Sunday, December 15, 2013

Handling collaborative projects in the age of the cloud - how to manage without spending any money!

As a clinician cum scientist, I am almost always involved in collaborations with multi-disciplinary groups.  And, as an american who has done some training in the UK, many of those groups are NOWHERE REMOTELY NEAR one another.  To keep myself, and my collaborators sane, I've been constantly experimenting with different combinations of cloud based repositories so that we can all work on things together.

So, I started out with +Dropbox by itself, but quickly found that, without spending money, I quickly filled up my allotted space.  Even after totally maxing out the freebie upgrades (I think I have like 35Gb or something thanks to some promo at Oxford, referring everyone I know and being a shameless social media promoter - every Mb counts!), I still am limited.  So I added +SugarSync for my personal files, to clear off dropbox and leave it for just collaboration.  Well, that filled up pretty quick too, and as I haven't found sugarsync to be as easy to navigate as dropbox, I haven't worked as hard to maximize my space there.  I muddled along with these two but still struggled with things like version control and the desire for contemporaneous editing.  Which led to...

Google drive and google docs.  So, with google drive, you get 25Gb or something, but I started to get confused as to where things were.  When I wanted to collaborate real time though, I was stuck, as google docs was really the best thing going (and still is, for many things).  However, in the last few months, I discovered +writeLaTeX, which has really changed the way I do business.  This is a (mostly) free service (I haven't hit the wall yet, and it is pretty big considering that you don't store data there really) in which you can create, manage and edit LaTeX documents with as many people as you like. For me, as a mathematical biologist, this was great - so long as my collaborators were also theorists who were tex-savvy.  I first used it doing a little hack-a-thon with some friends that I blogged about before. Recently though, as in, in the last few days, they have added a Rich Text formatting layer which, I think, will change everything again.  No longer will I have to give the link to a document to my biological/clinical collaborator with the caveat 'just ignore everything that isn't text - squint a bit if you have to'.  Now, they can just go ahead and edit away just like they are in word or whatever, but I can come in behind and have the full functionality of LaTeX.  So that aspect seems solved.

**note - I LOVE writelatex, and you should check it out... and if you do, use this code and I get 50MB more space.  https://www.writelatex.com/signup?ref=5ed6049a8e70 **

The only issue left for me is, now that I'm cranking TONS of simulations in (what I hope is) the final push toward this whole PhD thing, how to work on a DATA-HOG of a project on two computers (I can't stay at work late because I want to go home to my super sweet little kids, but then I DO want to work after they go to bed without driving back to work) or to share this with someone else.  I thought, at first, that I'd just clear out my dropbox a bit, and try to be parsimonious with what I saved...  but, as the output for my simulations is \mathcal{O} 30Gb/simulation, this quickly became unfeasible.  Enter bittorrentsync - my saviour.  This little gem (free) lets you have a synced drive on as many computers as you like.  The freeing move here is that there is no cloud interaction, so there is NO SPACE LIMITATION.  (There is also no auto-backup for the same reason... but this is obviated if you use Time Machine or something similar).  All you do is download the software (a couple Mb's) and then, to set up a shared directory, it generates a 'secret' which you share with whomever you want to share.  The secret is a massive string of letters/numbers that is autogenerated - and I bet even the guy at xkcd.com would approve.

So, now I'm set and I don't have to think too hard.  I also don't forsee ever having to spend any money (until maybe I have a lab or my own and lots of people, but by then - if that time ever comes - I will have grant money to spend on such things...  ?).

Summary: I use dropbox for shared/travelling talks, figures, syncing my papers library, shared simple code (matlab, etc).  I use sugarsync for my personal documents (though I might phase this out...). I use writelatex for (now and going forward) ALL papers I write and bittorrent sync for shared working directories (also home/office syncing of working directories).

As an aside, dropbox has a cool feature - since everyone can modify things at will, they have a nice cloud backup system, which includes who the last person to interact with a document was, so in case your supervisor starts randomly deleting hunks of your thesis, you can call them on it - not naming any names...  +Alexander Anderson :)



To be fair, I don't think Sandy actually deleted this stuff... and I was able to recover it, but it was pretty funny :)