Should LINQ to SQL Go Open Source
More and more, LINQ to SQL is being seen as a temporary solution. With the impending release of ADO.NET Entity Framework, a lot of people are worried that development on LINQ to SQL will fall by the wayside. With Microsoft's long history of developing and discarding database access technologies, these concerns are not without merit.
Further compounding the issue is that LINQ to SQL has been transferred to the SQL Data Programmability team, the very same group working on ADO.NET Entity Framework. With their own project filling a similar role, it is hard to see them wanting to devote a lot of time to the adopted project.
So this raises the question, "Should LINQ to SQL Go Open Source?" Leon Bambrick asks that question and provides some analysis. One of the biggest concerns he raises is the liability issue, wherein Microsoft may be reluctant to ship something with the framework that has had external contributions. On the other hand, it might be what's needed to push through community-driven features such as mockability and multiple providers.
Not completly temporary
LINQ to SQL serves its purpose pretty well seen from that angle, so its probably fine.
LINQ to SQL was never even supposed to exist
My biggest objections to LINQ to SQL are following:
1. LINQ to SQL works only with Microsoft SQL Server
2. LINQ to SQL is that it is no 100% POCO. Framework forces you to use System.Data.Linq.EntitySet type on "many" side of one-to-many relationship and EntityRef on the "one" side of one-to-many relationship.
3. Single table inheritance
These are not the only ones, but you have to give them some slack since this is the first version of the product.
Some interesting background on LINQ to SQL origins:
The LINQ to SQL was never even supposed to exist