MHJohnson@HI-MULTICS.ARPA (Mark Johnson) (12/17/85)
Having written my trip report, I have gotten a few impressions together about DECUS & the DEXPO exhibits down the street. These include: 1) The 68mbyte 8650 was very fast with the forty or fifty users that were on the system. It was running VMS 4.3 which has fixed the infamous logical name table bug reported a month ago. It didn't fix a few other bugs I found, so I got to resubmit some SPR's on software I haven't received yet. 2) There were a few `non-products' in existance including a color VAXstation II (not the 520), a diskless PRO 380, & a PRO 380 w/ the new 20mbyte half height drive. 3) Some of the MicroVAX talks indicates that the uVAX II w/ KDA50 controller & RA81 disks is almost 97% of a 780 (running a finite element benchmark). A number of users also complained about the excessive cost of the one packaged system w/ that controller & disk. We noted the performance gains & bought another company's controller & an Eagle instead. 4) VMS 4.3 is mostly bug fixes and will be out next month. Major enhancements include 8650 support & a number of improvements in the VAXstations software including the programmable interface & resizable windows. It was sort of cute to enter "SET TERM/PAGE=66" and see the window grow to full height. 5) VMS 4.4 will be improvements (upgrade time again...) and be out four months later. It is providing 802.3 support in DECNET, cluster node names, FAL performance improvements, descending keys in RMS, DCL subroutines, & a number of cluster performance improvements. There was also supposed to be a `VMS Development News' session, but it was cancelled. 6) The changes being made to support the ISO network protocols seem to be making a broad impact on how DEC is implementing networks. Several speakers emphasised that existing systems & software will continue to be supported but that all new products will move to the new protocols. Things to look for in '86 include 802.3 support, X.400 message system implementation, more support for MAP, and some X.25 improvements. 7) The magic session was excellent but omitted the UPG demo due to some preparation problems. A new and improved (even videos!) UPG presentation was promised for the Spring '86 DECUS. The Advanced Q&A was horrible due to a very cumbersome connection between Anaheim & the developers in Massachussets. I hope DEC sends out more responsible people to following DECUS meetings to help resolve this problem. 8) DEXPO was worth a couple of hours to see the sales people & their equipment. Systems Industries was very pushy & their system crashed while the salesman was showing off the disk cache box. Most of the other sales people were quite reasonable with a lot of good information. For me, I thought that Interleaf (documentation system running on a uVAX-II) & RTI (INGRES version 4) had the best stuff for our needs today. Overall, most of the sessions were very good. Only a couple clinkers in the week for me. The food was AWFUL at lunch & was the worst part about the whole week. Perhaps the next few symposia could have a `birds of a feather' meeting for INFO-VAX participants. It was interesting to meet some of the people who occasionally put in messages to see what they are REALLY like. --Mark <MHJohnson @ HI-MULTICS>