First, I agree with @kibitzer that the survey itself is likely to have suffered from significant selection bias, in that the respondents were gathered based on ads on Stackoverflow. I'd bet the percentage of people who said "no" to "are you interested in continuing developing on the Salesforce platform" would have been much lower if the respondents were gathered from salesforce.stackexchange.com.
That said, even with the SO bias, Salesforce scoring less than the likes of Visual Basic, Wordpress, and Sharepoint should be cause for concern among the developer evangelism team and general management at Salesforce. At the very least I'd suggest it warrants a more targeted Salesforce developer survey that tries to determine if there is a significant developer community dissatisfaction issue, and what its roots are.
Personally, although I've spent close to 10 years in the Salesforce ecosystem, I definitely find frustrations in the platform as a developer that don't exist in other programming or cloud platforms I have used. I'd love to see Salesforce close those gaps.
If I was able to prioritize Salesforce development team dollars with developer joy in mind, my priorities would be:
- give the community a set of development tools that include a robust and officially supported IDE, tight source control integration, continuous integration support, and better support for multi-developer/multi-team environments (e.g. a java-style folder/package structure for Apex/VF, sub-org/multi-org hierarchical environments)
- take a torch to the platform limits. I understand there are architectural reasons for many of the limits, but there are too many of them, they are a constant frustration to developers, and this aspect of the platform is significantly behind other cloud platforms like AWS. The "Limits Quick Reference Guide" is 47 pages long.
- this is probably a pipe dream, but providing a local execution emulator that doesn't require server round trips for saving Apex/VF, running tests, etc would be AWESOME and clearly is technically feasible. Even if it had significant limitations on what it could emulate.
- provide an on-platform data store that doesn't guarantee the ACID-style robustness of the core SOQL data store, but that is fast, cheap, and capable of solving big data problems. Something like DynamoDB but that is integrated into the platform such that it can augment and/or join with standard SObjects/UI elements with a much lower cost per GB of storage.
- expand the exposure of the core Java APIs into Apex. Binary file manipulation, image/sound APIs, and core utility APIs would be awesome. Thread.sleep would also be nice :)
Anyway, I think it could be treated as fact that the numbers in the SO survey have numerous flaws; that said, being at the bottom of the heap in any survey is never something you want to see. Hopefully it gets noticed within HQ and we developers get rewarded with some moves up the priority list for developer-centric features!