-------------------------------------------- 1990 Winter USENIX: AT&T UNIX pc BOF Minutes -------------------------------------------- Sponsored by ICUS Software Systems Lenny Tropiano (lenny@icus.islp.ny.us) Gil Kloepfer, Jr. (...!ames!limbic!gil) Wednesday, June 24, 1989 Omni Shoreham Hotel Calvert Room The 1990 Winter USENIX conference once again was the home of the UNIX-pc BOF session. As always, our thanks are extended to those who attended the BOF for making it the technical exchange forum that it is meant to be. As is always the case, there are those who cannot make it to the conference for one reason or another, and we hope that this summary again provides you with a little insight regarding the event you missed. For those who did attend, we hope that this summary is something you can keep in a file on your computer somewhere for reference and possibly some ideas for the next BOF. What we tried to do, and succeeded to some extent in doing, was to spice-up the BOF with examples of the more obscure or unusual capabilities which can be exploited using some of the expansion hardware. Those who came to the BOF early had a unique opportunity to see what went-on "behind- the-scenes" during the BOF set-up, and took the time to ask questions about the demonstrations. In Gil's opinion, these people got the most out of the demonstrations. The BOF opened as usual with Lenny's electronic artwork on the overhead transparencies and a summary of what was to be covered during the BOF. This was: Hardware o ICUS 2nd drive/P5.1 upgrade o Bigger hard disks o WD2010 (seek performance) o Voice Power boards o Botton's VIDPAL o Custom UNIX pc cases (Dapoz design/ICUS Design) Software o "Jukebox" (Voice Power) demo o Answering Machine (Voice Power) demo o USENIX FaceSaver Display demo o MGR ("manager") demo o UNIX 3.51 FIXDISK 2.0 o Public Domain (wmgr, ph, login) o "Archived" UNIX-pc Software o Alex Crain's sockets, symbolic links Miscellaneous o UNIX-pc Technical Newsletter o UNIX-pc Reference Manuals Unfortunately, what disappointed some was the fact that there were too many things planned to be covered, and many of them did not get covered as completely as they were during the BOF in June. Lenny opened the discussion by making some brief comments regarding the status of getting the UNIX-pc kernel source code (no further along than before), how to obtain the hard disk upgrade, the success rate of the UNIX-pc hardware reference manual deal, and touched briefly on the WD2010 modification and its benefits. All of these have been discussed on the net in detail, but some had some questions and this was the time to have them answered. The first demonstration and, perhaps, the bulk of the BOF was the port of the MGR windowing system to the UNIX-pc. Lenny showed how MGR was implemented on the UNIX-pc, and the various utilities with which MGR comes packaged. Among the more familiar MGR bells-and-whistles were the X-like clock, a numeric digital clock, and a Mickey Mouse clock, as well as an icon editor facility. This, again, brought-up the proverbial X-port discussion, and those who wanted to know "why do I need MGR?" Discussions about porting X-windows to the UNIX-pc can be summarized in one sentence, "X windows can be ported to the UNIX-pc, but we need someone to write it." Nobody at the BOF volunteered to port X, but the MGR demonstration did spark some additional enthusiasm about doing it. The second question regarding the necessity to use MGR over, say, the current windowing system or a port of X windows was a bit more involved. Gil's one-line answer was, "...because someone wrote it already, and it provides a good porting base to other systems." However, this answer didn't seem to be the answer that most of the group was looking for, and further discussion continued. In general, although there was an air of apprehension about how useful MGR would be, there was a feeling of awe in seeing the speed and responsiveness of a public-domain windowing system. MGR definitely out-performed the current window driver in reacting to menu requests, window resize/move, etc. There are yet problems in the speed of the pty drivers used to communicate between MGR-windows and the operating system. During the next six months, we hope to see various incarnations of MGR in the works, as well as some work in porting X windows. During that time, Alex Crain was questioned about his socket library. He commented that there were several people using it currently, and they seemed to have little trouble to date. The MGR demonstration was followed by Gil's demonstration of some unique (and not-so-unique) applications using the AT&T Voice Power speech processor board. The demo began with a synthesized speech recital of the Pledge of Allegiance with a crudely-drawn American flag on the display. Funny as it may sound, it was an attempt at showing just what progress has been made in voice response systems over the past 7 years. The second voice response demonstration was an answering machine which used the Voice Power board as the sole form of communication over the phone lines (ie. no tape recorders were used). Simple as it sounds, the answering machine program can be a more complex application as it needs many voice response messages to communicate with the user. As one attendee noted, there were computer synthesized messages for the answering machine attendant. A reason for these messages, as Gil explained, was to keep the program complexity down. Removing the synthesized speech would have involved digitizing many digits, providing a means to parse the digits (with the proper inflection), then adding the "filler" voice response messages to both ends of the recited numbers. Gil noted that an interesting "feature" of the room where the BOF was held was that there were two hotel telephone lines. This made the answering machine demonstration a little more realistic. The last voice response demonstration was a computer- controlled jukebox. The software provided a means of storing music in digital form recorded from a home stereo system in a specialized archive format. A user-interface, which looked much like a jukebox as would be seen in a local diner, allowed the 15 megabyte archive of songs recorded on a streaming floppy tape system to be retrieved and converted back to audible sound. Of all the voice demonstrations, the answering machine was the most popular. After seeing almost 75 minutes of demonstrations, the group decided to break for 10 minutes. During the break, many of the attendees took a moment to experiment with MGR and the voice demonstrations, and the new-improved USENIX FaceSaver Display program. Following the break, Mark Dapoz discussed how he re- organized and designed the UNIX-pc cabinet improvements for his UNIX-pc. Gil and Lenny also presented their ideas for a tower cabinet, and Mark reflected on both ideas with some of his design considerations and initial failures, in the hopes that others could learn from his experience. Mark's design goal was to provide a more workable cabinet for adding additional hardware and modifications, but to leave the option to return the cabinet to its original form if necessary. He sternly recommended that those planning to add multiple hard disks also add an additional 200 watt supply. He shared a story of how he overloaded his UNIX-pc supply by attempting to power two hard drives and several expansion boards at once. He recommends that the two hard disks be powered by a separate supply, while the remainder of the stock UNIX-pc hardware be powered by the stock supply. A day or two before the BOF, AT&T officially released the long awaited FIXDISK 2.0. Lenny summarized what was included, specifically the oddities of the "metermaid" and 3-key- functions. He did mention that he would detail all of the specific changes in a future posting on USENET (which he did). Everyone seemed anxious to receive their copy and try out the new "features" of release 3.51m. Next, Lenny went on to talk about where one could get UNIX-pc Software. He mentioned several archive sites, the ICUS Archives (uucp access) run by Gil Kloepfer, osu-cis Archives (ftp/uucp access) run by Brant Cheikes, and several other ftp-able sites. Although this information is frequently posted to USENET, it was important to discuss. As always there was a small population of BOF attendees that didn't have USENET access. Finally, Lenny discussed the possibility of creating a UNIX-pc technical newsletter, to be distributed for a minimal charge on a monthly or bi-monthly basis. The idea seemed to be accepted by the those attending, however it was preferred overall that the newsletter be in electronic format. The problems with an electronic format would be the inability to include illustrations and diagrams where appropriate, and the prerequisite of receiving the network newsgroup that the newsletter was posted to. Someone mentioned getting an advertiser or some sponsorship of the newsletter, but the overall consensus was that it would detract from the newsletter, and make it harder to produce. Following the BOF, several individuals met with Gil and Lenny to exchange ideas and software. The after-the-BOF meeting continued in our hotel suite (what Gil and Lenny dubbed the "ICUS Hospitality Suite") until about 2AM, at which time everyone felt a need to get some sleep. Last, but certainly not least, special thanks go out to the USENIX association for allowing us to hold the BOF during the conference, Gil Kloepfer & Lenny Tropiano for coordinating the BOF and demonstrations, Mark Dapoz for sharing his work he did with the modified UNIX-pc case, Brad Bosch for providing beta-copies of his MGR port, Brian Botton for his work on the VIDPAL board which is making porting MGR and X-windows more of a reality, and most of all the BOF attendees, for without you there would be no reason to do this! In closing, the BOF seemed to be a success. However, it did seem like there were too many items to cover. In planning the next BOF, it would help if more people provided some input as to topics to be discussed, and a good format for discussing them. We would also appreciate any feedback you have regarding how the BOF was presented and how appropriate you felt the demonstrations were to the discussions. 8/5/90: [Corrections noted since then my email address has changed to lenny@icus.ICUS.COM]