Drat. Drat drat drat. Drat. I've just been reading a mystery story, a good one. With a strange death and a scroungy detective and an alluring love interest, who is actually not who we think she is, I mean he is, I mean – but that isn't the point. The problem is that I got to the last page and there was no last page. Must have gotten torn out somewhere (maybe when I stuffed it between journals and my laptop in my carry-on bag. But I mean, drat. It's killing me.

When we envision writing a paper, we often think of it as a mystery story. In our outline we set up the scene (our figure 1) and then we explore the territory (figure 2). Then, wham, we describe the mystery – the body in the locked room with no windows (figure 3). We examine the crime scene, finding clues (figure 4). We connect the dots (figure 5). And then we reveal (ta da!) the answer (figure 6) and wrap everything up neatly (figure 7). At least, that's how we often think about writing up our cool mystery, knowing that the reader will be breathless with anticipation until we, the detectives, call everyone into the sitting room to reveal the culprit in the stunning climax.

But really? Bad idea. Because we (the readers) aren't going to curl up with your paper and patiently let you build to your climax. What we're going to do is read your work up to the first or second figure, and if it doesn't grab us, we're going to go to the next paper. I mean, you already blew the ending in the abstract, right? So what is the sense of taking us slowly to your big final scene?

I learned all this the hard way. Many, many years ago, when I was just a molet, a senior post-doc in the lab, Dr Red Panda (not to be confused with Dr Panda – different family altogether) read a draft of my paper and eviscerated it. Boring! I had a great result but RP told me to put it right up front, figure 1, to grab the reader. “But,” I cried, “then the rest of the paper is just analyzing the result and testing the conclusion!” “Exactly!” said RP. “Anyone who only gets to figure 1 will get the main message, and those of us who keep reading will learn more about it. What do you think a paper is for, anyway?” I never forgot it and, indeed, it is a lesson I teach to my own trainees now, and we go through the same dance. La dee dah.

But that isn't what I really wanted to talk about. What I really wanted to talk about is what actually happens when we've written our nice, lovely, exciting detective story. We have to publish it. And that's when it all goes wrong. First, the editors of the journal, and then the next one, and maybe the next, don't want to review it. And then it's reviewed (hopefully) and the reviewers want to know more. Much much more. Because even if we've told a good story (we thought that we had) it just isn't the whole story. And anyone (even a reviewer) can think of more we can do. So back to the bench and, hopefully, back to the journal, and maybe back to the bench a bit more. In the end, if everyone has done their job well, the story is, in fact, richer, better, more complete. Maybe it isn't the whole story, but it's more satisfying. Sure, by the time it's published, we're sick to death of it (“finally got that frickin' paper accepted”). But somehow, the system sort of works. Sort of.

Sure, I'm complaining, but there's more here. And something that might be terribly important (and, perhaps, wrong) about the way we do things. Everything. The way we do science. And I think I can prove it.

There is a growing perception that science, in general, and biomedical research, in particular, is simply too slow. Not the emergence of new information – that happens so quickly that it is nearly impossible to keep up with everything. But step back a bit and look at the big picture. Why we receive public support for our work (or, if you are in the private sector, why gobs of money is coming from investors). While an argument could be made that paying scientists so they stay in their labs instead of causing trouble in the community is ultimately cost-effective, I really don't think that is the rationale. The fact is, the public are paying in hopes that real, live problems will be solved through the application of the scientific method by people who have some idea of how to use it. And in reality, slowly but surely, some of these problems are being addressed. And of course, some aren't.

Sure, a huge part of the reason is that science is really, really hard. Even if we have what could be a genuine breakthrough, and we know how rare those are, it will take years and years before we know if this will actually solve anything (and meanwhile, those who translate the work to the problems that people have must choose, carefully, what might represent such a breakthrough). All of this is terribly difficult, slow and full of landmines that could blow up the effort (even if the basic idea was right). Breakthrough (or maybe breakthrough), translation, misstep, back two steps, forward a bit, ka-boom, back to the drawing board, and back to the bench. And meanwhile, your mother's friend is asking, “So how long before you cure this disease?”

I've been thinking about this a lot (hey, I'm the Mole, I can't help it). So we have this problem: lots and lots of new information, petabytes of information, are hitting the e-press every day. (I just checked: 75,302 publications in the past two weeks. Wait, 75,307.) New cures, not so much.

There has been quite a bit of discussion about this problem, even from me. Part of the problem (just part) seems to lie with the quality of what is being published. Some pundits suggest that too little of it is reproducible (see Replicant). Some suggest that too much of it is just faked (see The End of Science), but those people are just wrong. Some even think we should just give them the money, and they'll fix it (yeh, how could that go wrong? – but watch out for these people, some of whom are in the other two camps making a lot of noise).

I think that maybe it would be useful, instead of whining about not being able to repeat someone's figure 6b, or gazing endlessly at the smudge on the third loading control in the second western blot, to take another look at the way we write, review, rewrite, review, re-rewrite papers. Maybe the whole way we approach reporting on our findings. And. maybe, what we think papers are for, anyway (thanks, RP).

But wait! I just found a torn-out page tucked between my journal with the nice, soft pages and the one that is big and shiny! Maybe it's my long-lost solution to the mystery, the one that has been driving me nuts. I'll be back.