I've taken down the WURFL survey now, after running for a couple of weeks and gathering 136 responses. The full results are here, but a summary of interesting points I've noted:

1. Only 21.3% of WURFL users surveyed have never contributed to it.
2. It's used in the vast majority of cases by application developers (perhaps because they tend to be smaller and can more readily admit to using it - we know of lots of large organisations who use the WURFL, but quite quietly).
3. Overwhelmingly it's used for browser-based services.
4. Half of those surveyed use not just the WURFL data itself, but one of the bundled APIs.
5. Nearly half of respondents patch it with their own data - but fewer than this contribute regularly: naughty!
6. 95% of users rate the WURFL quality as average or better. 65% give it 4 or more out of 5. The WURFL gets similar marks for quality of data.
7. If it wasn't for the WURFL, most folks would be using an in-house solution. So it's clearly saving lots of people time and money.

Thanks to all who responded, and to SurveyMonkey for hosting the survey.