It's time to reinvent the wheel - a plead for disruptive changes in the file format department
- Speaker(s): Matthias Köfferlein
- email: matthias@klayout.de
Downloads
Slides (FSiC2025_Matthias_Koefferlein.pdf)
Abstract
EDA is a slow innovator. There are a number of reasons. Beside a general conservative attitude, there is little interest in disruptive changes by the EDA vendors. Why invest, when you can safely harvest your profit by locking your customers into your ecosystem?
Open Source does not suffer from this attitude. We are free to re-invent, with interoperability, extensibility and free access in mind. There are many drivers for such a venture: technologies, algorithms, emerging markets, the education sector and - in particular - interoperability. Files and - and in a broader sense – design databases are crucial for the interoperability of tools.
While the least common denominator for such databases are the “established” file formats, their interpretation and implementation often accounts for a major part of the tool development efforts. In practice, we see partial or buggy implementations, incompatible extensions and abuse of file formats for purposes different from their original ones. I don’t claim to make an exception in my case.
Solutions - like OpenAccess - exist, but not in a way compatible with the FOSS ecosystem. In parallel, modern schema-based serialization technologies emerged, that provide highly efficient alternatives to customized (binary) storage formats and reduce the implementation effort to a minimum.
With this talk I want to explore some options we have today, for driving a disruptive change in the full custom layout domain, potential benefits and ways of integration with a design flow and the tapeout processing chain.