Vendors were thoroughly under control - we had the most popular vendor and sponsor presentations ever heard, mostly less than one minute long and accompanied by useful little presents for delegates. This might have been a result of our long and impatient training over the years - the water-pistols, the heckling and the jeers - or it might have been thanks to the 200-odd, very noisy Special Prevention of Indignation Tools (SPITs) imported for the purpose by Carl Cargill and distributed to delegates right at the start.
Speakers were almost all of a high standard. There was, as usual, plenty of discussion during and following papers and BOFs followed through on various topics. Our overseas visitors were exceptional value, this year, not only giving very interesting presentations but also staying on and joining in.
During the conference over 2GB of data moved over the link, and about three quarters of that was incoming. The inbound traffic was kept down by the successful operation of an on-site Web cache that achieved a 40% hit rate despite the wide variety of interests of the delegates.
Our 1994 Conference saw the first broadcast of a conference presentation on the Internet MBone. This year two full days of the conference were broadcast for those that couldn't attend. Despite the heavy use, several conference committee members feel that we aren't making maximum use of the network at the conference. Next year, look forward to some new activities, and hopefully some ambitious demonstrations from vendors.
As usual, some presentations were classified in both `best paper' and `worst paper' categories.
In the popularity stakes for the exhibition, also, things were less clear cut this year, with the honours being shared almost evenly among SGI, DEC, and BCL. CompuWare and VUW also received some 'votes'.
In analysing the responses to the other questions, the most highly rated items this year were the organisation of the conference (averaging 4.6 on a scale from 1 to 5 - even better than last year), the conference meals, the tutorial presenters, the social events, and the overall rating for the conference (all 4.2). The least highly rated items were the interop and the BoFs (both 3.4). Note that no items rated below "appropriate" on average.
88% of the responses approved of the venue as a suitable location for the conference, the main reasons cited being the location (40% of positive responses) and the facilities provided (25% of positive responses). Of those who did not think it suitable (6 responses), half cited the location, and half cited the fact that the hotel facilities were stretched.
Finally 98% of responses said that they would (or would like to) attend next year's conference, while everybody who responded said they would recommend the conference to others.
Overall, it would appear that almost everyone enjoyed the conference, and we can look forward to trying to do even better next year, since there are some little things raised in the comments that can be addressed.
There were many very good papers presented, quite a lot of personal networking, and a few dud comments made. Of all the statements I heard last week, I'd like to reinforce a few and challenge a few.
There were two themes articulated by some last week that I believe could lead to serious tactical and strategic errors. First, that mainframes are evil, being wholly abandoned, or dead. Second, that Microsoft is evil, ignorable, or laughable.
Mainframes. Well, to start out, I have to admit that my background and employment are mainframe-biased. Although my current position is Manager of Open Systems, I have been associated with computing for a long time, since the time it meant only mainframes, and for over twenty years I have worked on mainframes, mostly as a systems programmer (one of the dreaded mainframe priesthood'). I have watched several cycles of centralisation/de-centralisation, and I can recognise the current ebb and flow of client/server for just what it is: the usual over-exuberance for a new idea that is inevitably followed by the wise and rational use of the resulting tools.
Mainframes are not evil. Just because a bus is big, won't fit in every garage, is more expensive and far less convenient than a personal auto doesn't mean that buses should never be built, owned, or used by anyone.
The honeymoon is over - the real world is finding out that there is still a place for centralised data on a mainframe. (If you don't like the m-word, then call it an enterprise server.) Oh, sure, the old hardware is dying, but that's been a constant fact of life about cars AND computers for as long as I can remember. The basic concept of critical data and function being centrally held and managed is sound. Do you keep your personal money in the bank, or scattered around your house, car, and workplace in every drawer, cupboard, and teapot because you might need some wherever you're standing? Which is more convenient? Which makes more sense? The client/server model itself demands that the server be managed in the way most appropriate to the criticality of the data and applications being served.
Mainframes are not dying. After a down year in 1993 when the world all seemed to be in love with downsizing and rightsizing, mainframes made a spectacular comeback from the dead in 1994, and are continuing their growth this year. The largest users of computing are expanding their usage of the mainframe (enterprise data server, remember?) as they make their data available more widely within the enterprise. While we are seeing a dramatic growth in client/server and an ever expanding use of data in daily decision-making at every level in most businesses, the protection and management of the data is more important than ever. It seems strange, but client/server and mainframes are BOTH growing.
Mainframes are not being abandoned. Don't mistake the normal evolution of technology to smaller, cheaper, and more powerful for a revolution from central management/serving to total de-centralisation. Client/server is a new buzzword for a very old concept, and let's grant that the information technology industry had at least some of their facts right.
Evolution and revolution are similar in result, differing only in rate of change. We can see around us in New Zealand some results of violent revolutionary changes made in some companies that didn't work.
Gartner Group, a world-wide industry watcher, puts the typical cost of an open, fully-distributed, client/server system at four times MORE expensive than the "legacy" mainframe system it is replacing. Draw conclusions with care, though. Our technology will mature and keep changing at an ever-increasing rate, and evolutionary changes will begin to resemble revolutionary ones soon.
Microsoft is not evil. Some people see goblins under the bed and in every cupboard. Microsoft is only a business, albeit a big one. They do some things very well, other things badly. They are, as nearly every speaker (grudgingly) had to admit, here to stay, so we had better make plans that reflect the reality of a dominant Microsoft desktop. Let's NOT cede the server world to them, though! We must promote the strengths that we know all too well, and put all our energies into marketing (in the generic sense) Unix.
Microsoft is not ignorable. We do not want to share the fate of the ostrich that sticks its head in the sand to hide from the approaching Microsoft steamroller (sorry for the mixed metaphor, but I couldn't resist the mental image).
Microsoft is hardly laughable, either. If we stand laughing, we will still be flattened by the steamroller! We have to run to escape, and laterally, not along the steamroller's chosen path. The Microsoft desktop is a reality; business today has accepted Microsoft Office as the defacto desktop workplace, much as the world has embraced TCP/IP and the InterNet as the defacto network. Sure there are more elegant answers, and other contenders. Let's stop wasting our energy weeping and wailing and gnashing our teeth, and buckle down to inter-operating with the real world better and cheaper than Microsoft itself. Hundreds of companies and products have made a career out of being better and easier to use on IBM systems than IBM programs, so why can't we do the same to Microsoft?
There is a fleeting window of opportunity to implement better networking and servers than Microsoft can deliver in the short term, and firmly establish Unix as the flexible, economical, sensible platform of choice for data and application servers. Don't be fooled, Microsoft WILL eventually get their act together. (Remember, 1000 monkeys on typewriters...)
In my daily life I have witnessed time and again how hard it is to maintain NT servers when compared with the network-ability and ease of Unix. Why can't we explain the business issues (dollars and cents!) involved, so that the correct business decision is obvious? We need to stop taking the moral high ground and just compete on business issues.
Peter Blake didn't rely on arguments centering on Conners' personality; his team just sailed a better race five times in a row. Forget Bill Gates' personality, just sail better races. It should be dead easy.
Board members John O'Gorman and Martin Lennon ended their two-year terms and did not stand for re-election. Noel Cheer (IBM NZ) remains for the second year of his term and Brenda Lobb (The Seabrook Group) is immediate past-President.
Kaye Batchelor (EDS) was elected President, Bruce Miller (Excalibur Consulting) takes up the position of Vice-President and Rob Pascoe (Koppens Pascoe) once again will fulfil the dual role of Secretary/Treasurer.
Elected Board members are Ray Brownrigg (VUW), John Hine (VUW), Julie Jones (Integral Technologies), David Carmine (Data General NZ), Peter Preston (Optimation NZ) and Alby Cartner (NZ Police).
Return to home page