The unplanned release of hydrocarbons from a pipeline can lead to significant pollution and/or the horrific potential of explosion and a very large fire,depending on the transported product.For this reason,the manufacturing procedure specification(MPS)developed to ensure the design requirements are met by the steel and pipe-making process is a critical element of the fracture control plan whose purpose is to protect the environment and ensure public safety.Pipelines transporting compressible hydrocarbons like methane or high-vapor-pressure liquids at supercritical conditions are uniquely susceptible to long-propagating failures,which opens to complications in both the MPS and the procedure qualification testing(PQT),which are compounded for designs that involve higher pressures and larger diameters.This paper considers pipeline steel specification with particular concern for long-propagating shear failures in advanced-design larger-diameter higher-pressure pipelines made of thinner-wall higher-grade steels.Assuming that the arrest requirements can be reliably predicted it remains to specify the steel design,and its MPS.While standards exist for use in PQTs to establish that the MPS requirements have been met,very often CVN specimens remain unbroken,while DWTTs lead to features that are inconsistent with the historic response that underlies many standards.In addition,sub-size specimens are often used,whereas there is no standardized means to scale those results consistent with the full-size response required by some standards.Finally,empirical models like the Battelle two curve model(BTCM)widely used to predict required arrest resistance have their roots in sub-size specimens,yet their outcome is widely expressed in a full-size context.This paper reviews results for sub-size specimens developed for steels in the era that the BTCM was calibrated to establish scaling rules to facilitate prediction in a full-size setting.Thereafter,issues associated with the use of sub-size specimens are reviewed and criteria are developed to scale results from such testing for use in MPS,and PQT,which is done as a function of toughness.Finally,issues associated with the acceptance of data from unbroken CVN specimens are reviewed,as are known issues in the interpretation of DWTT fracture surfaces.