The project I'm currently working on has a phased release. This is a decision the business has made and they intend to launch a mobile site that caters to the majority of mobile use cases, but there are some journeys unsatisfied in this release. Eventually the entire site will be mobile optimised but until then,the solution is to where features/content is missing we direct users to the desktop site. I'm unconvinced by this approach. We've worked hard to optimise the site for key user journeys and I think linking users out to the desktop site where they will get a horrid experience isn't what we should be aiming for. I would rather keep the mobile site limited and wait till future releases to make those options available? I'm wondering what the best approach is?
If you have a separate site for mobile and non-mobile then yes, most definitely yes, let them go to the full site to complete what they wish if it isn't offered on your mobile one.
Mobile doesn't just mean 'people sat in a coffee shop for 2 minutes looking at their iPhone'. These days Mobile means 'anything that isn't a laptop/desktop' (and even that line is a bit blurry with chomebook, MS Surface and similar devices nowadays).
Really I would question the logic behind actively deciding to prevent people from completing their tasks just because they're using one device compared to another anyway, If you can think of any valid reason where someone should be prevented from completing a task just because they are using one device instead of another then sure, go ahead and stop them doing it, but I'd wager the only reason you'd come up with would be 'we don't think people on a mobile would want to do this task', and that isn't a valid reason.
Possibly the only valid reason for no allowing it on mobile is the cost of development / support for doing so. And therefore that brings it back to your main point; providing a link to the desktop version. If you can't offer the option bespoke in your mobile site then yes, you should allow them to complete it by going to the main site to do so.
A poor user experience that allows someone to complete a task is preferable to a non-existent user experience that means they can't do anything at all.
You could also have some analytics in place to see how many people come to those sections on a mobile device to take that to the stakeholders and say 'look, X% of visitors are trying to do this on a mobile, lets make it even easier for them and optimize this area too' as it gives you scope for future work and helps you prioritize what sections of the desktop are in higher demand for mobile users than others.
Heck, people even use the US Passport application site to apply for a passport while using a mobile device, (that isn't optimized for mobile) and I can't imagine any more painful task than that to complete on a mobile, but people do.
I think there are two points here:
Get something out there: You can launch a mobile site with limited user journeys as long as those journeys are complete and standalone. If you are following a lean approach you should not need to wait until a site is fully developed as there may be a lot of value in delivering core journeys before then.
Give the user the choice: I don't think there is anything wrong with providing a user with the option of using the full-site should they wish to where doing so will add significant value. For example, the banking mobile site allows me to sign-in on a mobile friendly version and then use the full site to do everything else. Clearly for a banking mobile site this is a core piece to the user journey so is justified and allows me to mobile bank. Where there is a journey which sits outside of the core set of journeys (e.g. changing password, or adjusting some settings) then these I would generally opt out of including all together. In both of these cases, providing clear messages to the user is the important thing so they can decide what they would like to do.
I have the same experience as you and I agreed with you that linking out to desktop site only result in bad UX. Best approach is to not launch the mobile site till it's fully developed or to a phase where it's capable to fulfill at least 80% of the use cases. I would very much preferred a full desktop site than a mixture of both.
As long as you're not redirecting users to the homepage of whatever version of the website you're serving them, you're better than 90% of webpages, and don't deserve to be in that special hell.
I'd say link to the desktop site, but add a warning that the user is leaving the mobile-optimised part of the website.
With a decent website structure, the desktop site should be perfectly usable on most mobile devices.