~ / books / visual-basic-history / origins / vb-history-the-propaganda / history v1
⚠ Viewing archived v1 · 3 h ago · [back to current]

The propaganda — Microsoft pitches the future of programming (1989–1991)

The propaganda — Microsoft pitches the future of programming (1989–1991)

By late 1990 Project Thunder is feature-complete behind Microsoft's NDA boundary. The form designer works, the Embedded Basic interpreter is wired in behind the events, the runtime DLL compiles, the IDE runs on a 286 with 1 MB of RAM under Windows 3.0. The engineering arc is, for practical purposes, finished. What turns it from a Microsoft internal tool into a developer-platform shift is not more engineering — it's the marketing-and-distribution apparatus Microsoft was already eighteen months into building around it.

The marketing case for what became Visual Basic was, in Microsoft's hands, openly a Bill Gates production. Microsoft did not stumble into the launch pitch in 1991 — Gates had been telegraphing it in print since May 1989, and the launch itself was a near-verbatim delivery of a design argument he had already published. This is the section where "Microsoft as commercial entrepreneur, not inventor" lands at full strength. The technology was assembled — Cooper's form designer, Microsoft's BASIC dynasty, the Thunder team's integration. What Microsoft genuinely built itself, and what Microsoft genuinely deserves credit for, is what came next: the propaganda machinery that turned a tooling release into a wave.

May 1989 — Gates in BYTE Magazine

The clearest single piece of pre-launch propaganda is Gates' contribution to the May 1989 issue of BYTE Magazine, the BASIC 25th-anniversary issue. Dartmouth's Kemeny and Kurtz had launched the original BASIC on May 1, 1964; BYTE's editorial choice for the May 1989 cover was the language Microsoft had been mining for fifteen years. The issue was the natural forum for any "future of BASIC" argument anyone wanted to make in 1989. Gates took the slot.

In that piece, Gates laid out the design pitch that the eventual product would deliver almost verbatim:

After all, the best way to design a form is to draw the form, not write code to reproduce it. With a mouse and a palette of pre-drawn graphics images, you should be able to combine lines, boxes, and buttons interactively on a screen to design a form for a program.

Read it now and it sounds anodyne. Read it in 1989, against an industry where building a Windows program meant a C compiler, the SDK, and a thousand-line WinMain skeleton with a hand-rolled message pump, and it reads as a deliberate provocation: we are about to make this trivial. Gates was not bluffing. Microsoft already had Cooper's form designer and Embedded Basic in the building; John Fine's January 1989 proposal had become Project Thunder, and Scott Ferguson's team was eight months into the integration work. The BYTE article was a public down-payment on a product Microsoft was already most of the way through shipping.

The article is also a clean piece of editorial timing. BYTE's BASIC-anniversary issue was the only place in the trade press of 1989 where a piece arguing for "the next BASIC should let you draw the UI" would have looked unremarkable rather than speculative. Gates seeded the audience that was already paying attention to BASIC, in the issue they would all read, eighteen months before he could prove the claim.

A note on the quote's provenance. The wording above is the version reproduced across the modern retrospectives — Retool's Something Pretty Right history is the proximate source most secondary references rely on — and the quote is widely treated as canonical. The exact text against a primary scan of the May 1989 BYTE issue is a citation gap that has not been closed for this section; archive.org and the BYTE archives are the obvious targets and are flagged as such in the source audit. The substance of the argument is not in dispute; the wording is what's pending.

May 20, 1991 — Windows World, Atlanta

The launch itself was theatrical. Gates demonstrated Visual Basic 1.0 on stage at Windows World '91 in Atlanta, Georgia, on May 20, 1991. The video survives. He drew a form, dropped a button onto it, double-clicked the button, typed a single line of BASIC, and ran the program. The crowd reaction is exactly what you'd expect from an audience that had spent the eighties writing Petzold-style C against the bare Windows API.

The venue choice is worth pausing on. Windows World '91 was a developer-and-channel-partner conference, not a consumer launch. Microsoft was not pitching VB to end users at a Comdex-scale event; it was pitching it to the people who would build with it, sell it, and write the books and the magazine columns that would teach everyone else how to use it. Choosing the venue this way is itself an editorial signal — Microsoft understood, in 1991, that VB's adoption story would run through the developer-tools channel and the trade press, not through end-user retail. The product would be in PC Magazine and BYTE and on the shelves at Egghead Software within weeks of Windows World, but the wave started in a room of people who already knew what WinMain was.

Three things to notice about the demo, from the surviving footage:

  1. It was short. The whole "build, run, demonstrate" arc fit inside a few minutes. Microsoft's pitch was that VB collapsed something that took weeks into something that took an evening; the demo embodied the claim by example rather than asserting it.
  2. The code was trivially short. Gates typed essentially one line of BASIC behind the button. This was deliberate — the visual half was doing the heavy lifting, exactly as the 1989 BYTE article had argued.
  3. The audience was the right audience. Windows World '91's attendees were developers, ISV decision-makers, and channel partners. Microsoft was not pitching VB to end users; it was pitching it to the people who would build with it and resell it.

The contemporary trade-press coverage of the launch — InfoWorld, PC Week, BYTE, PC Magazine in their May and June 1991 issues — is the strongest independent corroboration of how the demo played in the room and how the press translated it for the developer audience that wasn't there. Those issues are accessible through archive.org's periodicals collection but have not yet been pulled into this section; they are flagged in the source audit as the next research target. The demo's effect on the room is documented in the surviving Microsoft-friendly retrospectives (Retool, Forest Moon); the independent contemporary read on it is the gap.

The three-part pitch

Microsoft's launch messaging had three explicit parts and they were not subtle:

  1. The hardest part of writing a Windows program — the UI — should be drawn, not coded.
  2. The language behind the UI should be the BASIC every Microsoft customer already knew.
  3. The whole loop — design, code, compile, run — should fit in one IDE on a 286 with 1 MB of RAM.

That last claim was the operationally sharp one. Visual Basic 1.0 retailed at $199, required a 286 with 1 MB of RAM and Windows 3.0, and was achievable on hardware sitting under a desk in any office in 1991. Each of those numbers was carefully chosen — the price was an impulse-buy for a department, not a procurement event; the hardware spec matched the median business PC; the OS dependency aligned with the Windows 3.0 wave Microsoft was already pushing through every channel partner.

The $199 price-point sat below every credible Windows-development alternative on the shelf in 1991. Microsoft's own C/C++ 7.0 (the contemporary professional Windows toolchain) and the Windows SDK on top of it were a multi-hundred-dollar professional purchase by the time you assembled a working setup; Borland's Turbo Pascal for Windows, the closest like-for-like RAD competitor in the same window, was priced as a professional development tool rather than a departmental impulse buy. VB undercut all of them — and the market it undercut into was the part of the developer population that would never have bought a C/C++ compiler at all. Microsoft was not competing for SDK customers; it was creating a Windows-developer demographic that hadn't existed.

The propaganda machinery beyond the BYTE article and the demo

Microsoft did not run the launch on a magazine piece and a stage demo alone. The pre-launch period also carried a developer-evangelism apparatus that would, by the mid-nineties, formalise into MSDN: NDA betas distributed to selected ISVs and trade-press reviewers in late 1990 and early 1991, channel-partner briefings through Microsoft's existing distribution relationships, and an emerging Microsoft Press / DevDays evangelism programme aimed specifically at the developer-tools audience. Mapping that machinery in detail against contemporary trade-press archives is research that hasn't been done for this section; the source audit flags it as a gap rather than an asserted result. It is named here because the BYTE article and the Windows World demo were the visible surface of a much larger run-up, and it would misrepresent the launch to imply they stood alone.

Reception

The trade press received Visual Basic 1.0 enthusiastically. PC Magazine awarded it Technical Excellence in its December 1991 issue, the highest editorial honour the publication carried. Visual Basic Programmer's Journal — a magazine launched specifically to cover this product — published its first issue in 1991. The existence of a dedicated trade-press title within months of launch is itself a reception data point: VB had a deep enough reader base, a fast enough early adoption curve, and a busy enough custom-controls ecosystem to sustain its own monthly publication from year one.

Hobbyist and shareware uptake was the strongest of any Microsoft developer tool to that point. Drag-and-drop form design lowered the barrier to shipping a Windows program from "weeks of C" to "an evening," and the shareware catalogues filling out through 1991 and 1992 reflect that immediately — VB became the default tool for anyone who wanted to put a working Windows utility into the wild without the SDK overhead. That award, plus the Cooper/Ruby form designer doing the visible work, plus Microsoft's existing fifteen-year BASIC user base looking for a Windows on-ramp, plus a $199 price point sitting comfortably below the threshold for departmental purchase approval, is what turned a tooling release into a wave.

The professional-developer reception was slower and more divided. C developers writing against the SDK looked at VB1 as a hobby tool; in 1991, that wasn't an entirely unfair read of what the product could do — VB1's database story was effectively absent, custom controls were a nascent third-party market rather than a mature one, and the runtime DLL footprint was a real deployment concern in the Windows-on-floppies era. The professional-developer story would not solidify until VB2 (November 1992) and especially VB3 (Summer 1993) brought MDI, ODBC, and the Jet Database Engine. By then VB had bypassed the question of whether it was a "real" tool by becoming the tool everyone was actually using.

Cooper's own retrospective verdict, looking back from the other side of all of it:

Had Ruby gone to the market as a shell construction set it would have made millions of people happier, but then Visual Basic made hundreds of millions of people happier. — Alan Cooper, on the Tripod-to-VB pivot.

That is the honest accounting. The Microsoft pivot from "ship Ruby as the Windows 3.0 shell" to "ship Ruby + Embedded Basic as a development product" is the call that changed the order of magnitude — and the pivot was Microsoft's, not Cooper's. The form designer was Cooper's, the BASIC dynasty was the Microsoft of the seventies and eighties, the integration was Ferguson's team, and the product would not have reached the audience it reached without any of them. But the wave was Microsoft's.

But before any of that, Microsoft had to answer a question its own dynasty was forcing on it: what about the BASIC customers still on DOS? That's the next section.

Sources

Source note. This section relies on Microsoft-friendly retrospectives (Retool's Something Pretty Right, the surviving Microsoft-produced demo footage on YouTube) for several of the framing claims about how the launch played in the room. Treat those as the institution's voice, not as audited fact. The independent contemporary corroboration — InfoWorld, PC Week, BYTE, PC Magazine, Visual Basic Programmer's Journal — is the next research target and is the level at which the institutional framing should be tested.