Wang VS

The Wang VS Community

Ashley Adlam wrote:

> We've all spent too many years supporting the venerable VS to
> just cease any association with its legacy in our later years! Why,
> it has shaped our very psyches - and that actually explains a
> bloody lot. Seriously though, keeping in touch this way may help
> to alleviate your sense of loss from not getting any more of those
> challenging and enjoyable band printer calls. Bring back the 5573,
> 5574, 5574-1, and of course the mighty 5575 (arrrrgh!). And
> remember when some customers replaced the standard plastic
> Daisy Wheel font disks with metal ones so that every fourth letter
> was smudged? Ahh, the good old days! Actually, wait a minute,
> they weren't really that....

This touches on something significant. Yes, I think many of us were indeed shaped by the VS. I credit Wang with having created the best, most efficient, easiest to use business data processing environment the world has ever seen. It converted me from a minicomputer, command-line person to a block-mode, named parameter person, and I have never looked back. PACE got me used to high-quality database features such that in 1989 when I became involved with Oracle 5.1, I was bitterly disappointed by the lack of features and poor quality of utilities in Oracle. I heard just a few years ago about a shop that uses the VS and Oracle and sends their database guy to Oracle seminars every once in a while. Each time he returns saying about Oracle... "Well, it's ALMOST there..."

My memories of 25 years working at various VS sites do not include difficult recollections of any hardware problems worse than anything else I've encountered in my career. If anything, with the exception of printers, most of the Wang hardware was built like tanks and was exceptionally reliable. Printers, being highly mechanical, gave more problems than most other equipment, but nothing worse than I've experienced with other vendors and often better. And in any case, one could lay the responsibility for most Wang printers at the doorstep of Data Products, as Wang didn't manufacture any but a few of the VS printers.

For me the VS has been a totality of experience, a gestalt, an IT worldview from which no individual components can meaningfully be separated. I think it is for this reason that conversions and substitutes, all of which fail to duplicate some aspects of the VS, always fall short. I had occasion to consult on a project for a conversion house, a followup to their Y2K remediation. Studying their proposal for conversion to a Unix system I was struck by the large number of footnotes... it seems that minor feature after minor feature would not be carried forward in the Unix conversion. I realized that while no one feature was critical, the totality of the effect was a death of a thousand cuts -- that the final result would be nothing like the VS and would lose the magic that made the application excellent on the VS.

My thinking and my expectations have been substantially shaped by the VS. I don't mind admitting that even though it has set me up for many disappointments with non-VS systems and environments and products to which I have been exposed. Wang, for all its faults, was the source of an exceptionally well-done data processing platform and environment and many component technologies. If that means I have to be bitterly disappointed by almost everything else I encounter, so be it. If that means that I have had to devote the latter part of my life to bringing the VS back to life, so be it. I only wish that the New VS had been possible earlier, that it could have saved even more VS customers from "life after Wang," which has almost never been good.

I am happy being an unashamed, unabashed, enthusiastic devotee of the VS. I am unafraid to point out that GUI is a bunch of unnecessary crap in the heart of almost any business, where characters are still the business of business. I am unafraid to point out that databases, while neat and cute and sometimes very useful, are horribly inefficient and often unnecessary to the conduct of the back office business of business. I don't hesitate to point out that client/server was a non-solution to the wrong problem, that it only complicated design and programming and implementation, all of which it was supposed to improve. I readily point out that application portability is a false objective, that almost any application, once hosted on a platform, can happily remain on that platform as long as the platform is viable, and that no one in his right mind would "replatform" an application without a damned good reason.

One of the first VS systems, the VS80, released in 1977, served up to 32 users and countless peripheral devices in a system maximum of 512 KB -- half a megabyte. That is so much more efficient than almost anything seen since that it's difficult to make any meaningful comparisons. Today a PC can't serve even one person without hundreds of megabytes of memory, while a modern New VS can serve many hundreds of users with that same amount of VS memory.

The VS systems at most of my clients, companies ranging into the half-billion-dollar per year size, have had no more than 15-20 GB of storage handling all their business functions. A single disk of that size would today be considered too small for a single PC serving a single user.

The VS CPUs, never objectively very fast, the reason we are now able to virtualize them at several times the peformance of the fastest legacy VS ever released, were nonetheless able to comfortably serve tens, scores, hundreds of users. The fastest legacy VS, the 1999 VS18950, had a clock speed of 58 MHz. No PC user today would be caught dead with a PC running only at 58 MHz, yet the VS18950 was and is able to support 500 users in real world environments. The New VS today can provide performance 2.4 times greater, and with the new Intel QPI machines, probably over three times greater. The era of VS systems capable of supporting 1-2,000 users is upon us.

So... I agree that the VS has shaped us. I admit it. I embrace it. I celebrate it. Long live the VS and rational, sensible, efficient data processing.

Thomas Junker

Views: 23

Comment

You need to be a member of Wang VS to add comments!

Join Wang VS

Comment by Ashley Adlam on September 6, 2009 at 2:11am
A great addition to this ongoing theme, Warwick. Now that you have steered the emphasis in this direction I'm going to shamelessly recycle the latest comment I had made to Ronan, into this blog as it does relate to what you have noted. Ronan touched on the fact that many of the older devices were interesting in that when you opened them up there was more variety in the way they were designed and worked - sure not always that good, but often quite different from each other.
Indeed, when we started working in computing, devices, just like the people, where allowed to be different and eccentric. As Ronan alluded to, we'll soon have to be visiting retirement homes if we want to reminisce about such things. Of course younger people in the industry have different expectations and computers have lost their mystique - they are now everywhere (every house has TVs, toaster, PCs, - no big deal), but ironically, there are definitely less opportunities to be creative at work - to come up with new and novel ways of doing things on the job, where colleagues would say, "Wow, that's clever - how did you do that?".
Anyone doing things differently today would more likely be in trouble for employing a non-compliant methodology. And you can't arbitrarily blame the employer. Customers, just like the major corporations, may have shed most of the workers but they’ve retained the lawyers. In today's distorted legal/ethical ethos, “non-compliant” sounds inherently bad – worse than murder, although not as bad as perceived racism.
Certification is "training" not education but it can go a long way to effecting a general compliance, and every certified sys administrator knows the correct and standardised way to do things, which explains why it always takes two weeks to get the permissions sorted out to the point where we can actually log on and do the requested work on the customer’s site.
I have worked, and continue to work with some marvellous young IT people that are a joy to be around but increasingly I notice that perhaps even the majority are also studying for or planning to work in other areas. There is not much of a career path anymore for most IT roles. It was interesting to note that in at least one major US survey of college graduates' career satisfaction, IT had the highest "regret" score, wishing they had made a different choice. To sum up the foregoing ramblings, I think one BIG difference was that in the past we implemented "computer based systems" to make work easier, faster, and more accurate. Today, IT projects are implemented to cut costs, reduce overheads (i.e. people), to be able to off-shore, retrench, consolidate. That's where the action is if you're a CEO looking for the big bonuses! And in many business sectors a CEO who is not in on that action, is irrelevant because his/her corporation will be out of business due to the competition’s lower costs.
Comment by Warwick Halcrow on September 5, 2009 at 8:32pm
Dear Contributors

Congratulations on a most thorough statement on the last generation ( 25 years ) of cost blow-outs, corporate inefficiency and "smoke and mirrors" financing. This is most eloquently presented in the context of IT marginal and unproductive utility.
I thought for many years that I was the only person who had the above opinions and was condemned as old, out of step and an IT irrelevant. Our only crime was that we tried to protect our employers and clients from ultimately their own demise. You cannot separate the stupidity of executives and managers from their "golf buddy" experts. The more concerned and conservative we became, the more and more exotic and enticing they believed the "New Economy", service industries and Wall Street analysts was the way of the future. However little will change as the arms and legs of operational stability have been amputated while the heads of "the fast smart buck" has not been removed. There are rarely poor sailors just poor skippers!

Regards
Comment by Thomas E. Mitchell on September 3, 2009 at 3:24pm
Regarding COBOL and how easy it is to learn, I am instantly reminded of Y2K remediation and the shortage of COBOL programmers. Some companies were specifically targeting disillusioned teachers, offering to teach them COBOL *AND* pay them more than they were making teaching. Based on my experience with the newer, hotter languages I'd have to say, I seriously doubt the same thing would be practical with those languages.
Comment by Thomas Junker on September 3, 2009 at 2:19pm
Thank you for your comments, Ashley. Just my own two cents about industry standards and portability and such:

Industry Standards

"Industry standard" is largely an illusion, except perhaps for low level devices such as disk and tape drives and, to some degree, memory. System manufacturers try madly to proprietize their products and lines even in the face of alleged standards. Even memory, once somewhat standardized for a time in the age of SIMMs, is now somewhat peculiar and unique to brands and models of server such as Dell and IBM.

Platforms, while superficially similar, and while using Intel or AMD processors, are anything but interchangeable. One cannot back up, say, a Dell server and restore the files or image to, say, an IBM server -- the result will not run. Some platforms, such as the populous AS/400, are entirely proprietary and have no direct counterparts from any other vendor. And who, in his right mind, would even want to run the enterprise on Brand A one week and on Brand B the next week? Platforms, even in an age of alleged industry standards, tend to have lifetimes of many years, for a variety of reasons. The best we can look forward to today is that the platform vendor will have more or less compatible models every year or two so that the enterprise can relatively painlessly upgrade the platform(s) within its service lifetime. If anything, platform lifetimes have shortened considerably, to three to five years, while those in the VS community became accustomed in the 1980s and 90s to 100% software compatibility between VS models and potential lifetimes of 10-20 years. TVS recently replaced a VS100 with a New VS. The VS100 was released in 1979 and was superceded by the VS300 in 1985. Nonetheless, a customer was still running his enterprise on the 1979 Wang VS100, well more than two decades after the machine was new, and more than 20 years after it had been superceded.

Even in low-level devices such as disk and tape drives, or memory, any industry standards we appear to have are bypassed by the necessity to have vendor support of entire systems, or at least major subsections of entire systems. Customers want to have entire systems serviced by one vendor, often the manufacturers of the systems. Customers who succeed in securing maintenance contracts from vendors such as IBM for multivendor systems are no doubt constantly pressured to make their systems more uniform, to eliminate the odd men out in their multivendor lineups.

This is something that has come full circle since the multivendor approach gained popularity in the 1970s, when third party manufacturers began to offer considerable savings in peripheral products such as printers, OEM-compatible memory, disk drives and, eventually, storage subsystems. Those who mixed vendors learned that maintenance could become a nightmare, with customer engineers pointing the finger of blame at "foreign" equipment in a system, a common scenario in the 1970s.

Eventually the third party equipment model did prevail in many parts of the industry, culminating with substantial subsystem products and vendors such as enterprise RAID and EMC. The 1990s, however, and particularly the dawn of the new Century, saw system manufacturers counterattacking with their own substantial subsystem offerings and various tactics to implement a strategy of forcing, as much as possible, single-vendor supply and maintenance once again. Microcode loadability, pioneered by Wang as a way to offload work from the central processor and a way to fix and enhance functionality without hardware changes, has become a tool for control and exclusivity in the industry at large. IBM, largely unnoticed by anyone outside their large system customer base, has taken this so far that customers must not only purchase a processor, they must also license its use for a particular purpose, and bypassing that is not only virtually impossible due to unique microcode and other software necessary even to have access to the processor, but is also impeded by copyright and license terms. Some of this is done under the guise of availability and reliability, as when extra processors are provided but only switched into use when needed to recover from a failure, or licensed for full time use, but what is really going on is a significant change in the way hardware is regarded by vendors and customers. In at least some circles, such as IBM mainframes, hardware is coming to be more like software, requiring a license for its use. In such an environment the opportunities for third party participation and competition are severely curtailed.

The marketplace acquiesced to the obstacles to multivendor systems raised by system manufacturers and service arms, and today many customers strongly prefer to have their entire systems, or at least the core, maintained by a single vendor. The largest exception is that of sophisticated storage subsystems, where EMC has created its own market for storage products. At the low end, it is now common that printers, especially thanks to networking, may come from vendors other than the main system vendor, but printers are not completely interchangeable.

Just as our automobiles no longer have standardized headlights thanks to computerization of manufacturing and the advent of widespread custom manufacturing, our computer systems no longer have standardized components except most superficially, when viewed from a distance. Servers now all offer integrated RAID, but only from the server manufacturer. Memory is usually unique to particular models of servers, and only a large third party like Kingston or its counterparts has any hope of keeping up with peculiar memory and being able to offer equivalents. Remote control features, crucial to the deployment of the New VS, are peculiar to brands and models of server. Buying RAID disks from anyone other than the RAID vendor is inadvisable, sometimes even impossible, due to customized disk drive microcode.

So where do industry standards help IT? I'm not at all sure. The selection of a platform vendor often determines the source of most of the rest of the equipment except perhaps a storage subsystem, networking and printers. The selection of an operating system is largely illusory, too. While we have some choices we didn't have decades ago, it's not as if we can switch operating systems midstream any more easily than we can switch hardware platforms. And, um, how many varieties of Linux exist to choose from? So many, and with so many differences, that the idea of "standard" is a bad joke.

Software Portability

Software portability is another illusion. In the old days software was not portable from platform to platform primarily because the machine languages were vastly different. With the advent of compilers such as COBOL, some degree of portability became possible, but operating system and system services differences still posed formidable obstacles. With eventual widespread adoption of the Intel x86 and, now, x64 processor chips and compilers such as C and C++, programs can often be compiled for various vendors' machines, but to what point? As with platforms, who in his right mind wants to run an enterprise on one platform this week and another next week? There has to be a degree of stability and predictability in platforms and software environments, and even in this day of alleged standards, moving a complex application anywhere is a big job, even if a compatible compiler exists for the new platform.

The one place where portability has shown itself to be useful is at the Internet client. Java, a language criticized as being most suitable for mediocre programmers, has proved useful in enabling Web servers to run applets on a wide variety of Internet clients that connect to the servers. The idea, though, of using Java or any other supposedly portable language for enterprise applications, is ludicrous.

What is important in languages for the enterprise is stability, and readability, which ties into matainability. Many IT projects today are houses of cards, built with multiple platforms, languages, tools and skillsets, some of which become obsolete even before project completion. Wang VS IT projects, in contrast, have been charaterized by long term stability of platform, languages, tools and skillsets, allowing many VS applications to enjoy useful lifetimes of 10, 20, even 30 years. And unlike IBM mainframes, which are frightfully expensive to acquire and operate, Wang VS systems have been extremely economical to operate, often achieving operating costs only 10-20% of their counterparts.

As a software designer and programmer, I have little interest in software portability at the enterprise level. I want stability in the platform and tools, and stability in the personnel working in IT. The Wang VS has provided stability since it was introduced in 1977, and through a record-holding succession of 100% compatible models. The New VS virtualizes the stable VS environment and allows it to run on modern platforms without ever having to deal with large scale rewrites or modifications to stay on current hardware platforms.

A Word About COBOL

COBOL is not going away, the wishes of PC and Unix weenies notwithstanding. C, C++ and Java and related newish tools are horrible for business data processing. COBOL is excellent for business data processing. COBOL is also easy to learn without courses or extensive collaboration. The newer languages are difficult to learn and almost impossible to use without recourse to collaboration, due to the vast amount of nonintuitive, nonobvious features and techniques they involve.

The idea that it is difficult to find anyone to program in COBOL today is absurd. COBOL is from a time when languges were straightforward enough that only a manual and access to a system were necessary for any of us to learn the language without any help. Too, if employment today in IT is difficult, those seeking jobs should be willing to do whatever is necessary to win a paycheck. It's supremely inconsistent to say that programming jobs are difficult to find today but that it is also difficult to get anyone to program in COBOL.

Maintainability is key to the long term life of enterprise applications. I have found it generally much easier to maintain COBOL code written by others than code in just about any other language.

Another Word About GUI

GUI (Graphical User Interface) is great for web browsers, and somewhat great for the desktop. It has little use in the back office where the business of business is done. One does not keep customer records using GUI, or colors, or pictures. Nor does one keep inventory or sales orders using those things. The business of business is characters... letters, digits, punctuation and special characters.

The Wang VS and COBOL deal exceptionally well with the world of characters. All the allegedly unique features of GUI such as checkboxes, dropdown lists and such have counterparts in the character world. Wang PACE, still a formidable application development tool, easily does picklists, which are the equivalent of dropdown lists, and checkboxes as simple one-char fields. I have programmed popup picklists in COBOL and BASIC.

For those who insist on seeing their Wang VS / New VS user interface in a fancier form, there is Clearview, a middleware product that supports the VS. At its simplest level it allows the user interface to be customized to use visual controls and appear to be GUI. At its intermediate level it allows the host (the VS or New V) to interject visual controls into the user interface to have greater control over the interface. At its most complex level Clearview offers the ability to combine data from multiple host platforms into unified user interfaces at the client.

Also, an alternative Visual Basic New VS client is on a track to become available. It was developed by our largest New VS customer and made available to TVS. We are presently evaluating it.

Further, changes have already been made to support very large workstation I/O in the New VS. The ultimate objective in this is to support our own GUI enhancement in a VS workstation product.

I'm not a dinosaur against GUI in general. I just insist on keeping a clear view of where GUI is appropriate, especially in the face of criticism of the VS that it is old fashioned and doesn't directly support GUI. Those of us who have been around for decades are painfully aware that most GUI data entry and retrieval today is horribly inefficient, involving frequent hand movement between keyboard and mouse, and that most GUI screens fail to make the best use of tab order and default selection. Data entry and retrieval in the days of the Wang 2200 and later the Wang VS were much, much faster. This is part of why I say that GUI is misplaced in the back office, where all the essential data in most businesses is comprised simply of characters. To criticize the VS for an inapplicable failing is just silly.

Stability of Personnel

One thing we cannot solve with the New VS is the high rate of turnover common in today's IT world. Turnover depletes the knowledge that is resident in the staff. In today's fast moving world in which very little is documented, and what documentation is produces is almost instantly obsolete and even false, the knowledge and experience in the people who comprise IT is more vital than it has ever been. Reducing turnover and ensuring stability of IT staff is vital to the long-term survival and success of IT operations.

Since the 1960s companies have increasingly treated their people as expendable. Employees, in turn, have responded by showing decreasing levels of loyalty to their employers. In this, companies have reaped what they sowed. I can only think that this has been an evolutionary change as each generation of management has replaced the previous one. In many aspects of human life time scales have shortened, and in business the short term has almost completely replaced the long term in planning and execution. IT has been adversely affected by this.

I place some of the blame at the door of Wall Street and the investment community. Today's public corporations have been steered strongly in the direction of satisying Wall Street's concern about the current quarter, with disappearing concern about the long term. Company management has simply complied with the demands of the investment community, and the effects have propagated throughout company management, even to where bonuses and stock options are geared to short term results instead of long term survival.

This is a topic about which entire books could be written, and probably have. I can't do it justice here, and can only point out that the shift to short-term thinking is one of the factors that has changed the IT landscape for the worse.

In my experience stable IT staff working with stable platforms and tools can work miracles, while unstable IT staff and platforms chew up company resources without producing much in the way of lasting benefit. The New VS offers part of the answer, but unless companies learn to once again look at the medium and long term, they will be lost. I know of real-world cases in which clients of mine have spent tens of millions of dollars to replace Wang VS systems worth only a few hundred thousands of dollars, only to discard the new systems in the course of being acquired by other brain dead companies with their own inefficient IT operations.
Comment by Ashley Adlam on September 3, 2009 at 7:09am
Hello Thomas,
Marvellous blog and some profound observations on how our long associations with the VS have influenced our lives and wider outlook. The implications of that go beyond just the experience (as significant as it was) of having supported the Wang VS for some years. The VS saga has been a multidimensional, technological and social phenomenon that is not over yet, thanks to TVS. During its halcyon days the Wang VS encompassed the best of so many outstandingly innovative and novel achievements, all based on some pretty rock-solid hardware design and manufacture. And that spawned a whole range of world-class third-party software products. Of course the VS shared that limelight of innovation with a number of other notable computer manufacturers - I still have some fond memories of Burroughs Medium Systems too.
I agree with you - the VS could do just about anything and we had that amazing diversity of client enterprises to prove it. And for a couple of decades, we really could tailor the VS offering to assist companies to enhance their own distinct business processes in a way that could be genuinely unique to each site. I think having been part of that is much of what has shaped us in so many ways. But that was before "industry-standard" became the paramount attribute - might not be as clever or efficient as other innovations but it's what everyone wants and "needs". Aside from the overwhelming rate of technological change for its own sake, this evolutionary shift to emphasis on "standards" rather than purpose (I deliberately didn't use the word 'outcome'), has significantly altered what it means to work in IT. Still, as one Technology analyst recently commented, "The great thing about having standards is that there are so many to choose from".
Ashley

© 2017   Created by Thomas E. Mitchell.   Powered by

Badges  |  Report an Issue  |  Terms of Service

Hit Counter by B-25