Last weekend, I spent a good portion of my "free-time" watching the one-hour-twenty-minutes Google Wave demo from the Google I/O conference.
While I found my self getting caught up in the excitment of it all, I couldn't help but feel a sense of DejaVu...as if I had seen something like this before, that had failed miserably. And then, I checked my bookshelf:
Dreaming in Code by Scott Rosenberg (of Salon.com fame). This book was the real-life chronicle of Mitch Kapor's attempt to create an Open-Source, cross-platform, Personal Information Management (PIM) tool.
For those who have not read the book, it is a frustrating and predictable (in 20/20 hindsight) failure to redefine email, tasks, notes, calendars, etc. as generic items that a user could work with interchangeably. They understood even then that people worked with email and those other "Outlook/Exchange managed items" differently than was originally intended. That the email often turned into tasks, documents, appointments, and near real-time conversations, and historical records. And, it was all Open Source. Thus, my comparison with Google Wave.
The Demise of Chandler
The system was called Chandler and the project had many issues. From a management perspective, the project suffered from:
- Ill-defined (or undefined) scope and features
- The Architect in the ivory tower
- Lack of consistent direction to even incremental delivery
- Going too deep into niche parts of the solution without starting to solve the core issues
- Eventually picking regular delivery dates but not delivering any real value that could start to replace Outlook/Exchange
- Trying to run OSAF so differently from other Silicon Valley startups that things ran slowly and without consistent direction
From a technological perspective, the solution was problematic because:
- In an effort to avoid being Microsoft Exchange (and paying consultants ridiculous amounts of money every month to come in and do care-and-feeding), it was decreed that Chandler would be a peer-to-peer system with no server.
- Web access was an afterthought, and then they had to figure out how to provide web access to a non-centralized peer-to-peer system
- To enable the client (and really, the WHOLE system) to be cross-platform, they chose to use Python and a Python GUI toolkit (but most people do not have Python on their machines, the GUI toolkit did not fit their needs--especially considering that the UX was not designed until after they had already chosen that toolkit). As a side note, the GUI toolkit was also Open Source and they often had to decide whether to wait for that team to fix issues or to make the changes themselves, which often delayed the fixing of bugs.
As the book finishes, all Chandler ever became was a poor Calendaring system. Disappointing.
What, If Anything, Does Chandler's Failure Predict for Google Wave?
As I watched the demo, I was acutely aware of the overarching goals shared between Chandler and Google Wave, so I started looking for the signs of failure that were so obvious with Chandler...and I saw none of them. In fact, I saw many reasons to be encouraged that Google Wave was not going down the same paths and has a great chance of success and adoption!
- Even though the project is Open Source, it has been incubating for a year and a half and has been actively used by the team as they go
- It is a platform and protocol so others can contribute to, build on top of it (ala Google Maps) or adapt it to a non-web-browser environment if they needed
- Much like AdSense, Wave has the ability to be embedded into and published out to any web page
- Having proved the basic concept of a Conversation with GMail and the insight of how much time is wasted reading "So-and-so is typing..." is quite unique
- It can run in the cloud but is also federated and can communicate to peer servers, just like email--which has been EXTREMELY successful
- Real-time language translation
- They have already solved many of the challenging issues (real-time collaboration across servers, languages, and business boundaries)
- The browser is the cross-platform programming toolkit and runtime (so no additional installation required and accessible from anywhere!)
- Even though they only showed the collaborative creation of a document, Google already has Email, Docs, Spreadsheets, and Calendar, and I'm sure they have the ability to figure out how to tie Calendaring and Office Documents. If they succeed at that...look out Microsoft!
- This is the team that provided Google Maps, that enabled the coolest Mash-ups, and that clearly recognize that they don't even quite know what Google Wave is going to become. But they recognize that, have solved some very hard problems, and are open with it so we can participate in the conversation about what it should be.
I definitely feel the chances of spectacular failure similar to Chandler are nil. It is still early and it will be facinating to see what tie-ins to existing email and calendaring systems will be during some type of transition time. It will be bizarre to see which companies develop their own servers based on the specs and protocols. It will be awesome to see Waves embedded in blogs and news sites and that we can continue to participate in community and conversation anywhere we happen to be on the web or on our mobile devices.