NSF Application Tips: LaTeX edition

I’ve previously written about applying for an NSF grant here. But for those applying this year, I have a few further technical tips concerning the technical specifications of your LaTeX document. If you are the type of person who uploads all their files at the last moment (not me), then you could be in for a rude shock if you haven’t written your proposal up to code — rules are being checked by computer and are much more stringent this year.

The first requirement is that the various subject headings “Intellectual Merit,” “Broader Impacts,” etc. need to be on separate lines and without any further characters. If you make some error (say write “Broader Impact”) your file will not be accepted and you won’t be able to submit it. (You will get an error message.) Of course, these things are easy to fix. There were, however, a few other problems which stumped me for a while.

The margins cannot exceed the allowed specifications. In particular, if your document has page numbers, the machine will interpret your file as extending too far in the lower margin. Hence you need to manually remove the pagination. This can be done, for example, by including \pagenumbering{gobble} in the preamble.

As for the left and right margins, at least 1 inch margins are required. You might think this is easy to enough to ensure by including (for example) the command

\usepackage[margin=1in]{geometry}

However, that didn’t work for me. I *could* fix it temporarily by extending the margins slightly further, but that had the problem of making the proposal too long. For me, fitting my proposal into the 15 allowable pages is pretty much like trying to write a tweet: make a draft, then find it is way too long, and then painfully eliminate unnecessary sentences until it *just* fits. The difference, of course, is that it’s worse than a tweet — at least a tweet is very transparent as to the allowable number of characters; LaTeX is completely mysterious (to me) in its algorithm for spacing and paging, so I end up do a lot of back and forth editing paragraphs by making minor word changes so that they end precisely at the right margins (instead of wasting almost an entire line by dribbling a few words onto the next line). Then, of course, there are those times when your document is three lines too long and so you savagely cut three lines from your beautiful introduction only to find that it is still exactly three lines too long. All of this is to say that extending the margin to 1.1 inches was not an acceptable option for me. So the algorithm was to do a binary search (try uploading the first half of the proposal, then the second quarter, etc.) until I found the offending section of the file. The left margin issue was ultimately caused by an enumerate environment where I had used the following construction:

\item[{\bf Coherent]}

LaTeX had automatically adjusted the location of the word Coherent slightly into the left margin space, which was making the NSF system decide that my entire proposal was unacceptable. This was fixable by adjusting the enumerate environment thus:

\begin{enumerate} [leftmargin=2cm]

The right errors could ultimately be resolved by addressing all the “badness” in the output file. (For example, I had used a few \mbox commands to squeeze things onto a line and that came back to bite me.)

At this point, the file could be uploaded, but I it was still generating severe warnings about fonts, namely:

• Your file contains font type(s) that are not supported. For details about font types currently supported in the system, please refer to the Proposal Preparation Document Upload FAQs. NSF may return a proposal without review if it includes font types other than those specified in the Proposal Preparation Site Document Upload FAQs.
• Your file contains font size(s) that are not supported. For details about font sizes currently supported in the system, please refer to the Proposal Preparation Document Upload FAQs. NSF may return a proposal without review if it includes font sizes other than those specified in the Proposal Preparation Site Document Upload FAQs.

Returning again to the binary divide and conquer method, the first problem was isolated to the use of the LaTeX symbol \boxtimes. The second I never completely work out but at least determined that

\sum_{n=0}^{\infty} \frac{q^{n^2}}{(q)_n}

(which comes out to)

$$\displaystyle{\sum_{n=0}^{\infty} \frac{q^{n^2}}{(q)_n}}$$

caused an error (the thought is that the subscripts are being identified as too small a font). Now these were only warnings and did not necessarily mean there was going to be any actual issues with submission, but just in case, I emailed NSF technical support, who after five days eventually got back to me with the following:

Thank you for contacting the Fastlane Helpdesk. Regular document text needs to follow the font and font size requirements specified on https://www.nsf.gov/pubs/policydocs/pappg19_1/pappg_2.jsp#IIB2. The formulas aren’t restricted by these requirements (including inline formulas). Formulas that trigger the warning messages can be ignored as only some latex formats are currently supported.

This response is honestly not that useful given that the error message doesn’t indicate where in the file the problem is — perhaps the use of \boxtimes earlier on is acceptable whereas the fact that the entire Broader Impacts section is written in Zapf Dingbats is not?

This entry was posted in Mathematics, Uncategorized and tagged , , , . Bookmark the permalink.

12 Responses to NSF Application Tips: LaTeX edition

  1. DS says:

    Did the ‘intellectual merit’ and ‘broader impact’ subheadings in the results from prior support section also have to be on their own lines, or just the headings for the two main IM/BI sections?

    • Persiflage says:

      Don’t forget the s! (“broader impact*s*) My prior support sections are as follows and work ok:

      \section{Results from Prior Support: Intellectual Merit}
      \section{Results from Prior Support: Broader Impacts}

  2. MS says:

    I guess this is what those of us outside the US have to look forward to in a few years.

  3. John says:

    I have a naïve question: what kind of file is uploaded (.pdf or .tex)? My memory says I uploaded .pdf files in the past, but your first sentence scares me that they changed the rules (or my memory has gone!).

  4. just different says:

    Not to be heretical, but doesn’t all this make word processing your application documents seem more attractive? Is it still true that people will think of you as “not serious” if you do this?

    • Persiflage says:

      First question: it does not. Second question: it’s very hard to separate causal and correlational factors since it is rare nowadays to come across mathematics in a format other than LaTeX (and its relatives) that is not nonsense. But I feel confident that a great proposal would be funded regardless of whether it is in LaTeX or not.

      • just different says:

        If “nowadays” means in the last five years and not the last fifty, word processing has improved mathematics handling a lot and other options like Markdown + MathJax are readily available. Journal articles fit for publication are in LaTeX because that is (sensibly) the standard for publication, but it might be time to start bucking that expectation for documents that aren’t meant for public presentation.

  5. Drifter says:

    Thanks for sharing the info!

    I am running to the same problems. How did you manage to fix the font type(s) error?

    I am uploading only the summary of the project, which doesn’t include any math formulas. The following is the error:

    “Your file contains font type(s) that are not supported. For details about font types currently supported in the system, please refer to the Proposal Preparation Document Upload FAQs. NSF may return a proposal without review if it includes font types other than those specified in the Proposal Preparation Site Document Upload FAQs.”

    • Persiflage says:

      I didn’t resolve the warnings. But since I could localize it to LaTeX which was clearly not “supposed” to be a problem I’m not worried about it. Moreover, a program director offered the following feedback:

      If FastLane accepts it, we will review it, unless something is *really* wrong & non-compliant.

      So I would trust your judgement.

  6. Stumped Me For a Bit says:

    I recently (not with NSF) had a similar experience.

    I had a math expression in the paper’s title, say with a variable $x$, and LaTeX for some reason capitalized this on the even-numbered pages in the running title (but not odd-numbered). Somehow the tech-specs software noted the discrepancy (are the old school marms who measured my dissertation margins 2 decades ago now unemployed by tech advances?), and of course didn’t give any specifics as to what the problem was.

Leave a Reply

Your email address will not be published. Required fields are marked *