For two years I was on the Canadian NSERC committee that reviewed individual grant applications. Fascinating.

After reading dozens of applications you can begin to see patterns emerging. I’m going to review some of these  patterns, all but one of which I don’t recommend. No guarantees but I hope this helps.

This is common but very naive …

I will solve X. 

In  other words, I don’t  need or have a good idea. Why should we believe you?

Slightly more sophisticated is:

I will apply A to solve X. Still no good idea in evidence. A is usually standard technology, e.g. eigenvalues.

Also not good:

I will try to solve X until I have a good idea. 

At least you admit you need a good idea, although you also admit to not having one yet.

More optimistic is:

I’m bound to have a good idea at some point. 

And here is one of the most common patterns:

I plan to have a good idea.

As above, but I know exactly when. Typically, the applicant presents a multi-year schedule with what might as well be labelled “have a good idea” in e.g. the third month of the second year.

Even more optimistic:

I have several ideas and one is bound to be good.

That remains to be seen. A very common variant is:

I have an indexed family of  ideas, one of which must be good.

Here they describe a general idea with a parameter, which might e.g. be a search method. They are assuming (hoping) that for some value of the parameter the idea will be good. Again, no guarantee any value of the parameter will work.

I  once had a good idea. 

Good for you. But is it still relevant to the problem you’re working on?

I have an idea which I will investigate for goodness. 

Sounds risky.

Here’s one I’ve used (and gotten away with):

I have this fascinating idea that must be good for something.

Anyway, I don’t recommend any of the patterns above. The one you should aim for is

I have an idea, here it is, and here’s why it’s good. That’s the ticket!

And this twist is almost universal:

I have an idea but first let me give you some background.

This (anti- , i.e. bad) pattern is so common in other kinds of writing that journalists have a name for it: “burying the lede”. It means beating around the bush, procrastinating telling people what it is you’re up to, or what it’s all about. Instead, the writer begins establishing background information, and in technical writing this means giving a little tutorial before getting down to business.

I often did a little test to see how far I have to read through the application till I find out what the applicant in fact intends to do. In practice this could be any where between 1/3 and 2/3 the way through. I would often mark the  point with a pen so that when I take another look at the application I don’t have to hunt to remind myself what the applicant is proposing.

Burying the lede is also common in abstracts of talks. In fact it’s become the de facto standard form. I had a student who in his first draft of his dissertation abstract buried the lede and was puzzled at my objection. All the grad students he knew used this pattern, as did visiting speakers in their colloquium abstracts. I told him I don’t care how many people use it, it’s not as effective as stating up front what you’re doing and why, and expanding on the explanation after. He grudgingly agreed.

The latest tendency in colloquium abstracts is to leave the “in this talk…” sentence to the very end. At least you know where to look.

The very best pattern is to begin by saying something like “I intend to solve X using idea I”.

Dog-gone it, people will like you!