To quote from Sharon Lamb's position paper: The "grand challenge" for ESP developers is to provide a complete distributed system: hardware, kernel, compilers, tools etc. such that programming distributed...
详细信息
ISBN:
(纸本)0897915445
To quote from Sharon Lamb's position paper: The "grand challenge" for ESP developers is to provide a complete distributed system: hardware, kernel, compilers, tools etc. such that programming distributed computers is convenient and makes optimal use of resources. This workshop, I believe, showed that the challenge has also been taken up by others across industry and academia. It is clear that issues of scale and distribution have ramifications in a range of sub-disciplines of computer science, from software engineering to language design, influencing everything from operating system mechanisms to object models. Although it is hard to draw a single point out of the web of inter-related material presented during the day, perhaps one repeated point is worth noting. Orthodoxy in O-O work has it that 'everything is an object', providing a simple unifying model. However, throughout the presentations it was clear that there is a need to explicitly recognise the changes that occur as one moves up through a system: from hardware to o/s abstractions, to multiple address spaces and distributed systems (which exhibit complex failure modes) and then finally to a global vision of cooperating autonomous establishments using highly interconnected and interactive, yet widely distributed, applications. Identifying the mechanisms that support the transitions from one layer of this view to the next, providing efficient and effective implementations when low-level design decisions have major ramifications at the highest levels, and building applications at each of these scales, will occupy us for some time to come!
暂无评论