(You might drop the paper, fix the basic idea, or in some casespick . If youdrop it, the writeup will still come in handy in a year or two whenyou realize that the idea could be applied elsewhere.)
Assuming the idea is indeed a good one, then writing a draft makes yousharpen the message of the paper. Then you can figure out what workneeds to be done to support .
Honest writing may lead you to realize that proving your pointrequires more work than you'd thought (which is why you'd better writeearly). Remember RichardFeynman's on how to do honest science: "The first principle is thatyou must not fool yourself—and you are the easiest person tofool."
Writing first also helps you gauge the scope of the paper. You maydiscover that you can't fit your intended presentation into the pagelimits. That may cause you to divide up the work into two or morepapers, each of which has to make a separate clear point. So it willgreatly affect your planning.
Most readers decide whether they like a paper before they get tothe experimental results. Show people a draft of the first 2-3 pagesbefore you do any experiments. If they tell you that your basic ideawill evoke skepticism or yawns from reviewers, it's better to knowthat you waste your time.
Of course, neither coding nor research is purely top-down—in practice, there's feedback. Just as coding will make you rethink parts of the spec, certainly experimentation will make you revise parts of the paper. But crucially, you'll keep the code and the paper in sync.
But let's assume you are aiming for both clear writing andcomplete experiments. Suppose you have to finish the paper onno sleep. Which would you rather have left for the last minute,writing or hacking?
The latter case is more frequent—and there, the writeup the research (just as in History or Literature). The ideasin your paper will live on if they get picked up and used in othersystems. They have to be clear and convincing.
Aren't the experiments important too? Yes, but mainly to persuadepeople that the ideas are worth picking up, and to illustrate how theywork out in practice. They merely demonstrate whether your methodworked as you claim. Certainly experiments take a ofwork—and you should be careful, painstakingly honest, thoughtfulin your analysis of results, and willing to give out your code. Butother people will often prefer to write their own code based on yourwriteup. Your code usually dies; your writeup needs to live.
Clear motivation and exposition are more important thanresults for getting your paper accepted. If you run out of time, itis better to have a great story with incomplete experiments than asloppy draft with complete experiments. A good paper builds its casewith the accumulated weight of several experiments, so missing a fewis not fatal (and you can finish them for the camera-ready version).But a confusing, unconvincing, or incomplete writeup is fatal.
Of course, the experiments are important in theirown right if the point of the paper is to disprove the existence ofProto-Indo-European or demonstrate that MySpace users are functionallyilliterate. That's actual science where the experiments tell yousomething about the world.
A draft is concrete, visible, and understandable by others, so itis something that you and others can discuss, debate, and improve. Ifyou have a paper draft early, then you can give it to other people(including me) for feedback. And we can improve the paper togethervia a git or svn repository and its issue tracker.
White papers are great tools for generating credibility. Customers respond better to informative write-ups than they do to blatant ads. The trick is to make sure your white paper is organized and well thought out so that you will create a natural and genuine interest in your services.
If you run into trouble while doing the work, then I may havedifficulty diagnosing or even understanding the problem you arefacing. We may waste a whole meeting just getting aligned. But if youcan show me a precise writeup of what you're doing, then I can be muchmore helpful. In general, meetings are very productive when we have aconcrete document in front of us.