Friday, October 17, 2014

A Non-technical Overview of the SageMathCloud Project

What problems is the SageMathCloud project trying to solve? What pain points does it address? Who are the competitors and what is the state of the technology right now?


What problems you’re trying to solve and why are these a problem?

  • Computational Education: How can I teach a course that involves mathematical computation and programming?
  • Computational Research: How can I carry out collaborative computational research projects?
  • Cloud computing: How can I get easy user-friendly collaborative access to a remote Linux server?

What are the pain points of the status quo and who feels the pain?

  • Student/Teacher pain:
    • Getting students to install software needed for a course on their computers is a major pain; sometimes it is just impossible, due to no major math software (not even Sage) supporting all recent versions of Windows/Linux/OS X/iOS/Android.
    • Getting university computer labs to install the software you need for a course is frustrating and expensive (time and money).
    • Even if computer labs worked, they are often being used by another course, stuffy, and students can't possibly do all their homework there, so computation gets short shrift. Lab keyboards, hardware, etc., all hard to get used to. Crappy monitors.
    • Painful confusing problems copying files around between teachers and students.
    • Helping a student or collaborator with their specific problem is very hard without physical access to their computer.
  • Researcher pain:
    • Making backups every few minutes of the complete state of everything when doing research often hard and distracting, but important for reproducibility.
    • Copying around documents, emailing or pushing/pulling them to revision control is frustrating and confusing.
    • Installing obscuring software is frustarting and distracting from the research they really want to do.
  • Everybody:
    • It is frustrating not having LIVE working access to your files wherever you are. (Dropbox/Github doesn't solve this, since files are static.)
    • It is difficult to leave computations running remotely.

Why is your technology poised to succeed?

  • When it works, SageMathCloud solves every pain point listed above.
  • The timing is right, due to massive improvements in web browsers during the last 3 years.
  • I am on full sabbatical this year, so at least success isn't totally impossible due to not working on the project.
  • I have been solving the above problems in less scalable ways for myself, colleagues and students since the 1990s.
  • SageMathCloud has many users that provide valuable feedback.
  • We have already solved difficult problems since I started this project in Summer 2012 (and launched first version in April 2013).

Who are your competitors?

There are no competitors with a similar range of functionality. However, there are many webapps that have some overlap in capabilities:
  • Mathematical overlap: Online Mathematica: "Bring Mathematica to life in the cloud"
  • Python overlap: Wakari: "Web-based Python Data Analysis"; also PythonAnywhere
  • Latex overlap: ShareLaTeX, WriteLaTeX
  • Web-based IDE's/terminals: target writing webapps (not research or math education): c9.io, nitrous.io, codio.com, terminal.com
  • Homework: WebAssign and WebWork
Right now, SageMathCloud gives away for free far more than any other similar site, due to very substantial temporary financial support from Google, the NSF and others.

What’s the total addressable market?

Though our primary focus is the college mathematics classroom, there is a larger market:
  • Students: all undergrad/high school students in the world taking a course involving programming or mathematics
  • Teachers: all teachers of such courses
  • Researchers: anybody working in areas that involve programming or data analysis
Moreover, the web-based platform for computing that we're building lends itself to many other collaborative applications.

What stage is your technology at?

  • The site is up and running and has 28,413 monthly active users
  • There are still many bugs
  • I have a precise todo list that would take me at least 2 months fulltime to finish.

Is your solution technically feasible at this point?

  • Yes. It is only a matter of time until the software is very polished.
  • Morever, we have compute resources to support significantly more users.
  • But without money (from paying customers or investment), if growth continues at the current rate then we will have to clamp down on free quotas for users.

What technical milestones remain?

  • Infrastructure for creating automatically-graded homework problems.
  • Fill in lots of details and polish.

Do you have external credibility with technical/business experts and customers?

  • Business experts: I don't even know any business experts.
  • Technical experts: I founded the Sage math software, which is 10 years old and relatively well known by mathematicians.
  • Customers: We have no customers; we haven't offered anything for sale.

Market research?

  • I know about math software and its users as a result of founding the Sage open source math software project, NSF-funded projects I've been involved in, etc.

Is the intellectual property around your technology protected?

  • The IP is software.
  • The website software is mostly new Javascript code we wrote that is copyright Univ. of Washington and mostly not open source; it depends on various open source libraries and components.
  • The Sage math software is entirely open source.

Who are the team members to move this technology forward?

I am the only person working on this project fulltime right now.
  • Everything: William Stein -- UW professor
  • Browser client code: Jon Lee, Andy Huchala, Nicholas Ruhland -- UW undergrads
  • Web design, analytics: Harald Schilly -- Austrian grad student
  • Hardware: Keith Clawson

Why are you the ideal team?

  • We are not the ideal team.
  • If I had money maybe I could build the ideal team, leveraging my experience and connections from the Sage project...