Skip to main content

Google Docs began as a hacked together experiment, says creator

Google Docs began as a hacked together experiment, says creator


Sam Schillace talks about the challenges of collaborative editing and why people get freaked out when there's no Save button

Share this story

sam schillace box
sam schillace box

If Google Docs is the mother of modern word processing, Sam Schillace is probably its father. Back in 2005 he built Writely, a web-based text editor, which Google acquired in 2006. One month later, 90 percent of the company was using it. Writely formed the foundation of Google Docs, a product Schillace led for years before heading to Box to be its senior vice president of engineering and take on the next age of online collaboration. He took a few minutes to talk to The Verge about the building blocks of Google Docs, the persistence of email, and why killing the Save button still freaks people out. You can find him on Twitter at @sschillace.

Back when you were at Writely, why did you try to build collaborative document editing? Is that even what you were trying to build, initially?

"When we went to Google, Writely was internally adopted very quickly."

It got started as an experiment — I wanted to work with what was then called AJAX (Javascript in the browser) and noticed the “content editable” functionality of the browsers. I’d been working on word processing software for a long time (my first job was working on the second version of FullWrite, an early Mac word processor, in 1989, and my partner and I built Claris Home Page in 1994), so it was natural to see what kind of document editor we could pull together from the two technologies.

Gmail was about a year or so old at the time, so we knew the browsers could support the kind of code we needed to write (which they could, but just barely at the time). Collaboration on documents wasn’t really a thing we thought of as a first-order goal, we just thought locking was kind of gross (locking a document so others can’t accidentally overwrite your edits) and thought that it should feel natural if you were working on something with someone else.

When we went to Google, Writely was internally adopted very quickly — something like 90 percent of the company was using it in the first month. Which always made it funny when someone would tell me that browser-based apps were a bad idea that would never work.

What’s the technology look like behind collaborative web page editing? Was it straightforward, or hacked together using new technology?

Hacked together, mostly. The browser did most of the heavy lifting with the editor surface itself (though in later iterations the Google team had to write a new editor completely in Javascript because of limitations in that system). The client-side Javascript was pretty straightforward, only about 10 pages of code, mostly just dealing with merging changes from other users and generating deltas from your own changes. That turns out to be one of the hard parts — you’re editing HTML, and different browsers treat document features differently in the HTML itself.

IE used "p" and " /p" to denote an empty paragraph, for example, and everyone else used "br." So you can’t just merge the text on the server — you have to do a “normalization” step into a common, stable dialect of HTML, do the merge, and then “de-normalize” back into whichever browser dialect the specific client needs. If you get this wrong, you get “browser fights” where two different users lob the same non-change back and forth endlessly, because the act of applying the delta “generates” a new change.

"We made a deliberate bet that users would want speed, convenience, and collaborative features."

What kinds of decisions did you have to make along the way, like do you keep margins? Page breaks? Do you kill the Save button once and for all?

We made a deliberate bet that users would want speed, convenience and collaborative features enough that we could ignore richer functionality like rich formatting, margins, pagination, etc. This turned out to be true in the early days but as we got deeper into the market, we realized that some of that functionality was either important enough to everyone, or helpful enough in opening up the market to new users, to be worth the pain of implementing it.

The Save button is kind of funny. It’s obviously (to me) better to never have to think about saving, but we’re all conditioned to think about it, worry about it, etc. We’ve been trained to do work for the computer, which sucks. So, we really wanted to take that out, but it freaks people out when they “can’t save” their document. I think it’s funny — as though, if save were broken, your act of clicking on the screen would help the computer get that job done magically.

Will email always exist or will we move to even shorter forms of communication like Twitter, text, etc?

"Email will exist as some kind of communication substrate for a long, long time."

I think email will exist as some kind of communication substrate for a long, long time. It’s a classic example of a “worse is better” technology. The abstractions of domain, address, mailbox, message, attachment, reply, etc are so fundamental and useful. And I think the openness of the system will keep it around — I love some of the things like the social asymmetry of Twitter, but I really wish it were based on an open protocol that others could build against, instead of being proprietary. Not because I’m a hippie — but because I want to see those systems thrive and survive for a long time, and I believe openness is a critical part of that durability.

How does working at Box differ from working at Google? How do the two companies’ philosophies differ?

There’s a lot of Google culture in a lot of the more interesting companies in the Valley these days. Box is like Google in that it aspires to be a world-class engineering organization, spends a fair amount of energy being introspective about engineering culture and process (especially for a company at this size), and is filled with really smart, passionate, and interesting people who are open to learning from and collaborating with each other. I think Box is more focused on the needs of enterprise users which can be a little “messier” and more complex than the kinds of problems Google likes, which tend to be more amenable to generalization and automation.

Is there anything that could’ve saved Google Wave, or was it just ahead of its time?

I think it was trying to solve a lot of problems all at once, and this is really tough. I don’t think they articulated a clear user benefit that was universal to their whole user base (Writely: collaborate on a document, Twitter: broadcast your message, or consume news from others, Facebook: keep tabs on your friends, Gmail: never worry about managing a mailbox again…). The combination of something disruptive (so, by definition, hard to understand because it breaks assumptions) without a clearly articulated problem statement, I think was fatal.

What science fiction technology would you most like to see become reality?

"We’re idiots as a species."

Cold fusion. Though I’m afraid that if we do actually ever unlock some kind of energy source like this, we’ll have true global warming through the process of actually heating up the planet with waste heat. We’re idiots as a species — we'd totally do that if we could. “Hey, honey! Let’s fly the house to Paris this weekend!”

Having worked in tech for so long, do you ever feel a need to disconnect? If so, how do you do it?

It’s funny — we have a place in Mendocino County that’s in the hills, so it doesn’t get cell coverage. And we always resisted putting in internet access there — it was really nice to go up, be offline, and let everyone know we were out of pocket. We’ve had to install a line now because the kids need it to do homework (amazing that something that didn’t exist for the first two startups I did is now a requirement for middle-schoolers). That’s changed the tone of the place, and I miss being out of contact. I don’t do anything active to recreate it — I feel less the need to disconnect as an individual, and more a desire to be disconnected to interact with other people.

One thing my family does though: We always have dinner together, every night. We ban phones at the table, and we usually light candles on the table, which seems to encourage everyone to linger and chat for an extra 20 minutes or so. It’s very nice — I highly recommend it. I don’t know what it is about the candles, but it totally works.

Read more 5 Minutes on The Verge