Some of you know I'm involved in converting a large system from mainframe to web technology. I've noticed in other threads some comments saying one of the benefits of fixing the Y2K bug often is replacement of "obsolete" hardware and software with "state-of-the-art" stuff. Is this really a benefit? Our mainframe is an IBM 4341, early 80's technology, a descendant of the IBM/360 going back to the early 60's. We run the VSE/SP operating system, 12 years old. Funny thing about this old configuration is that it is rock-solid. Hardware failures are very rare, even considering the age of the equipment. The operating system, even being 12 years old, has a 25 year history of being tweeked and tuned, making it very stable. Basically the only time we reboot the machine is after a power failure.

Now lets take a look at windows technology. Two days ago another Y2K bug was found in Win/98, although considered minor. 98 is also famous for the problems is caused when users tried to upgrade from older versions. Win/NT 4 now has 4 service packs, some introducing new bugs. I keep hearing story after story of constantly rebooting hung servers. When Win/2000 does ship, it is sure to introduce a whole bunch of new problems. So far each new release of Visual/J has required most programs to be modified in order to continue functioning.

So, I ask you, is new technology really better? Do you think it will EVER be as stable as the tried and proven big iron, considering the pace of change? Any other comments?

<:)=

-- Sysman (y2kboard@yahoo.com), February 19, 1999

Answers

I think you're referring to a comment I made.

The comment was not made in reference to hardware, but to the applications themselves. Personally, I think the *best* server will eventually be the mainframes.

Although I now work with client-server, I had (and to some extent still have) the same reservations, especially around backups, security, scheduling, etc. But consider, those "big-iron" systems weren't rock-solid out of the box. The features of those systems grew out of necessity, and I suspect the same will be true of the client-server systems as well.

AG

-- (AG@BFI.com), February 19, 1999.

Sysman,

Sent a private message to your yahoo address this morning - hope you received it.

I agree with you - the key is system evolution - It takes years to stabilize hardware and operating systems. I also worked on a 4341 in the early 70's and had great respect for the stability of that system. I got into LAN technology during it's infancy and really enjoyed the constant danger and excitement involved in keeping those systems up and running, but the fun wears off after a while.

The older I get, the behinder I get, but the principles are all the same, no matter the platform.

"We might as well be walking on the sun" - Jim Morrison (I think).

"If I woke up this morning with my head sewn to carpet, I couldn't be more surprised." - Clark W. Grizwald.

Mike Cumbie.

-- Michael H. Cumbie (Mikecumbie@yahoo.com), February 19, 1999.

Sysman - I hear ya, and I agree. Me... while I have never been a fan of big blue, but of the DEC line myself, understand all to well. I know that every single application which I have designed and written in the past 21 years is already Y2K. Designed and programmed that way. One good thing about the operating system which I work on. :) It is just a phase (hopefully), and will go back around. I see the swing to windows and various Unix flavors to be the "in" thing to do with companies. While granted, windows (and some of it's apps) are good for things like word documents, I do not think it was meant to run a business. I know that I would not base MY business upon it. Unix is still growing and is very hot right now, but I and others think that the swing will return back to our ways. Gotta get back to moving some code into production.

-- (cannot-say@this.time), February 19, 1999.

As with life, new ain't always better, and dedicated followers of fashion rarely have much between their ears.

It can be lots better, provided the person making the decisions on what to replace (and what to replace it with) is competent. If it's a phm, then most probably all he "knows" is that Microsoft is wonderful and that everything else is obsolete... and that what he says, others do (unless they see what's coming and bale out)

So come Y2K, the new system won't be working and the old system will die. Particularly sad if the old system was reliable and easily remediable but the phm thought upgrading to the flavour of the month was preferable! The legend doesn't record whether the emperor burnt his old clothes...

-- Nigel Arnot (nra@maxwell.ph.kcl.ac.uk), February 19, 1999.

Well, the new hardware is a lot cheaper, runs faster, doesn't require so much cooling, it's a lot smaller. And the new software tries to do a lot more (whether you need all the functionality or not is another question). Programmers aren't any smarter, of course, though there have been some improvements in programming techniques. But open systems are an order of magnitude harder to work with.

So long as the old stuff does everything you need, why not stick with it?

-- Flint (flintc@mindspring.com), February 19, 1999.

All depends on what you want the system to do. (And that is from a guy who thinks VM is the best OS ever!) If you want a tightly controlled central data processing environment with users limited as to what they can do, then you want a mainframe with terminals. If you want the users turned loose to follow their fancy, you hook up to a server for shared databases and files, email and internet access, get the best workstations you can afford, load all programs locally (if possible) and turn the users loose. In one case the IM dept has total control, in the other you just provide user support. And you can strike a balance anywhere between the two extremes. So 'better' all depends on what you are trying to do.

-- Paul Davis (davisp1953@yahoo.com), February 19, 1999.

I'm pretty sure no businesses are converting their mainframes to Win98 :-)

Big Unix servers, that's another matter. Or web as the front end, which could still use a mainframe in back--get rid of dumb terminals, use the Internet (with encryption) and you're world-wide at low cost. Relational databases in place of old hierarchial-type gives you vastly more flexibility, admittedly at some cost in performance. A straight conversion with no changes may not seem a major benefit, but the big win is on development cost savings with future changes. And there are ways of getting around even NT's reliability problems, with redundant servers and clustering.

-- Shimrod (shimrod@lycosmail.com), February 19, 1999.

I agree with most posters here. Unfortunately, if the enterprise systems go down hard (read: IRS? some of Fortune 500? banking?) and we are forced to island IT for next few years because the cross-system interfaces don't work, the technical and economic recovery could be forced onto small PC/LAN environments. Windows2003? Yuck.

-- BigDog (BigDog@duffer.com), February 19, 1999.

canis major -

I'd be willing to bet that a fair number of those "islands" will be UNIX-based. Sun, HP, and even Apollo (pre-merger) were pushing their way into this market when Microsoft was still focussed on owning the desktop, and there have been too many horror stories about NT migrations for every company that needs to move to just jump onto the Bellevue bandwagon. Lots of mature tools in the UNIX environments now, along with experience and stability. May be a major opportunity for UNIX to make inroads at the enterprise systems level.

I still can't get over the fact that Linux is reportedly making some headway at the desktop. Anyone who predicted that five years ago (and I know a couple of 'em) got laughed at. Kinda like Y2K...

-- Mac (sneak@lurk.com), February 19, 1999.

Mac --- Agree about Linux, actually, and it is a serious point to make. In a hard down scenario, Linux and Linux community exactly the right kind of system/people to rebuild a shattered infrastructure.

-- BigDog (BigDog@duffer.com), February 19, 1999.

Does my handle give you a hint of my opinion? :-)
All my experience has been on PCs (with exception of HP series 80 back in early 1980s). First DOS, then Windows. I remember all the complaints about the "complexity" and "unfriendliness" of DOS. And the hype at the time was that Windows would bring about a new era of "friendly, intuitive" personal (and business) computing. Intuitive my ass. Intuitive like riding a bicycle -- that is, intuitive after going through a (often painful) "learning experience."

Buggy, unstable -- if this is an improvement over DOS, I guess I'm a retard. (Remember, DOS had memory management stuff from outfits like Quarterback, and graphical interfaces like GEM. Plus the so-called "toy" language BASIC had graphical functions -- could draw all the stock market charts you wanted, etc.

I'm ready to retire (and maybe buy a Mac?)

-- vbProg (vbProg@MicroSoftSucks.com), February 19, 1999.

Hey VB, I got an old mainframe you can have, if you come haul it away! (grin). <:)=

-- Sysman (y2kboard@yahoo.com), February 19, 1999.

Sysman:
I'd love it. I understand they make great space heaters (if you can get the juice).   :-)
By the way, I meant "Quarterdeck" above not Quarterback.

-- vbProg (vbProg@MicroSoftSucks.com), February 19, 1999.

Hi VB, I've always loved QEMM. Life was so much easier in the DOS days. Much more stable, easier even for us ASSEMBLY freaks.

Thank you all for your remarks. Got to go for now, but I'll check back here when I get home. <:)=

-- Sysman (y2kboard@yahoo.com), February 19, 1999.

Re: Linux

Japan has always lagged in software and Intellectual Property Rights but those days are changing - and fast. The major hardware manufacturers hired some sharp kids (and many have been hiring only systems and software grads for two years now) and the work they're doing now may throw Silicon Valley into a tailspin. I expect it to be similar to to the impact Japanese cars had on the U.S. auto industry in the 70's. Quality is the key. Look at the makeup of the auto market now and transfer that image to the software market of the future.

The U.S. software dominance and free ride on atrocious quality and instability is almost over. The programmer community and end users will react just they did to Japanese cars (my projection only). It's going to be an exciting time.

Linux is getting alot of attention throughout Asia and the current work here on slim, fast and clean GUI's for Linux will shock some sleepy people in the U.S. Linux was the window of opportunity to battle Windows and NT. The opportunity was and is being taken.

-- PNG (png@gol.com), February 19, 1999.

Good point that the Linux community has exactly the right sort of people to rebuild. A whole community of people who give away their software, just for the satisfaction of a job well done! (And the esteem of their peers.) There are a couple essays out there by Eric S. Raymond that provide a kind of anthropological analysis of the open-source culture. They're well worth reading:

http://www.tuxedo.org/~esr /writings/

Click on The Cathedral and the Bazaar, and Homesteading the Noosphere. Or, for a quick intro, How to Become a Hacker.

In any case, I'm sure they'd all be thrilled to get the whole country running on solar-powered Linux servers. Makes me want to learn Linux....if only I wasn't spending all my free time learning wilderness survival. Ahh, well, if the grid and the networks survive, maybe I can get a late start...

-- Shimrod (shimrod@lycosmail.com), February 20, 1999.