Wednesday, June 13, 2012

Research protip #1: always save multiple manuscript file versions

It's been a month since my last post, and I apologize: I have plenty of new material and topics to discuss, so stay posted. I've been busy juggling two manuscripts, preparation work, and catching up on new literature. An interesting problem hit me last night while manuscript writing, and I thought it would be an excellent talking point.

Writing up a manuscript for journal submission naturally takes a bit of time, patience, and lots of energy - because it can take so damn long, you should expect your ideas to evolve - I'll give an example. I've been working on this juggernaut of a manuscript for one year (started last June) and have more or less been working constantly on it since then, without taking much longer than a week at a time away from it. In cases like this, most of the material stays relatively fresh in your mind - even still, you'll look back and be puzzled why you wrote something a certain way, and want to change it. At the same time you should ask yourself "well... I must have said this for a reason; what was it?" and if you didn't write it down, then you've got a minor problem. These sorts of 'bugs', if you will, crop up all the time while writing - so it's best to save multiple versions of your document file as you go, so in case you need to go back and look at an older version - you can, as a sort of manual backup. Note that this is a separate issue from just backing up documents on a flash drive or an external hard drive - that should be done as well. With this long ass manuscript, I now have about twenty versions saved, all with the same title, with v.1.0-1.20 on the end. When in the course of writing should you move on to the next version? I tend to save the next version after typing up a few pages or more, or work on one file version for about a week or two (i.e. I don't save unless I've typed more than a few pages). If you want, you could keep with v.1.0 until you had all of the major parts of the manuscript filled in (abstract, intro, methods, results/systematic paleo, discussion/conclusions, acknowledgements, references) and only update it once you're revising and tweaking the "completed" manuscript - this works, but I only do this for really short articles where I can get all of that done after a week or so.

What if you have coauthors? I get the impression that despite things like google documents, cloud-based services, and new shiny stuff like that, the easiest thing - if you have only two main authors - is to just email it back and forth. Each time Morgan Churchill and I email a draft, the version number increases by one. Or, alternatively, I'll send a manuscript file titled "pinnipeds are awesome v.1.12 RWB comments" - something like that. This actually works out quite nicely, and you can separate out the individual drafts for easy access. Most people know about this, but 'track changes' (i.e. in microsoft word) is the best way to send an updated manuscript to a coauthor so they know exactly what you've changed.

Now comes the problem. Lets say you have three versions of your file - "pinnipeds are awesome v.1.19", on your flash drive, laptop at your office on campus, and another on your home computer (actually, this is me as of yesterday). You have forgotten to save the newest version (the one on the flash drive) with a new number: this version and the one on the laptop have the most recent round of revisions, but have the same version number as the un-updated, un-revised version on your home pc. And you remember "Ah, I need to update the figure captions, because I annoyingly added a figure after fig. 15, and everything after I needed to increase the number by one (e.g. fig. 16 becomes fig. 17). This is exactly what happened to me last night - and I spent a good hour updating all of the figure numbers, only to realize that I had done it on a manuscript file lacking the most recent round of revisions.

Fortunately, there is a nifty tool called "document compare" in microsoft word which can help you if you screw it up like this. Go to Tools>Compare and Merge Documents, and an "open document" folder will pop up; check the box that says "legal blackline" and your document will show up, with the changes highlighted in the same way changes appear with the "track changes" tool. You'll see every change between the two, and will be able to reject/accept each one; in this case, all I need to do is reject all of the changes involving figure citations, and accept all of the other changes - which are the most recent revisions. It depends upon what version you open up first and which you compare it to; say you make revisions A to document 1, and revisions B to document 2; if you open up document 1 and compare with document 2, you will need to reject all revisions A and accept the rest from revisions B (whatever revisions are already in document 1 will be overwritten when doc. 2 is opened up in document compare), and vice versa. If that makes any sense - it may seem confusing now, and maybe for the first few minutes of trying document compare, but it's not too complicated after thinking about it and experimenting with it for a few minutes. Now, if you'll excuse me, I've got a boatload of figure captions to fix.