Note to self: when submitting NSF proposals through Fastlane -
if you do a last minute revision then run LaTeX twice before generating the new PDF file, or the bibtex cross-cites break... ugly.
Apparently the panel really hates that. Really, really hates that.
Never had such a unanimous referee response.
Doh. What a bummer. It's so hard to get those damn grants anyway, and when something dumb like this also gets in the way....
When I was in grad school, I wrote a perl script called "laytecks", which would run latex, then bibtex, and then run latex as many times as it took until the .aux file was identical from one run to the next. It may have run it one extra time just to be sure.
It is possible in a long document for accumulated size changes as a reference gets inserted to move things from one page to the next, and for it to take more than two runs for things to stabilize. Rare, but I think I saw it happen at least once back during my thesis days.
-Rob
Ouch. Kinda like when referees kinda get pissed when you accidentally send them the wrong version of a paper draft.
This is like traveling--you should forget one obvious thing so you don't forget the subtle crucial thing.
Oh, man, that REALLY sucks.
Yea, I have my little alias (latexem) to run latex, bibtex, and latex again twice.
Thanks, you've added a new thing for me to be paranoid about whenever I submit something.
Although like nearly everyone else here, I have my unhelpful description of what I do. I only use LaTeX through a Makefile that runs it and bibtex a whole bunch of times. Plus a perl script that I use to package things up and do some tests before submitting to places (cause for some reason submitting to the arXiv makes me nervous about making mistakes).
yeah well, I was using TeXShop, working in Europe from my laptop...
it was late evening, I think I added one more "cover my butt" reference, reran latex and the PDF looked good - but of course that was the old PDF, never saw the new PDF with mismatched aux files and all the cites missing after page 3 - or so I'm told.
Having seen that from the other side I thought the panel was a wee bit harsh in its irritation, but I cannot disagree that it was a "hasty" proposal.
Of course with the success rate well below 20% in this round, any excuse to reject, even the most brilliant proposals, would be seized upon.
What is really irritating is that it was a good idea, but another group has moved in on it, they already have people ready to go, and they're going to use our code...
so time to walk away from that particular idea and figure it will be done by the time I get to another cycle.
Bugger.