The last couple of days I’ve been working directly from computer notes, either elicitation or Laves story correction. I’m trying to work out if I like it, and if the disadvantages outweigh the advantages (and/or vice versa :) ).
On the one hand, it’s nice to enter sentences directly onto computer without having to retranscribe. And since I don’t have a printer, if I was preparing “printed” English prompts, they’d be printed by hand, not from computer. And I have a nice laptop with a very clear screen and we blow up the text nice and big so it’s easy to read.
I don’t think I like working this way, though. I’m in the habit of writing while the speaker is talking – that’s fine if it’s pencil on paper, but the keyboard is more noisy. If I prepare sentences in toolbox I can’t see what the corrections are (and track changes makes things a bit clumsy too sometimes), whereas crossing something out is easier to read. Scrolling up and down is a bit off-putting, I think. And I have an irrational fear that the notes will just ‘disappear’, whereas fieldnotes in real notebooks are a bit more tangicle. (It’s symptomatic of the irrationality of this fear that it does not apply to Bardi digital recordings, all three and a half copies of which, I must admit, are still on the same damn table. I think I preferred my Yolŋu Matha PDA setup for some reason (although it had all these disadvantages, plus the inability to display text in the proper font).
It works a bit better for Laves text proofing, where the lines are in different colours and the work is more taking notes of variants and correcting Bardi sentences.
Other people have been swearing by their field computers for years, though, so maybe I’m missing something.
Anggarrgoon 
For me it depends on the type of fieldwork session. For certain elicitation tasks it sure is a lot easier to input directly into the computer. I’ve automatized the task to some extent by putting a Word template on my desktop called ‘New Fieldwork Session’. If I doubleclick the icon, a new document is opened in which today’s date is automatically generates as well as a few lines to quickly enter metadata (assistants, type of task). It’s a quick and easy procedure and it prevents me from forgetting to enter the relevant metadata. (In my paper notebooks, on the other hand, I sometimes forget the write down the date.)
In special lexical elicitation sessions I often input vocabulary directly into the database. This is less risky than it sounds, as Fieldworks Language Explorer has a few wonderfully streamlined procedures for this.
But hey, you can’t always walk around with your laptop, and a sturdy Moleskine paper notebook nicely fits into the back pocket of my trousers. I always take it with me.
“as Fieldworks Language Explorer has a few wonderfully streamlined procedures for this”. True, but it also crashes fairly frequently, feeding into my data eating nightmares…
I’ve had some nightmares like that last year, when I was using Fieldworks 4.0 (or something) which was very crash-prone. Even then, however, I found the data to be safe because of the particular set-up Flex uses (storing everything in an SQL Server database).
From version 5.0 on, the crashes have disappeared for me and FW is extremely stable. The newest version is 5.2.