Abstract: The Y2K problem has analogies in the mask-making world. With the Y2K problem where a date field has just two bytes for the year, there are some cases of mask-making data in which the file size cannot exceed 2 gigabytes. Where a two-digit date field can only unambiguously use a limited range of values (00 to 99), design coordinates can only cover a range of about 4 billion values, which is getting a little uncomfortable for all of the new applications. In retrospect, with a degree of foresight and planning the Y2K date problem could have been easily solved if new encodings had been allowed in the two- digit field. Likewise, in the mask-making industry we currently have the opportunity to achieve far superior data compression if we allow some new forms of data encoding in our data. But this will require universal agreement. The correct way to look at the Y2K problem is that some information was left out of the data stream due to common understandings that made the additional information superfluous. But as the year 2000 approaches, it has become widely recognized that missing data needs to be stated explicitly, and any ambiguities in the representation of the data will need to be eliminated with precise specifications. In a similar way, old mask data generation methods have had numerous flaws that we have been able to ignore for a long time. But now is the time to fix theses flaws and provide extended capabilities. What is not yet clear is if the old data generation methods can be modified to meet these developing needs. Unilateral action is not likely to lead to much progress, so some united effort is required by all interested parties if success is to be achieved in the brief time that remains. !0
展开▼
