---------------------------------------------------------------------------- The Florida SunFlash S U N E R G Y E M A I L : NEWSLETTER 7, January 1993 (1 of 2) SunFLASH Vol 49 #8 January 1993 ---------------------------------------------------------------------------- From: Miyong.Byun@Corp (Miyong Byun, User Programs) ---------------------------------------------------------------------------- ********************************* * * * S U N E R G Y E M A I L * * * ********************************* NEWSLETTER 7 January 1993 Distributed by the Press Relations, Sun Microsystems Computer Corporation, A Sun Microsystems, Inc. Business; all rights reserved. Director of PR: Kay Hart; Sunergy Manager: Larry Lettieri; Managing Editor: Miyong Byun. Contact the Sunergy editors over email at: sunergy_information@Sun.COM. =========================================================================== = CONTENTS - ISSUE #7 = =========================================================================== Part 1 of 2 ANNOUNCEMENTS *1* Satellite Broadcast Recap *2* Next Sunergy Broadcast ROADMAP TO RESOURCES *3* List of Object References TECHNICAL TOPICS *4* Tips -N- Tricks: Restarting OpenWindows Without L1A *5* "SunWorld" Magazine: Tog on Interface Part 2 of 2 SIG NEWS *6* SunPro SIG NEW PRODUCTS *7* SunPics Announcement *8* SunExpress Announcement UPCOMING EVENTS *9* Tradeshow Calendar --------------------------------------------------------------------------- - ANNOUNCEMENTS - --------------------------------------------------------------------------- ******************************** *1* Satellite Broadcast Recap * ******************************** The most recent Sunergy(SM) satellite broadcast featured our familiar host and a very focused agenda. John Gage, Director of Sun's Science office, brought together a panel of experts to present and discuss the current state of object-oriented technology. The panel included: * Peter Deutsch, SunFellow, Sun Microsystems Laboratories, Inc. * Arvind Deogirikar, Manager, Market Development, SMCC * Steve MacKay, Vice President, User Environment Software, SunSoft The panel overviewed Sun's activities concerning object environments, and also reviewed industry influences and other related topics. Questions called or FAXed in were addressed during the live broadcast. Here's a brief summary. - Evolution of system software - Conceptual ways of thinking about objects: merging function with data characteristics and relationships. - New ideas introduced by objects: dynamic polymorphism, inheritance, secure data storage and manipulation, distribution over the network. - Multithreading and objects - Where objects can be applied - How do objects impact Solaris: DOE will fit on top of SunOS(TM) and ONC(TM) at first. - Tools for object-oriented application development: Sun products, third-party offerings (object creation and management). - New courses for learning about objects: methodology, analysis, design, database design, C programming, C++ programming, C++ for non-programmers. - Tools still needed: design and analysis level tools, managing existing code for realistic re-usability (cataloging, retrieving, and adapting existing code). - C++ shortcomings (predicting the evolution of C++) - SPRING: SunLabs(TM) research projects resulting in IDL and DOE today, future technology for Solaris(R) (goal: total object approach for system, from the bottom up). - Barriers or difficulties for adopting object technology: differing opinions for the object environment model, interface design, performance, GC and type safety, source code. - ToolTalk(TM): an object-oriented approach for integrating applications today We have listed object references the panel mentioned during the broadcast. See the ROADMAP TO RESOURCES section of this newsletter for listing. ***************************** *2* Next Sunergy Broadcast * ***************************** Our next Sunergy Broadcast is slated for March 29, 1993 at 1700, GMT, LIVE from the CeBIT show in Hannover, Germany. If you would like to arrange a downlink in your area, please contact Denise Marcucci at: Phone +1 415 355-9193 FAX + 1 415 738-9210 Email sunsat@woodwind.Corp Overview: Telecommunications networks are a critical part of any nation's infrastructure. Major shifts are taking place in the technology for these networks. Sunergy Live from CeBIT takes a look at what's going in Europe to put in place the next generation telco networks, how countries are coping with the change, and what can be expected in the future from these new networks. Also, if you have any suggestions for our future broadcasts, including topics, please let us know. --------------------------------------------------------------------------- - ROADMAP TO RESOURCES - --------------------------------------------------------------------------- ******************************** *3* List of Object References * ******************************** Here is a list of references mentioned during the Sunergy Broadcast on December 2. Book List ========= - Semiology of Graphics. By Jacques Bertin. University of Wisconsin - Presence: Teleoperators and Virtual Environments. MIT - Object Oriented Technology Portfolio. Sun - Object Oriented Design With Applications. By Grady Booch. Redwood City, Calif.: Benjamin Cummings, 1990. - Object-oriented modeling and design. By James Rumbaugh, Michael Blaha, Willim Premeriani, Frederick Eddy, and William Lorensen. Englewood Cliffs, N.J.: Prentice Hall, 1991. - The Common Object Request Broker: Architecture and Specification published by Object Management Group, Tel: 1 508 820-4300 - Journal of Object-oriented Programming - Object Oriented Programming - An Evolution Approach. By Brad J. Cox - Object-oriented Software Construction. By Bertrand Meyer - OOPSLA '92 Conference Proceedings. Addison-Wesley Publishing Co. Tel: 1 800 447-2226 FTP/EMAIL ========= -FTP iftp omg.org OMG_IDL_CFE_1.0.TAR.Z from the directory pub/OMG_IDL_CFE_1.0. -EMAIL send email with the subject `help' to omg_idl@omg.org (for OMG information available) -Case Interoperability Message Set Document Anonymous ftp gatekeeper.dec.com Get /pub/case/casemsg.ps doc For all questions relating to CASE messages send mail to: casemsg.ext@sun.COM -General DOE Information send email to DOE_info@Sun.COM for autoresponse list of information sources Trade Shows =========== Below is a list of the key US trade shows and events: -CASE World - occurs twice a year, usually in Spring and Fall -CASE Expo - occurs twice a year, usually in Spring and Fall -Triada- Department of Defense show, Fall -Embedded Systems Conference- September -OOPSLA Conference-aimed at object oriented programming -Design Automation Conference -FCC-Federal Communications conference -AFCEA-Air Force conference -International Conference of Software Engineering (ICSE) -Object World -Object Expo -CASE Japan Journals ======== The following list represents the key journals and publications in the CASE market: -Dr. Dobb's Journal (Software Tools for the Professional Programmer) -Software Magazine -AI Magazine -Embedded Systems Programming -IEEE Software/IEEE Computer -Journal of Object Oriented Programming -Programming Update -CASE Outlook -CASE World --------------------------------------------------------------------------- - TECHNICAL TOPICS - --------------------------------------------------------------------------- ********************************************************** *4* Tips -N- Tricks: Restarting OpenWindows Without L1A * ********************************************************** by John Lewis II, john.lewis@sun.com /|||||\ Has your system ever frozen up? |||| |||| ||/\\ \|| || \\ || Did you have to reboot using L1A? /| \\ |\ || | \| /\ || || | |\ /__\ || Doesn't that take a while? || |_ |\\/ \ || \| \\ |/ || \\ || Isn't L1A bad for your system? ||\ \\/|| |||| |||| \|||||/ Well, here's just the tip you need! Most people enter OpenWindows by typing "openwin". But if you type "openwin -escape", you will be in luck the next time your system freezes. Instead of pressing the "L1" key and the "A" key at the same time to reboot the machine, you can press the "Stop", "Alt", and "Delete" keys at the same time which will just exit you from OpenWindows. Then just type "openwin -escape" to re-enter OpenWindows and you'll be back in business in no time. ________________________________________________________________________ STOP ALT DELETE ________________________________________________________________________ An alias provides a way to rename a long command to a shorter command that is easier to remember. So, instead of having to remember to type "-escape" with the "openwin" command, you could just create an alias. Start by putting the following line in your .cshrc file located in your home directory: alias openwin 'openwin -escape' Now, after logging out and back in, you can type "openwin" as usual, and have the "-escape" option added automatically. If your system is set up so that you enter OpenWindows automatically without typing anything, the command "openwin" is probably in your .login file located in your home directory which is read each time you log in. You could add the "- escape" option to the "openwin" command in this file, or the alias will work if you added it to your .cshrc file. If you are in OpenWindows and you are wondering if the escape option is in effect, you don't have to press Stop/Alt/Delete to find out. Instead, in a command or shell tool, type "ps" to list the processes running. At or near the top of the list you should find the "openwin" process: prompt% ps PID TT STAT TIME COMMAND 382 co IW 0:00 /bin/sh /usr/dist/share/openwin,v3/bin/openwin -escape : : In this case, you can see that the escape option is in effect, so if the system freezes, pressing Stop/Alt/Delete will take me out of OpenWindows, and I won't have to reboot. This tip should save you time, and be an easier shock to your system than rebooting with L1A. *********************** *5* Tog on Interface * *********************** "Good interface designers reduce developer engineering costs. Speculation? No. I can prove it." BY BRUCE TOGNAZZINI Sun Microsystems (From SunWorld(TM), December 1992; copyright 1992 SunWorld Magazine, reprinted with permission.) Tog, I read your recent article, "The myth of ease of use," (SunWorld, August 1992) with great interest. I pose to you the following question which reflects a current dilemma many companies are facing today. How does one quantify the cost of neglecting human factors engineering (specifically in software) in terms such that it will help build a strong enough case to help justify expenditure of full-time resources to focus specifically on this area? In your article, you state that... "Sun has assembled what is likely to be the largest human engineering team in any computer software- development organization anywhere." Can you share any specific information that... inspired the assemblage of this large human engineering team? Many ISVs today are failing to dedicate even one full-time body to this end. This, in my opinion, is a huge quality exposure to our collective software product offering. Rich Pusateri HaL Computer Systems Inc. Austin, TX Tog replies: The people who make the buying decisions on computer systems and software are becoming sensitized to the real costs of ownership. Some surveys I have seen have pegged the cost of ownership for a standard PC at upwards of $35,000 per year. That's a lot of money for a $5,000 computer. Where's the money going? A disproportionate percentage can be traced to direct and indirect training costs. Direct costs might entail people taking time from their regular jobs to attend training classes. The indirect costs are more complex: There's the time users spend pressing buttons and flailing through manuals, trying to figure out what went wrong with the machine, when the problem is that they have inadvertently triggered some unknown and less-than-obvious system state. Then theres the time they spend wandering around looking for a warm body in another office to ask for help. Finally, theres the time other users spend being the warm bodies those poor souls find when searching for an answer. Information resource managers are right now putting a lot more emphasis on total cost of ownership, rather than initial capitalization. They are willing and able to spend more money for software that will ultimately cost them less money to maintain. They are no longer prepared as in "the good old days" to accept software that may look good on spec sheets, but results in poor human productivity and high overhead. Good human interface designers earn their keep by lowering customer product training and support costs. They also reduce developer engineering costs. Everyone wins. Speculation? No. I can prove it. +Performing Design Evaluation+ Jeffries, Miller, Wharton, and Uyeda presented a study at CHI'91 ("User Interface Evaluation in the Real World: A Comparison of Four Techniques," Proceedings of CHI'91, Addison-Wesley, Reading, MA, 1991) in which they tested four groups using different approaches to user interface evaluation [Jeffries, et.al.]. Their first group, experienced human interface specialists, formed their evaluations based on experience and an understanding of human factors. The second group consisted of a trained usability tester with six test subjects. The third group, three experienced software engineers, applied a preformulated set of guidelines. The fourth group, also engineers, used a relatively new intellectual testing process called "cognitive walkthrough." When the smoke cleared, the human interface experts were the clear winners. They found almost four times as many problems as the next closest group, and they found more than three times as many serious, "core" problems. The only way to ensure your product design will offer high productivity and low training costs is to hire experienced, competent human interface designers, either full time or as consultants. +Ensuring safety, consistency, and explorability+ The nuclear industry, in its own good old days, paid little attention to human factors. Control room design was done with emphasis on efficient engineering layout, with little thought for the people who would be using the rooms. What little layout planning did occur was typically performed by people with no training in human interaction design. They would apply principles that "just seemed like a good idea." Don Norman, in his book, *The Design of Everyday Things* (Basic Books, New York, 1988, and in paperback by Doubleday, New York, 1990), showed a photograph of a nuclear power plant control room based on a naive understanding of consistency: Every control in the room looked virtually alike. Very clean. Very neat. Very uniform. Very dangerous. The control-room crew, desperate to make potentially hazardous controls stand out, unscrewed the uniform black balls from atop the most critical controls and fastened handles removed from barroom beer spigots. Pulling on a Heineken handle would lower the rods into the The nuclear industry didn't feel the need for human factors types. The control rooms were working out just fine -- right up until Three Mile Island. After so-called "human error" almost caused a disastrous meltdown, the nuclear industry went after human factors people in volume, resulting in major advances in power station safety. Our software is not in danger of immediate meltdown, but some of our users are. Some of the industrys most finely engineered products are all but impossible for people to learn and use. We have gotten away with this situation until now because there has been little software that was both powerful and easy to use, but that is changing. Over the next three or four years, there will be a revolution in high-end software as more and more companies build up their human interface teams. Those who hold back may suddenly find themselves in a noncompetitive position. Sun is taking an aggressive position toward increasing human productivity. We want our developers there with us. +Preventing "human error"+ Our society likes to identify users who have made predictable errors and brand them as culprits. Ten years ago in San Francisco, an armored car driver forgot to fasten the truck's back door, and a sack $100,000 fell out. The money was never turned in. The newspapers announced the verdict was human error, but the real reason the money was lost was lousy human interface design. People do make mistakes. One study cited by Abi Sellen in her doctoral dissertation, found that trained medication nurses gave one in seven doses in error (A.J., Sellen 1990, "Mechanisms of Human Error And Human Error Detection," Unpublished doctoral dissertation, UC-San Diego; K. Barker, W. Kimbrough, and W. Heller, 1966, "A Study of Medication Errors in a Hospital," Fayetteville, University of Arkansas). One of the things human interface designers do is to predict and prepare for those errors. Even the most novice human interface designer, faced with the back door to an armored car, would have fitted a $1.00 switch on the door lock, wired it to a doorbell by the driver, and connected them both to the ignition switch, so that starting the car without the door fastened would make the doorbell ring. A 100,000 to 1 return on investment. If people make "human errors" when using our software, it is our error, not theirs. I don't care what the situation is, the "customer is always right." Human interface designers are trained and talented in flagging such errors before the product ships. It saves users a lot of frustration, it saves information resources departments from having to train people, and it saves developers a lot of money in 200-pound manuals and technical support. +Right from the start+ Many people see the human interface as an evil to be tackled the last three weeks of the engineering cycle. Others prefer to wait until the first bad review. Human interface design must occur before any coding begins. Effective applications, like effective buildings, are designed from the users in, rather than the technology out. Who are your users? What are their needs, desires, abilities? What talents and training will they bring to the party. Until you know the answers to these questions, you cannot design responsive software. Human interface designers are trained to perform user and task analyses to pin down these answers. Just as engineers spend 16 or more years of their life developing an education that will enable them to do magic in the mysterious inner world of computer science, we human interface people spend our time learning about people and how to develop designs that will enhance their interactions with machines. A small investment in analysis, rapid-prototyping, iterative design, and testing at the beginning of a project can pay major dividends at the end. Human interface professionals should not cost you time and money, they should save you time and money. +Exploding the myth of the power user+ Many engineers panic at the sight of human interface designers, knowing these misguided do-gooders will not stop until every bit of power and performance is stripped from the product. Not true. Let's look at a classic example of user-centered design for the most macho engineers in the word -- astronauts. The time: December 1985. The project: Space Station. The various pods that make up the space station are to be attached to a framework, built in nine-foot sections, that will ultimately stretch over the area of two football fields. As originally engineered, a team of two astronauts would emerge from the Space Shuttle to assemble each nine-foot section, fastening together the section's struts by tightening bolts, using various optional tools slung from rubber cords on each astronaut's arm. When tested in the underwater facility at Lockheed, it took more than 50 minutes to assemble a single nine-foot section. At $250,000 per hour, the costs of putting up two football field's worth of these things in space was prohibitive. Enter a human factors engineer. He watched the tapes of these astronauts clumsily trying to use earthbound tools in the simulated "outer space." The astronauts, their gloves pressurized to 15 pounds per square inch, could barely grip the tool they were using, suffering fatigue and cramping. The rest of the tools, meanwhile, would float away from the astronauts' bodies, becoming tangled in the struts. The astronauts, in attempting a cranking motion, ended up tangling their own bodies in the struts. It was Laurel and Hardy at a quarter-million dollars per hour. The human factors engineer started over by studying what the astronauts could do when in a pressurized space suit. He measured the natural spread of their grip, the range of their arm movement, the effect of various types of movement on their body's stability. Then he worked in partnership with the engineers to build a whole new strut system requiring no tools. The struts themselves were the same diameter as the astronauts resting grip. The ends of the struts would slide freely into waiting holes, after which the astronaut would make a single, low-effort downward movement to slide a sleeve toward the hole, locking the strut firmly into place. When the astronauts hit the task again, it took them just over three minutes to assemble the new section. These ultimate power users performed their task more than 15 times faster. The human factors engineer was Dr. Robert Glass. He's now Director of Human Factors Engineering at SunSoft. We're working now in partnership with SunSoft(TM)'s engineers, approaching future Solaris designs just as Bob approached Space Station. We want all of you along for the ride. We hope to supply you some great design aids in the future, in the form of system software tools, manuals, and guidelines, but you, as companies, are going to have to develop your own in-house expertise. Human interface engineering pays off big. It's a new year ahead. Start investing now. Until next month, my e-mail address remains, Tog@eng.sun.com. E-mail your questions, comments, suggestions, tirades, and editorials on the state of human interface. Include your name (aliases acceptable), city and state, plus company name, if you wish. He will print some in SunWorld and pass the others on to Sun Engineering. You will be heard. He does not have the time to respond personally, but rest assured that your words will not go unnoticed. ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ For information send mail to info-sunflash@Sun.COM. Subscription requests should be sent to sunflash-request@Sun.COM. Archives are on solar.nova.edu, uunet.uu.net, sunsite.unc.edu, src.doc.ic.ac.uk and ftp.adelaide.edu.au All prices, availability, and other statements relating to Sun or third party products are valid in the U.S. only. Please contact your local Sales Representative for details of pricing and product availability in your region. Descriptions of, or references to products or publications within SunFlash does not imply an endorsement of that product or publication by Sun Microsystems. John McLaughlin, SunFlash editor, flash@Sun.COM. (305) 776-7770. TRACE: To: sunflash@suntri Errors-to: owner-sunflash@suntri.east.sun.com