Wang's Integrated Products
Past, Present and Future
By Vincent Flanders, ACCESS 86 Staff
(From Access 86: The Magazine for Wang System Users, December 1986, Copyright © 1986 Data Base Publications.)
Many personal computer users were first exposed to the concept of integrated software with Lotus 1-2-3. Released in 1983, this package combined individual spreadsheet, database and graphics applications into one coherent integrated program.
Users of Wang systems have been exposed to integration in a slightly different form -- modular integration -- for many years. As opposed to combining separate applications into a single program, modular integration combines several individual programs into a single environment. This Special Report will focus on the histories of four Wang integrated products -- one failure, one that never made it out the door, one success and one promise for the future.
In the late 1970s, Wang hired Amnon Waisman, a young Israeli who'd written his master's thesis on the relational database. As a member of the Alliance development team recalls, 'Amnon Waisman was the significant force behind Alliance. He was very heavy into database. At one point, he even came up with the concept, `The world can be viewed as a database. Not only that, but databases can be built upon Universal Data Items.'
"He used compressed B-trees to represent everything. If it had been up to him, he'd have changed the Wang document structure to a database. Very early on, Waisman was able to demonstrate that you could perform some really remarkable stuff with his B-trees. Harold Koplow, who was in charge of the OIS product line, became exceptionally enthralled with Waisman's concept of database. He gave Waisman the go-ahead to put this access method into just about every aspect of menuing and indexing on the OIS. The OIS 200 (Alliance) series was based totally on Waisman's database."
A former Wang employee elaborates. "One day Koplow said, `Let's build a general-purpose office system. In addition to WP and BASIC, let's use verbs.' He wanted to use the two-column interface; it had been only one column before -- Edit Document, Create Document. He changed it to Document, Visual Memory, etc., on one side. Then if you chose 'Document,' you'd get Edit, Create, etc. on the other side. So he had the object first, which was a noun, and the second column was the verb. That's what he wanted to do -- to create a new user interface with his database concept."
The Alliance was really an extension of the OIS. The core system was an OIS 140 or 145 with a 128K master processor and 64K workstations and printers. Initial Alliance software included System Security, Word Processing, Glossary by Example, Document Retrieval, Indexing (by user-specified keywords and Words in Text) and Spelling. The optional software included Visual Memory (database), Notebook, Calendar and Message System, Electronic Directory, Voice Documents (transcription and dictation) and List Processing (later dropped).
On November 3, 1981, Wang officially announced the Alliance: "To provide a total systems solution for all members of the office team, Alliance 250 features the integration of data processing, word processing, audio processing, image processing and networking. Driven by a new and innovative database technology developed by Wang, Alliance 250 is highly functional, yet easy to use, and responds to the needs of the entire office-management, staff professionals, administrative assistants and clerical support people." Interestingly, there was no mention in the press release as to what they meant by "image processing."
On March 1, 1982, Version 1.0 of the Alliance was released to Quality Assurance and alpha testing. It was scheduled to go to first-level beta testing on March 15, second-level beta testing on April 15 and first customer ship on May 15, 1982. Version 2.0 was scheduled for beta testing on May 1 with first customer ship on June 15, 1982. New features added to Version 2.0 were CP/M and BASIC.
"Alliance and tigers and bears, oh my!"
On paper, the Alliance was impressive, but serious problems plagued it from the start. Foremost was the issue of speed or, more correctly, the lack of it. An internal memo from early 1983 offers confirmation: "The speed/response time will be addressed with Release 3.0, which is due out in the field sometime in the March/April timeframe."
If Alliance software was slow on the OIS, it came to a grinding halt when ported over to the VS. It was so slow, in fact, that a home office programmer, tired of seeing the usual "Alliance Word Processing is loading" message hover onscreen for minutes, changed it to read "Alliance and tigers and bears, oh my!"
The software was delicate and bug-ridden, as well. As one analyst recalls, "This one site had incredible problems with the Alliance. I'd never seen anything like it in my years at Wang. There was this new release of software I think it was 1.E, that was the pits. Terminals would intermittently freeze if you'd move backward through menus. There were also lot of problems with the hardware. Engineers were always going out and working on Alliance systems -- it was a nightmare."
Another "bad dream" resulted from Wang's failure to bring out WP Plus with the Alliance. It's not widely known, but WP Plus was originally planned to be introduced along with the Alliance. Says one source, "If they could have brought WP Plus onto the Alliance, they'd have clearly separated the OIS customer from the Alliance customer. The OIS customer would just have regular WP; if you wanted advanced WP, database, messaging, calendar and everything else, you'd go to the Alliance. But since they weren't able to deliver WP Plus, what did you gain by buying the Alliance?"
So the Alliance bombed. Afterwards, an internal post-mortem was conducted to determine what went wrong. After detailed analysis, blame was assigned to functionality and system performance problems.
As far as functionality was concerned, the Alliance menu system was far too complicated -- the Wang menu system taken to its ultimate extreme. With so many menus and indexes, users became frustrated using the Alliance. It took too much time to go from one activity (creating a document) to another (printing it). It also gave menu-oriented systems a bad reputation with Wang Corporate.
At the performance level, speed was the main shortcoming. Since everything in the Alliance was based on compressed B-tree structure, every time a user went from menu to menu or index to index, the Alliance had to close the databases that were currently open, then open the others, prime all the buffers, search through several layers of indexes, expand the keys, do comparisons ... The whole operation was going on in a Z-80 with 64K of memory. As one programmer says, "We never saturated the data link, but in terms of throughput, there wasn't anything left. That's why it was so slow."
One individual involved in the Alliance project sums it up best. "Many times the team got caught up in the technology. Waisman didn't realize people would take things overboard -- totally invert documents and try to do things you weren't able to do on a Z-80 with 64K of memory. Theoretically, it was a nice idea. But what did it buy you? What did it do for the end-user except slow the system down? People got hung up on the underlying technology -- how you do it, how we get there, theoretical things. They didn't worry about things like, `How's it going to look to the end-user? What's the end-user going to be able to do with it?"'
Like most companies, Wang prefers not to talk about products that don't make it out the door. The Advanced Systems Architecture (ASA) project, later to be known as the Extended (Network) Office system, is one of them.
Unlike most unreleased products, ASA had major potential. It could have dominated the office automation market in the same manner as the IBM 360 architecture dominated the mainframe market. But due to factors unrelated to the product itself, it was never released. The story of its creation and demise (and subsequent "resurrection") depicts a rather bleak side of Wang Labs.
ASA is still such a touchy subject to many who were directly or even indirectly involved. One person refused to be interviewed for this article because "Our company does a lot of business with Wang." Another refused because "I liked the people and I liked the product. I didn't like what happened." Fortunately, other sources provided enough information to sufficiently determine what really did transpire. To protect the innocent, all quotes will be attributed to a "composite team member," the mythical Joe Smith.
"ASA came about because Wang was scared silly by the Xerox Star workstation. The Star came out in 1980/1981; it was revolutionary. From what we could see, Star could integrate diagrams, graphics and typographical word processing. Although it was never particularly successful, it had a huge influence on the computer industry. Look at the Lisa and Macintosh, look at Microsoft Windows.
"The Star caught the attention of Wang higher-ups. Dr. Wang called a meeting and asked everybody, `What'll we do? What'll we do?' One attendee got up and said, `Let's copy it exactly and sell it. We'll all be millionaires.' The Word Processing group thought they had the solution with Alliance and the V3 editor.
"The VPs went into a caucus and announced they were taking a number of people out of their current jobs to form a group to study the problem of the next-generation integrated system. There was no charter for the group and, looking back, that might have been the fatal error. This group included two vice presidents -- Harold Koplow, who had helped create the 700 series calculator and Wang Word Processing, and Bob Siegel, who was the instrumental force behind the VS. They also selected managers from different areas and put them into this group. Those `invited' were Amnon Waisman, who headed the Alliance database project; Dov Isaacs, who headed the VS office systems group; Jonathan Addleston, from the OIS group; and Bob Slezak. Later players of interest were Lee Story, Richard Drane, Thomas Greene, Deborah Mayhew, Stan Curtis and Bill Dressell.
"The first thing the group did was look around at what was already out there. We bought different products, stuff from Three Rivers, Apollo, other folks. We got an in-depth look at the Xerox Star before many others did. We looked at high-resolution screens and different processors. We sat back and looked at how the Fortune 1000 `did' office automation. Ultimately, we came up with ASA. Our architecture came down to two major components: the Office Processor, which was the desktop workstation, and the Node Processor."
The general goals for ASA were outlined by June 1982, according to Smith. "The advanced Office Processor was based on a 16MHz 68010 processor, which was created especially for Wang by Motorola. We were going to use it throughout the system because we needed all the power we could get. The workstation was going to be a touch-sensitive, 15-inch monochrome workstation with a portrait, rather than a landscape, orientation. The goal was to be able to see one, full-page onscreen at a time. The resolution, both horizontally and vertically, was 100 pixels per inch or something like a resolution of 860 x 1,024. The screen prototype we made looked better than the Sun or Apollo, which are available today. The workstation would contain 1Mb or 2Mb of memory on the motherboard
As compared to current systems, the following list of ASA's applications may seem "quaint." But note this list was proposed in mid-1982 -- before Apple's Lisa was announced, while IBM's PC was just beginning to make inroads and during the time period the Wang 128K PC was released:
• Advanced Word Processing - integrated text, voice, graphics, spelling and proofreading, facsimile, database query/input/reporting (forms) and visual calculation (similar to that offered by then-popular VisiCalc).
• In-Basket and Out-Basket -- for messages and electronic mail.
• Calendar Management -- the same type as available with the Alliance.
• Electronic Filing and Retrieval.
• Electronic Rolodex/Notebooks -- again, along the same lines as the Alliance.
• Interactive Data Management.
• Telephone -- including DVX capabilities.
• Work Scheduling -- triggered system activities, messages, batch jobs, etc., on the basis of predefined times and conditions.
• Data Computation -- batch and large-scale data processing.
• Program Development.
Other functions scheduled for incorporation included a "Camera Option" (which ended up as the PIC or Professional Image Computer) and Audio Option (July 1983); Twisted Pair Link via PBX and Half-Page Office Processor (January 1984); and Double Page Office Processor (19- or 20-inch monochrome monitor), Full Page Color Office Processor and WangNet Link (July 1984). According to Smith, the marketing people assigned to the group felt Wang could sell both the workstation and WangNet connection for $5,000 and still make money.
"The Node Processor was essentially the file server," says Smith. "It was a not-quite-infinitely expandable set of processors; you could add more processors as you needed them. Fault tolerance would be an option. In terms of how data was addressed, you could hook multiple node processors together with high-speed WangNet or fiber optics. Essentially, the entire Wang Labs complex would appear as one gigantic system."
Figures 1 and 2 show ASA screens. On the top row are icons representing a calendar, telephone, mailbox, blank forms book, file cabinet and calculator, which were system-defined facilities. The "AR Invoice" and "AP Check" represented specific user-defined facilities for Accounts Receivable and Accounts Payable. The shaded area was available for one additional activity. Certain functions, such as the mailbox, passively monitored incoming messages, and communications via audible signals and changes in the icons.
After the initial enthusiasm, interest in the project dwindled by Christmas 1982. "Things were nice and quiet between then and March 1983," says Smith. "Nobody expected anything. Our deadline for an internal demo was March 1983. As it turned out, Dr. Wang was scheduled to visit the Orient then. The day before he left, we invited him up -- we invited anyone up -- to see the demo. This demo had windows, applications, a crude editor and a 928 board hooked on to the system. We had the prototype screens. We had it; we were able to display a whole page of document onscreen in less than one second. When Dr. Wang left the demo, he was smiling and nodding to everybody. If it had been 1980, that would have meant you were in heaven, you could get everything you wanted to make sure it got out the door. But not then. Things were different."
So what happened? Why didn't it get out? Recounts Smith, "Fred Wang had come into power around 1981/1982. The Doctor was giving Fred more discretion. He was beginning to make more and more decisions about what was going to be a product and what wasn't. The Doctor stood in the background.
"Fred wanted `six-month solutions.' Fred's attitude toward the project was `You're not willing to commit yourself to give me the product in six months.' We told him, `Fred, it's not a six-month project. There's never been a new system done in six months here at Wang Labs.
Neither the initial word processor, nor the VS, nor the 2200 was done in six months. Some people like to glamorize about how they really worked hard over a few evenings and got a product out, but it never happened that way.' Fred didn't want to believe it. He thought it was a lack of commitment on our part.
"And there was another problem. Several VPs saw ASA as a threat to them. In fact, it was a big threat because it didn't come out of their group. The level of petty politics involved here had never been seen before. The only thing these vice presidents could agree on was to kill ASA. Kill it before it spread and took over. There were some meetings between Koplow, Siegel, Fred and other VPs; a lot of yelling and screaming went on. Eventually, Koplow was taken off the project and placed in charge of looking for PBX companies to buy. We saw his removal as the end. The project wasn't killed, we just didn't get the resources we needed. A lot of the ASA group got fed up and left the company.
"Looking back, the demo really was the beginning of the end for the project. People saw we could deliver something exciting -- something they weren't a part of. They put a lot of pressure on us to change it to fit what their groups were doing. Executives kept asking, `How does this fit in with the OIS 300? Why don't you change it to an 8086? Why don't you put it on the PC?' They withheld resources. We just gave up.
"ASA was scheduled to be released by February 1985. We were on schedule -- something unheard of at Wang. ASA would have been a tremendous success. If Wang could somehow come out with the ASA today, this very minute, they could still turn things around."
(As it turns out, Wang might actually agree with this prognosis. Like Lazarus, it appears ASA has been brought back from the dead and is now a project called the "42X," which will be discussed later in this Special Report.)
By 1983, Wang's internal communications systems were a shambles. According to an in-house study, Wang had over 225 Mailway installations, 180 VS EXPRESS sites, plus a heavy reliance on expensive mail services. In May 1983, Wang initiated a project to take care of their mailroom problem.
"Wang Labs' internal mail system was terribly slow," says one source. "Somehow they wanted to eliminate all the mail flow within the company or at least dramatically cut it back, so they conducted research on document flow. That project merged with a few other ideas, including an underground project by a marketing support group that had to do with putting office automation projects on top of the VS.
"On October 4, 1983, they announced Wang OFFICE. Well, it's probably more accurate to say they announced a bunch of preconceived notions. After the announcement, they put a group of people together to implement these notions. The implementation was after the announcement -- which I know is par for the course, but OFFICE was the worst example in the company's history. They had some screens they could show you, but it really was a Hollywood facade. The people who were `assigned' to the project were incensed -- the first they knew of it was after the public announcement."
What exactly is Wang OFFICE?
"Wang OFFICE," according to a Wang spokesperson, "is a comprehensive set of communications, data-processing and office automation applications that allows users to perform virtually all business information-processing tasks from a single computer terminal. Through its standard features, such as mail and messaging, directory services and time management, users can realize significant savings in time and money by accomplishing many common office tasks electronically. Users can exchange messages, memos and personal phone books."
Wang's public and private announcements about OFFICE indicate it is "the primary solution within which Wang will integrate all office information applications for delivery to the user's desktop." In a report just prior to rollout in 1985, Wang stated, "Wang OFFICE will form the basis for integrating future products, enhancements and new emerging technologies such as voice recognition."
Wang recently announced Wang OFFICE Message Notification (part of Wang Integrated Office System), Directory Integration, Wang OFFICE/Voice Mail and Wang Telephone Message Exchange (see News, page 54).
Wang OFFICE is an evolving product. Both vendors and users criticized the initial release's lack of features, but Release 2.0, announced May 29, 1986 and now shipping, offers new features and options including:
• Office Indexer -- lets users search for document files across libraries based on author, title, date created or word-in-text search criteria.
• IBM communications gateways -- allows users to file and retrieve documents and exchange messages with IBM Professional Office Systems and Distributed Office Support System users.
• Computer-based training -- provided online.
• Asynchronous Traveling User facility -- provides users with access to their home system with any asynchronous device that supports teletypewriter (TTY) scroll mode; users can send and receive mail and messages, display their calendars, and create and manage distribution lists.
• Application Management Utility -- provides menu bypass capability and the ability for administrators to custom design menus without programming.
• WP Plus integration -- WP Plus, Wang's premier word-processing program, is integrated with OFFICE so that an alert notifies users when mail has been delivered; with two keystrokes, users can temporarily exit word processing and enter OFFICE and, upon completion of any OFFICE tasks, return with one keystroke to the exact position from which they exited WP Plus.
• Application Program Interfaces (API) -- programming development tools that allow end-users and independent software companies to integrate their software applications with Wang OFFICE and produce customized office information systems. Although the feature most overlooked by users, it's probably the most important one offered. The goal of API is universal connectivity. Coupled with VS Data Exchange, applications can be written to send data from any PC source up to the VS, where it can be translated into VS format and inserted into VS applications. Significant developments in this area are expected within the next six to nine months.
It appears the ASA project is still alive at Wang: On the eighth floor of Tower Three, 65 engineers and programmers and involved in Wang's next-generation integrated product.
Like most projects-in-the-works, it has a code name -- the "42X." Because it's a future product (the final link scheduled for 1989 release), there's little outward enthusiasm among the ranks. They've seen too many similar projects dispatched. As one person familiar with the "42X" says, "In the past, every time we tried to create this product line we had to fight our VPs. This time, to avoid a similar occurrence, top management decreed that `there would be no vice presidents involved.' So, all the developers were gathered up, put into another Tower and made to work together.
"They've been working on it for about a year now. If you want to build a really revolutionary product that's also good, you have to spend a lot of time."
Although scheduled for internal demonstration on December 12, 1986, exact details about the "42X" are predictably scarce. Says one source, "Phase 1 is to ship it out as a bitmapped, 68020-based VS workstation. That model is called the `421.' They'll get it into the hands of the developers and by March, Wang will have a super spreadsheet that's downwardly compatible with Lotus 1-2-3. Phase 2, the `422,' will be released in 1988. It will offer desktop publishing-quality word processing. There will be onscreen proportional spacing, multiple columns and true fonts. Phase 3 is scheduled for 1989. It's called the `423' and will be a standalone system.
"The hardware's ready, the operating system's ready (code named Osprey, a subset of the VS operating system) and the tools are ready. They're going to hook it up as a VS workstation with the ability to go back and forth between applications. If you're in WP, you can jump to your graphs, edit them, exit and be back in WP. Nowadays, you have to hit CANCEL, CANCEL, CANCEL, etc. It will also be multi-tasking, which is something they have to have or they can't compete. It's a very difficult task, but they've got the best developers on it, including some of the original VS people, some originals from the OIS group and the former head of the PC software group. It's another example of the best and the brightest."
When the previous descriptions were read to a member of the original ASA team, he replied, "It sounds like they've resurrected the ASA project. Of course, it's four years too late. I don't understand what they're going to do with it. It's funny they're using `Osprey' as a code name -- it was the code name for a part of the ASA operating system kernel. At the time, it had nothing to do with the VS operating system. Why would they run a VS operating system in a 68020-based workstation?"
Wang has been concerned with the importance of integrated software solutions since 1979. As we've seen, there has been one failure, the Alliance; one that never made it out the door, ASA; one reasonable success, Wang OFFICE; and one promise for the future, the 42X.
We've also seen an integration of another kind, created from the inventories of Wang's mistakes: Portions of the Alliance survive, incorporated into Wang OFFICE and other VS products, and the ASA project lives, destined to finally make it out the door as the 42X.
Historically, OFFICE is Wang's biggest success in the area of integrated software solutions. Although set back with its premature announcement, there was never any real doubt about its future. It's a strategic product, providing the basis for the development of future product solutions and embodying Wang's strategy of openness, coexistence and flexibility.