At 10:41 AM 2/15/2002, Dimitrie O. Paun wrote:
Wine is a _collection_ of products, just like a Linux distribution. Which means that the viral aspect of the LGPL _stops_ at every DLL boundary.
Alas, it does not. First of all, there's the virulence that arises when a programmer of commercial software is "contaminated" by reading the existing code. The programmer cannot run the risk of reading the code or learning from it. Any improved DLL would have to be COMPLETELY reimplemented -- a highly wasteful and needless duplication of effort. Because this extra labor and expense will tax vendors' businesses, they will not have as many resources to contribute back to the project.
Nor will they be able to hire consultants such as Jeremy to help them with their products, because Jeremy will have seen the original source and thus cannot write commercially licensed DLLs for them. CodeWeavers will lose clients to consultants who have not seen the original code.
The contamination problem also means that any programmer of commercial products CANNOT debug the code and contribute fixes. He or she must treat it as if it were closed source. This hurts both the programmer and the project.
Finally, because the (L)GPL makes product differentiation at best and often infeasible or impossible, commercial software vendors that attempt to base products on it will not be viable or financially sound. Consultants such as Jeremy will find that they are often not paid due to business failures and will hence experience problems making their own payrolls. As a result, they, too, will not be able to contribute as much, hurting the project. Everyone loses.
In short, the (L)GPL will accomplish its stated goals: to put commercial software vendors out of business and to prevent programmers (including CodeWeavers!) from being adequately compensated for their work.
--Brett