope; b) add words to the existing directions that differentiate them from ours. Draft text ------------ Proposed Title: Response of SC24 to ISO/IEC JTC 1 N 4631 Based on the above rationale, here is some draft text organized as answers to the questions: 1. Do you consider that your present projects (or project groupings) align with the proposed recommendation? If not please provide rationale for a different alignment within the overall set of technical directions. No, almost none of the present activities of SC24 fit solely within with any one of the proposed Technical Directions. To explain this, it is necessary to carefully explore the proposed scopes of some of the new Technical Directions against the approved scope of SC24. First consider three of the proposed Technical Directions that have some relationship to the present scope of SC24: Multimedia & Representation: Coding of hypermedia and multimedia and coded representation of picture, audio and character information, including - sets of compression and control functions for use with such information, - interchange of such information. Programming Languages and Software Interfaces: Programming languages, their environments and systems software interfaces, including: - specification techniques, - common facilities, and - language bindings. User Interfaces: User-system interfaces for input devices such as keyboards, mice, pointers and pens; rules for systems control by voice recognition, presentation techniques, dialogues, graphic symbols and icons including those for people with special needs. In contrast, the present scope of SC24 is: Area of Work: Standardization of interfaces, in windowed and non-windowed environments, for: =B7 computer graphics =B7 image processing =B7 interaction with, and visual presentation of information Included are the following related areas: Related reference models; application specifications; functional specification; interchange formats; device interfaces; testing methods, registration procedures; presentation and support for creation of multimedia and hypermedia documents. The following observations can be made: a) The work of SC24 deals primarily with the functional specification of interfaces and with aspects of the presentation of and interaction with information that are driven by and derived from such interface specifications. There is only a minor focus on the coded representation of information and even there it is common for an interface of interchange format specification developed by SC24 to allow many different coded representations depending on the characteristics of the specific= application. b) Based on observation (a) above, we conclude that the Multimedia & Representation Technical Direction as presently defined by the JTC1 Reengineering Ad-hoc has only a minor relationship to the work of SC24 and its present projects. In our view, choice of coded representation should be driven by functional considerations and in this sense the work of SC24 would fall into a yet undefined Technical Direction that deals principally with the functional aspects of the Presentation of and Interaction with Information rather than the coded representation of such information. Alternatively, the Multimedia & Representation Technical Direction could be expanded to include both aspects, with a recognition that functionality considerations are an important input to the development of coded representations. c) The Programming Languages and Software Interfaces Technical Direction as defined by the JTC1 Reengineering Ad-hoc includes "systems software interfaces". This description is so ambiguous that some might think that the includes "systems software interfaces" being described include those used for the Presentation of and Interaction with Information. While it may be beneficial to recognize that there can be common generic approaches to defining such interfaces (for example, a common set of language independent data types), the functional specification of such interfaces should be defined by experts in the particular subject matter, not by programming language experts. This is true whether the "systems software interfaces" be those used for Presentation of and Interaction with Information, Distributed Application Services, Networking & Interconnects, Data Capture, Security, or another area of Information Technology. d) The proposed User Interfaces Technical Direction has aspects that overlap (but do not strictly contain) many of SC24=92s projects. This overla= p may be due to the poorly worded description that fails to differentiate those aspects that are purely a "user interface" concern from those that concern the Presentation of and Interaction with Information. It is widely recognized by Information Technology specialists that the design and implementation of effective "user interfaces" is a separate discipline from the Presentation of and Interaction with Information. The later has sub-disciplines called computer graphics, multimedia, interactive techniques, and so forth. The former consists of disciplines such as human-computer interaction and human factors and is as much concerned with psychology and organization dynamics as it is with Information Technology. A technical direction in Presentation of and Interaction with Information would provide the basic tools, techniques and infrastructure building blocks that could then be used within a User Interfaces Technical Direction to develop and define effective interfaces. By way of example, consider an interface that uses iconic representations of familiar objects, such a file folders. Experts working within a User Interfaces Technical Direction might choose the best iconic way (in terms of the "picture", size, shape, color etc.) to represent that object to a given class of users, perhaps dealing with cross-cultural issues in the process. A Presentation of and Interaction with Information Technical Direction would define abstract techniques that could draw pictures of icons and show the state of an icon (such as available, selected or disabled) based upon user interaction. These abstract techniques could be used to display and interact with a wide variety of icons and might by themselves be used to build either very effective or very poor user interfaces. By combining the knowledge and facilities developed by both Technical Directions a programmer could construct an effective an standard user interface with minimal effort. For these reasons, we believe that a new Technical Direction should be added. It would be described as follows: Presentation of and Interaction with Information Functional and interface standardization for the presentation of and interaction with information, including: - computer graphics; - multimedia; - image processing; and - interactive techniques The work in the proposed Technical Direction can be clearly differentiated from work in the Technical Directions proposed by the JTC1 Reengineering Ad-hoc as described above. Some aspects of some of the projects within at least the following SC=92s fall within this new proposed Technical= Direction: - SC18 (aspects of "user interfaces" that deal with functional or interface specification rather than "human computer interaction") - SC22 (systems software specifications for interfaces used for the Presentation of and Interaction with Information). - SC24 (all) - SC29 (functional specifications and architecture); This proposed organization aligns with the real-world situation in Information Technology. Both the Information Technology industry and the academic community recognize disciplines of Computer Graphics and Interactive Techniques and of Multimedia within organizations. These disciplines are supported by major professional organizations such as the Association for Computing machinery (ACM) and Eurographics. Within the ACM there is a Special Interest Group on Computer Graphics (SIGGRAPH) and another for multimedia. Most experts that contribute to SC24 belong to SIGGRAPH and many also belong to SIGMM and Eurographics. The annual meeting of SIGGRAPH each year is the premier event in the computer graphics industry worldwide and attracts approximately 5,000 fully paid technical attendees as well as over 30,000 exhibits only registrations. SIGGRAPH also sponsors more specialized conferences devoted to VRML (the subject of one of SC24=92s most active current projects) and 3D and Interactive Techniques (one of SC24=92s historical focus areas and the subject of standards such as PHIGS.) 2. Do you believe your current projects (or project groupings) within the proposed technical direction will benefit from a synergistic fit with other projects,* (if any) in your technical direction? If all current SC24 projects go within the Multimedia & Representation Technical Direction as presently defined by the JTC1 Reengineering Ad-hoc, there could be some synergistic benefits. To achieve this, first the scope of the proposed Multimedia & Representation Technical Direction would need to be expanded to include the work of SC24 (which as described in the response to Question 1 above is presently largely excluded the Multimedia & Representation Technical Direction even though the JTC1 Reengineering Ad-hoc indicates that is believes that all current SC24 projects fall within that Technical Direction). Second, several other important changes must be made in re-aligning the present work of other groups that fall in this area. Some of these are: a) There must be a strong commitment that in the future functional specifications will drive the coded representation(s), not conversely. This might in the future avoid the serious problems such as we see today in attempting to apply coded representations developed from a narrow perspective to the broader problems of the Presentation of and Interaction with Information. Just one example is the error of leaving compression of a fourth channel (alpha) out of representations such as JPEG which renders them sub-optimal or unusable in multimedia systems where generative computer graphics and images are merged at the client. b) The broad needs of Information Technology in general for Multimedia & Representation must be considered, not just the highly specialized needs of one market segment such as "set-top boxes". c) The manner in which work is organized and meetings are held within the Technical Direction must be carefully considered. In particular, "Working Groups" must be able to set their own meeting schedules, meet independently of any parent organizations (except for perhaps at most once per year.) The size of one of the SC=92s alone that is proposed for inclusion within this area (SC29) is already in the opinion of many too large to do the work in an effective and timely manner. Finally we note that if the work of SC24 is incorporated into the proposed Multimedia & Representation Technical Direction with the proposed scope, then most or all of SC24=92s present cadre of technical experts will not continue to participate in the work of JTC1 since they have generally little interest in the specialized area of coded representations. Many of these experts have worked together for almost twenty years to develop and validate the technologies of device and platform independent computer graphics, multimedia and interactive techniques that are the cornerstone of the networked graphics systems that support commerce and technology in almost all areas of human endeavor today. It would be a great loss to ISO and IEC if their efforts moved solely into the many Consortia that today vie with JTC1 for their attention. 3. Do you believe any of your current projects require significant interactions with projects (or project grouping ) in other technical directions? If so which ones? Yes. The focus of much of our current and future work is in the application of our traditional technology focus (the Presentation of and Interaction with Information) to distributed and networked environments, in particular the World Wide Web. Current projects such as VRML, newly proposed projects such as PNG, and our future plans for standardization in the area of WWW based presentation and interaction are designed for and directly applicable to the needs of these environments. Not only do the groupings proposed by the JTC1 Reengineering Ad-hoc not represent the correct subdivision of Information Technology into disciplines (as we explained in the case of our work in our answer to Question 1) but they do not minimize the interactions necessary in real world projects. For example, implementing VRML (Virtual Reality Modeling Language) based applications requires innovative and interacting solutions from at least the following Technical Directions: - Distributed application services: data bases, data structures and semantics, directory, etc. - Identification Cards and Related Devices: personalization of interaction (such as avatar characteristics) based upon identification of the individual and matching to previously stated individual preferences. - Multimedia and Representation: efficient coded representations of graphics, images, sound, genetic algorithms, etc. - Networking & Interconnects: protocols and services with throughput, latency and other characteristics to support real-time interaction in distributed and networked systems. - Programming Languages & Software Interfaces: languages and virtual machines that can execute on the client side to customize user interaction. - Security: identification and authentication of users; safety of executing software transferred on demand over a network. - Software Engineering: design and validation techniques for distributed applications. - User Interfaces: paradigms and rules for the construction of effective interfaces (such as navigation support in 3D worlds and the representation of interaction devices and other real-world objects within a virtual world.) - Presentation of and Interaction with Information (our proposed additional Technical Direction): functional, interface and interchange format specification of virtual worlds and their inhabitants; overall architecture for Virtual Reality applications. Moreover, many of the above interactions are not loose and decoupled ones, but are of such a tight and time critical nature that they require joint design and development activities with close liaisons. We believe that the only solution to this problem is to adopt an organization for JTC1 that is already common industry practice, that is, a matrix organization. In such an organization there would be technical discipline focused units (similar to but refined from the existing SC=92s) and a separate project structure where all aspects of project management would be focused and which could draw upon resources from SC=92s as required= . Such an organization is absolutely necessary because there is no one static organization that is an ideal match for all projects. 4a. Do you believe supplementary standardization activity should be undertaken to sufficiently address the technical direction and which adds value to the work of JTC 1? For example there appears to be added standardization needs in the technical direction "Security" beyond the present work program of SC27. SC24 is already very active in harvesting widely adopted industry specifications into ISO standards. This harvesting is presently hampered by the overlapping scopes of the present SC=92s and will only be exacerbated by the proposed Technical Directions. As an example of one such activity that is difficult or impossible to start within the present or proposed JTC1 structure is and integrated approach to the Presentation of and Interaction with Information over the WWW based on VRML. As described above, this work partially falls within the scope of and overlaps the activities of more than six(6) present SC=92s and new Technical Directions. SC24 cannot by itself initiate or coordinate all the needed work. Only a project organization could do this. 4b. Do you consider that the resources to address these areas will be available? This work is presently being done in various Consortia outside of JTC1, although the overlapping scopes of the Consortia themselves is hampering the work today. Since we regard the only proper role of JTC1 as that of harvesting technologies that are developed and proven outside of JTC1 and making them into ISO standards, the present core of experienced experts within our own area is sufficient to handle the workload (assuming that industry and government continue to recognize the value of and support our work.) The main problem is in attracting experienced managers of large scale projects into JTC1 at (the JTC1 level) to coordinate project-oriented standards development. The number of individuals required is not large, but such people are scarce and their companies must be convinced that devoting a large percentage of their time to a standards project will have appropriate rewards. Currently JTC1 level participants are mostly standards managers in companies, not managers of development activities. It is the later that are needed for an effective project approach to standards.=20 --------------------------------------------------------- Steve Carson phone: +1-505-521-7399 GSC Associates Inc. fax: +1-505-521-9321 5272 Redman Road e-mail: carson@siggraph.org Las Cruces, NM 88011 USA ---------------------------------------------------------