I can assure you that we thought long and about the issues you raise before going ahead with Glow, and we've had plenty of healthy debate internally along exactly the same lines.

A few things helped tipped the balance in favour of developing Glow, some of which I'll summarise quickly.

1. Glow is an evolution of several older libraries that we have developed over many years, it's not that we sat down one day and said "hey, why don't we build a JavaScript library, that'd be fun". We've been building libraries for our own use for years, this is not a new thing for us, it's just the first one we've open sourced.
2. Continuing to develop our own library means that we have complete understanding of the software, there are no "black boxes" which we cannot open.
3. This means the core development team is able to provide a valuable internal support role for the other developers around the BBC. We are able to give help and training to other teams, ensuring they know exactly how to use the library (and JavaScript generally) in the most efficient way. Open source communities are great, we're not knocking them, but they are not always able to respond with the speed and level of detail that as an organisation we require.
4. Creating the library ourselves means we have been able to ensure it contains exactly the features that our business requires. Even if we did adopt another library, we would most likely still need to create the specific widgets and other bespoke features we need on top.
5. Controlling the release schedule also means we have been able to make prompt updates with bug fixes as and when required, without needing to wait for agreement from external project owners. This is crucial to maintaining the level of quality our audience deserves.

These sorts of issues (along with browser support, the cost of maintaining forks, etc) all contributed to our decision to keep JavaScript library development in house. It was not a simple decision, but one I am happy with.

Finally, please note that we are far from closed to adopting open source solutions in the future, where it makes sense. For example, we are strongly considering using John Resig's Sizzle CSS selector engine in the next major version of Glow. This would be relatively easy to integrate, solve a discrete task very well, and we would donate back any code that helped keep it in line with our browser support requirements. If it saves money and helps maintain quality, bring it on.

All the best,
Steve

But putting it that way sounds like it's an even proposition should they want to proactively protect themselves from a possible "worst case scenario," this in spite of the other issues that make controlling your own framework at a Very Large Corporation (like oh say Google or Yahoo?) a good choice, as was already discussed recently here http://news.ycombinator.com/item?id=696623 and here http://news.ycombinator.com/item?id=694380 .

And wouldn't the folks at [insert any other framework here] question why the "people being paid real money, taxpayers' money" at the BBC were spending their working days contributing to jQuery and not [insert any other framework here]? As it is now all frameworks can freely take or contribute to Glow under the Apache 2 license. To me that's the important point here.

It'd be really nice to see a major player with real money behind it contribute to an open source project.