Support for Microsoft platforms is obviously important. Not just Azure, but even AWS keeps coming out with really cool stuff almost by the day. It is tempting to think that an army of engineers can shorten everything and get things done quicker (mythical man month?) - but I personally don't think that things will work out that way. There is no amount of money that can bring all the key contributors to Julia together - they are all doing dramatically different things. With $X million, one may be able to assemble a different team, but then the outcome will certainly be different.
Even with JuliaBox, we have built things just bit by bit as we need them, and that has turned out to be a more useful way to think about many of these things. It is not just about writing lots of wrappers, but thinking about the right abstractions, and that either comes from someone with deep experience and thinking, or by actually doing things and reflecting on how to do things better. The question I would rather ask is - what do we want in Julia 1.0, and the surrounding ecosystem (IDEs, packages, cloud platforms, software stacks, etc.). Like I said, I will shortly clean up what I have written up and post it here. From there, we should as a community ask about how to best get there - whether it is about finding funds, finding people, finding space, or finding pizza - most likely a combination of all these. -viral On Saturday, May 16, 2015 at 7:42:30 AM UTC+5:30, Eric Forgy wrote: > > Hi, > > I have never been a huge fan of Microsoft, but I have to admit that with > Nadella at the helm, it does seem like an entirely new company doing some > pretty amazing and innovative things. Windows 10, Surface Pro 3 and Azure > are quickly making a lot of the things we're doing in my FinTech startup > obsolete already. I'm increasingly tempted to ride their wave and hitch my > startup to MS technologies. > > Microsoft recently acquired Revolution Analytics > <http://blogs.technet.com/b/machinelearning/archive/2015/04/06/microsoft-closes-acquisition-of-revolution-analytics.aspx> > > and have integrated R into the Azure SQL suite, which is pretty awesome. > They have also done quite a lot with integrating Python. > > You can guess where I am going with this... > > I can imagine Julia would have to get to 1.0 before Microsoft would > consider incorporating Julia directly into their products. > > I've asked before > <http://blogs.technet.com/b/machinelearning/archive/2015/04/06/microsoft-closes-acquisition-of-revolution-analytics.aspx>, > > and Keno estimated back in January that 1.0 is 2 years away. In light of > the recent discussions around Julia Computing, can you think of ways this > could be accelerated or is it just something that takes that much time? For > example, if significant funds were raised and the core team could build a > small army of senior full-time developers, could this get done quicker? > > If we allow ourselves to dream, how much funds would be required and what > is the shortest path to 1.0? > > "With US$X million, we could produce Julia 1.0 in Y months." > > What is the most optimistic X and Y (allowing ourselves to dream)? > > Best regards, > Eric >
