It is often said that open source projects will "win" or "lose" based purely on technical merit. Experiences from the LSB Project's interface standardization efforts indicate there are some concrete steps an open-source project producing interface libraries for general use can take to make the project more usable for a wider audience, leading to greater chance of widespread acceptance. Such projects have a reasonable chance of becoming standards, whether de-facto or by inclusion in formal specifications such as the LSB.
The evidence is that projects ready for large-scale use typically meet most of a set of criteria that include: demand; stable, well-documented interfaces; comprehensive interface and regression tests; an easily-deployed (portable) working implementation; and an appropriate choice of license. With the exception of demand, most of these criteria can be consciously worked towards. The paper will present some case studies of libraries that have successfully been incorporated into the LSB specification. It will also discuss some tools the LSB has developed that may help in describing public interfaces and developing tests, and discuss some ways in which portability of the code base can be improved.