Blaise: A Modern, Zero-Legacy Approach to Object Pascal

The Object Pascal ecosystem has long been defined by a tension between stability and modernization. For decades, developers have relied on Embarcadero Delphi—a powerful but proprietary, Windows-centric tool—and Free Pascal (FPC), a robust open-source alternative that, while impressive, carries significant legacy baggage. Enter Blaise, a new self-hosting Object Pascal compiler that aims to provide a "clean slate" implementation of the language.

By targeting the QBE backend and focusing on a zero-legacy philosophy, Blaise attempts to solve the long-standing "warts" of Pascal while maintaining the performance and readability that made the language a staple of early software engineering.

A Zero-Legacy Philosophy

Blaise is not intended to be a drop-in replacement for existing Delphi or FPC projects. Instead, it is a modern reimagining of Object Pascal. The primary goal is to eliminate the "cruft" that accumulates over decades of language evolution. This includes a move toward more sensible defaults and a streamlined memory model.

Key areas of focus for the project include:

• Modern Defaults: Implementing UTF-8 strings as the standard, moving away from the complex string handling of legacy Pascal.
• Inline Variable Declarations: Addressing a common complaint in the Pascal community—the requirement to declare all variables at the beginning of a function—by allowing declarations anywhere in the code.
• Simplified Interfaces: Moving away from complex COM-style interface GUIDs to reduce architectural overhead.

Technical Architecture: QBE and Self-Hosting

One of the most significant technical choices in Blaise is its use of the QBE backend. While LLVM is the industry standard for many modern compilers, it is often criticized for being heavyweight. QBE serves as a lightweight alternative, allowing for faster compilation times and a simpler toolchain.

Furthermore, the project is designed to be self-hosting. This means the Blaise compiler is written in Blaise itself. While this creates a bootstrapping challenge—initially requiring another compiler like FPC to get the first version running—it ensures that the language is capable of expressing its own complex logic and that improvements to the compiler directly benefit the language's development.

Community Perspectives and Critiques

The announcement of Blaise has sparked a range of reactions from the Pascal community, reflecting the diverse needs of developers who use the language today.

The Value of Experience

Some observers were initially skeptical of the project's rapid early development. However, others pointed to the pedigree of the author, Graeme Geldenhuys, as a sign of the project's viability. Geldenhuys is the creator of the fpGUI library, a project that has been maintained since 2010. As noted by community member @ziotom78:

"He surely has a lot of experience with FreePascal and the Pascal language... this project seems to start on good grounds!"

Missing Features and Trade-offs

No "clean slate" project is without its trade-offs. Some developers have expressed concerns over the omission of certain legacy features. For instance, the with statement—a controversial but convenient feature in traditional Pascal—is absent in Blaise. Additionally, some users noted that the lack of a built-in GUI framework makes the language less attractive for those who primarily use Pascal for rapid application development (RAD) of desktop apps.

Ecosystem Gaps

While the project focuses on the core compiler, the broader ecosystem remains a point of discussion. Some users pointed out that the current landscape includes more than just Delphi and FPC, such as Oxygene, though it remains proprietary. The challenge for Blaise will be to build a library ecosystem that makes the language viable for modern production use beyond simple CLI tools.

Conclusion

Blaise represents an ambitious attempt to bring Object Pascal into the modern era by stripping away the legacy constraints of the 1990s. By combining a high-readability language with a lightweight backend like QBE, it offers a compelling alternative for those who want the performance of a compiled language without the baggage of a 30-year-old toolchain. Whether it can gain enough traction to build a sustainable ecosystem remains to be seen, but its commitment to a "zero-legacy" approach provides a refreshing direction for the language.