The Real McCrea

[20 Years Ago, Part 12. Other options: prior post or start at the beginning.]

The first mainstream and well-remembered “WYSIWYG”¹ HTML editor, FrontPage, was released in October 1995. Less than three months later, the small startup that created it, Vermeer Technologies, was acquired by Microsoft for a whopping $133 million! FrontPage would become a key weapon for Microsoft in its ruthless, monopolistic “browser war” against Netscape.

But FrontPage was actually not the first WYSIWYG HTML editor; it was just the first one on Windows. A full year prior to Microsoft’s acquisition of Vermeer, Silicon Graphics, at the time the hottest company in Silicon Valley, launched a product called “WebMagic”. It was the actual first WYSIWYG HTML editor, empowering designers and business people for the first time to create web pages without learning to code HTML. This is the previously untold (and rather improbable) story of how that historic application came to be…

I’ll start the story from hours after I had secured $2.5 million in funding from Tom “TJ” Jermoluk, SGI’s President and COO, based on my pitch/commitment to launch a product line for web authoring and web serving by the end of January 1995. That deadline was less than 80 days away (a period that would include Thanksgiving, Christmas, and New Year’s!).

And that means that these events took place in the single craziest time of my career. It was such a blur, that I’m sure I’ll miss some key facts and miss-remember some details. With hope, others involved in the story will keep me honest.

I think it was in the afternoon of the day of the funding that Way Ting sought me out. He was the Vice President and General Manager of the Visual Magic Division (VMD) that created the desktop software environment and a bunch of tools to showcase the differentiation of our powerful workstations. And after the amazing events of the morning, he was keen to sign up for a critical element of the plan.

“We want to create that SGI-quality authoring tool you described,” he said.

“That’s great news,” I said.

“But there’s just one problem,” said Way. “The timeline is too short. I don’t think it’s possible to develop such a tool by the end of January. Any chance we can add a few weeks to the schedule?”

Way knew a heck of a lot more about software development than I did, but I had a deep conviction that timing was vital in this rapidly emerging web market. I feared that Sun or Apple were about to beat us to market with web authoring/serving solutions, and I strongly preferred to launch first with an imperfect product than to launch second or third with everything on my wishlist.

“I’m sorry, Way,” I said. “I know it’s crazy, and it may not be possible, but the one thing that is certain for this project is that we’re launching by the end of January.”

“Okay,” he said. “I’ll see what we can do.”

To my surprise, within a few days, he reached out to me again, asking if I’d accompany him to a meeting in Palo Alto with a company that had a possible solution to our problem. The company was Enterprise Integration Technologies (EIT), and although I didn’t know it at the time, it happened to be the startup that Marc Andreessen had moved to California for, and the one from which Jim Clark recruited him just months later to co-found what would become Netscape.

Our host for the meeting was EIT’s founder and CEO, Marty Tenenbaum, a more-than-a-little smart guy, with multiple degrees from M.I.T. and a PhD from Stanford. EIT was a pioneer in e-commerce, having conducted the very first Internet transaction in 1992. Therefore, it was not a surprise that nearly every desk had an SGI workstation on it. They were the real deal.

I have no idea how Way had managed to make this meeting happen, but it was pretty magical. What Marty showed us that sunny afternoon looked like exactly what I had been advocating for – an SGI-native web authoring system, with WYSIWYG HTML coding and an intuitive, drag-and-drop user interface. It looked pretty polished.

But when Way asked, “What is it written in?” Marty replied, “WINTERP.”

Ruh, roh!

I had never heard of that, and I’m guessing Way hadn’t either. Marty went on to explain that it was an interactive, object-oriented user interface language that EIT had developed.

As we drove back from the meeting, it became clear that this was hardly an ideal path for achieving our goal. I resolved to pursue my Plan B with fervor – getting SoftQuad to port their less-than-ideal alternative HTML editor, HoTMetaL Pro, as the candidate for bundling with our web workstations. It wasn’t WYSWYG, nor was it going exploit the differentiation of our OS and tools, but it could allow us to check-the-box for “HTML editor”.

But then, to my surprise, Way reached out to me a few days later with a proposal for another outing, this time to Sunnyvale, to the headquarters of Amdahl.

“Amdahl?” I thought. “Did I hear Way right? Why would we go visit such a freaking dinosaur?”

For my younger readers, Amdahl was a maker of IBM-compatible mainframe computers. Hardly the most likely place to find cutting-edge web software!

You may be thinking my story is old-timey, but here’s how ancient Amdahl looked to me 20 years ago:

Much to my surprise, what we found at Amdahl was a decent HTML editor, still under development by a single contractor developer, David Koblas. It was a Motif-based program, written in C++, running on Solaris. In short, it was exactly what we were looking for – the right kind of code base, far enough along in its development, that turning it into a robust SGI-native application just might be possible in weeks, not months.

I don’t know all the details, but within days we had struck a deal with Amdahl that somehow brought us both the code and David. In a recent email exchange, David recalls that to avoid creating a taxable event, no physical media were involved in the transfer of the source code; the bits were passed from Amdahl to SGI via FTP.

A small team was quickly assembled around David, including Ken Kershner (who managed the team), and Ashmeet Sidana, Baron Roberts, and Victor Riley. By then, I had locked in a launch date for the project: January 25, 1995. So this newly formed “WebMagic team” would have just under eight weeks to port to IRIX, integrate deeply with SGI’s desktop environment and media tools, and polish code that David now recalls as “quite buggy”.

In addition, the team fully embraced an even more audacious goal – full “render compatibility” with the now dominant browser, Netscape Navigator. And what that meant was aiming at a moving target. The team at Netscape was extending the capabilities of the browser at a blistering pace. The WebMagic team wanted to support it all, including tables, forms, and, yes, even the blink tag.

It took long days and nights, with David and Baron often working until 3:00am. (On those nights, they would typically switch to pair programming at 1:00am to minimize bugs.)

The heroic work of the team paid off. When we launched, WebMagic had become all that I and the team had hoped: a true WYSIWYG HTML editor with the polished look and feel of a word processor, drag-and-drop integration with SGI’s media toolkit, and pixel-perfect compatibility with Netscape Navigator.

And that is the rather improbable story of the creation of the very first WYSIWYG HTML editor.

But what of that “project” that WebMagic was such a central part of? To be continued…

¹This is an old acronym for “What You See Is What You Get” which came into common use in the 1980’s during the word processing revolution. From Wikipedia “a WYSIWYG editor is a system in which content (text and graphics) onscreen during editing appears in a form closely corresponding to its appearance when printed or displayed as a finished product, which might be a printed document, web page, or slide presentation.”