Slashdot is powered by your submissions, so send in your scoop

 



Forgot your password?
typodupeerror

Submission + - C++ Commitee Prefers Bjarne Profiles Over Baxter Rustification

robinsrowe writes: No surprise, the C++ Committee is still trending toward C++ Profiles. It would have been a huge change had the Committee embraced Baxter's Rustification memory safety proposal. Would mean banning pointers. Making the C++ language much like Rust would deeply break every C++ program in the world. Article at TheRegister: “Rust-style safety model for C++ 'rejected' as profiles take priority” https://www.theregister.com/20...

The C++ standards committee abandoned a detailed proposal to create a rigorously safe subset of the language, according to the proposal's co-author, despite continuing anxiety about memory safety.

Article at Le Monde (in French): “The C++ standards committee rejected a proposal to create a secure subset of the language. Members prefer to focus on the Profiles framework pushed by C++ creator Bjarne Stroustrup.” https://www.lemondeinformatiqu...

"If you mark your code to apply a Profile, some features of the C/C++ language will stop working," he says. There is also a small problem, these guidelines were not integrated into version 26 of C++, but simply into a white paper. The controversy surrounding the security of C++ opens the door to another solution with the use of another language. The first advocated by several American authorities is Rust, but there is also Google's experimental Carbon project. Unveiled in 2022, it also aims to modernize C++.

If Profiles are eventually adopted, it may Balkanize C++ by dividing C++ into safe and unsafe subsets. C++ Profiles won't fix the issue of making C pointers memory safe. A proposal to implement pointer memory safety is TrapC, but for the C language, not C++. Some say make the switch to Rust, but that doesn't solve the safety problems lurking in billions of lines of existing C/C++ code.

Submission + - Math Genius AI to Co-author Proofs within Three Years (theregister.com)

robinsrowe writes: How must faster will technology advance with AI agents solving new mathematical proofs? AI today isn't very good at math. Vividly demonstrated recently, when the White House used AI to calculate "reciprocal tariffs" that made no math sense whatsoever. (AI doesn't know the math difference between a tariff and a deficit.) That AI today cannot mathematically reason is a rich source of AI hallucinations. DARPA, the research arm of the U.S. Department of Defense, aims to make AI math be much, much, much better. Not merely better at calculations, but to make AI do abstract math thinking. DARPA says that "The goal of Exponentiating Mathematics (expMath) is to radically accelerate the rate of progress in pure mathematics by developing an AI co-author capable of proposing and proving useful abstractions." Article in The Register...

The US Defense Advanced Research Projects Agency, aka DARPA, believes mathematics isn't advancing fast enough. So to accelerate – or "exponentiate" – the rate of mathematical research, DARPA this week held a Proposers Day event to engage with the technical community in the hope that attendees will prepare proposals to submit once the actual Broad Agency Announcement (BAA) solicitation goes out. Whoa, slow down there, Uncle Sam. DARPA's project, dubbed expMath, aims to jumpstart math innovation with the help of artificial intelligence, or machine learning for those who prefer a less loaded term.


Comment Re:What problem does this solve? (Score 1) 99

Thank you to Slashdot for posting about my work on TrapC. A friend who is on Slashdot daily, and was a reviewer on my TrapC whitepaper, wrote me to say, "the comments are unusually on-topic". Appreciate the 97 Slashdot comments. Added to the many helpful comments from reviewers and that were raised last week during my ISO C Committee N3423 presentation. While not everyone seems thrilled, the response more positive than I'd anticipated. Constructive criticism super helpful, appreciate you! As Bjarne Stroustrup has said, the most important thing in software design is to be clear about what one is trying to build. Pr. Robin Rowe TrapC Creator https://www.reddit.com/r/trapc...

Slashdot Top Deals

The IBM purchase of ROLM gives new meaning to the term "twisted pair". -- Howard Anderson, "Yankee Group"

Working...