From practical perspective we would be better off with Web SQL DB. Functionally it addresses everything a developer needs to create offline HTML apps. API is not perfect but you can always improve that with small wrapper libraries (like JQuery improves DOM API). You can always easily implement IndexDB-like API on top off SQL if you really want that. The opposite is probably possible but much more difficult.
Also I think IndexDB might turn out to be yet another compatibility headache because everybody will be creating their own implementation (I'm pretty sure there will be at least 3 - Microsoft, Mozilla & Webkit). It's a speculation but most likely this wouldn't be an issue with Web SQL DB because everybody would just use the same SQLLite backend.
And worst of all, this is delaying progress of web applications. Lack of solid cross-browser client-side storage solution is the last big gap for next generation of web applications. If everybody had just agreed on SQLLite we could be there today... now it's still unknown when we will get there.