Dehyped title: Gordon Bell, pioneering computer engineer and co-founder of major computer museums, dies at age 89.

Summary:

Gordon Bell was a pioneering computer scientist who played a key role in the development of influential minicomputer systems at Digital Equipment Corporation (DEC) in the 1960s and 1970s. He co-founded the first major computer museum in Boston, which later became the Computer History Museum in Silicon Valley. After retiring from DEC, Bell remained active as an entrepreneur, policy advisor, and researcher, including working on life-logging technologies at Microsoft Research. Bell received numerous accolades for his contributions to the computing industry, including the National Medal of Technology. Industry leaders remembered Bell as a brilliant, helpful mentor who had a profound impact on the evolution of technology.

Comments:

Gordon Bell, a prominent computer scientist and co-founder of the Computer History Museum, has passed away at the age of 89. Commenters share personal anecdotes and memories of Bell, describing him as a sharp, generous, and passionate individual who made significant contributions to the field of computing. The discussion highlights Bell's role in preserving and presenting the history of computing, as well as his continued engagement and curiosity about technology later in life. Overall, the thread celebrates Bell's remarkable life and legacy.

Here is a summary of the key comments, grouped by commenter:

jpgvm: I was fortunate to have Gordon Bell as the first investor in my startup. He was sharp, generous with his time, and lived up to the hype. I have fond memories of enjoying dinner with him and his wife, who made amazing cookies.

ChrisArchitect: The Computer History Museum has shared the news of Gordon Bell's passing. Bell was a brilliant computer designer and a major force in preserving and presenting the history of computing to the public.

workergnome: My father worked with Gordon Bell and told a story that the reason the ASCII bell character sequence is CTRL-G is because of Bell.

sctb: I remember watching a video about a decade ago where Gordon Bell spoke about his MyLifeBits project at the Microsoft Research office in San Francisco. I'm struck by his lifelong passion, curiosity, and engagement with technology.

dang: I enjoyed reading Gordon Bell's article "The No Output Division" from 1982, which was shared on Hacker News.

dcminter: Reading about Gordon Bell's dislike of "sewing circle" meetings resonates with me, as I attend many meetings in my work.

smarks: I am very sad to report the passing of Gordon Bell, a famous computer pioneer and co-founder of the Computer History Museum. There will be a "Celebration of Life" service held in his honor in August.

Dehyped title: A software bug caused a period to disappear from emails sent to a specific customer, which was resolved by updating the SMTP client code.

Summary:

The document describes a software project where the team built a system to manage document templates for a client. After the system was in use, they received a report that some emails were missing a period in the body. After investigating, they discovered this was due to the way the SMTP client was formatting the email body to comply with the SMTP specification. The SMTP client would add an extra period at the beginning of any line that started with a period, and the SMTP server would then remove the first period, causing the period to disappear. The team fixed the issue in their code, but later learned another team had the same problem, leading to some customer emails missing a decimal point in important financial information. The team was able to quickly patch the issue once they understood the underlying cause.

Comments:

The discussion highlights several issues with implementing the SMTP protocol correctly. Commenters emphasize the importance of using well-tested libraries instead of implementing SMTP functionality from scratch. They also caution against not keeping dependencies up-to-date, as that can lead to bugs even when the underlying issues have been fixed upstream. The discussion also touches on best practices for handling protocol specifications and the various quirks and edge cases that can arise when dealing with email standards.

Summaries by contributor:

teddyh: Implementing standards haphazardly is a bad habit. Dependencies should be regularly and promptly updated to avoid issues like the one described.

evmar: The real problem is that the code implemented an SMTP client directly instead of using a proper library. Implementing protocols correctly is hard, and bugs like this are common.

quotemstr: It's best to specify protocols at a high level, like using cap'n'proto, instead of expecting every developer to correctly implement a network exchange.

blueflow: You have to thoroughly read and understand the protocol specifications. Brute-forcing until "it works" is not an acceptable engineering approach.

jeffbee: If someone hasn't heard of "dot stuffing", they're unlikely to be aware of the many other horrors that exist within the email space.

jiveturkey: Automatically splitting a monetary value across a numeric decimal indicator and splitting lines for this use case seems questionable.

pwg: The period disappeared because it was placed at the start of the next line when the split occurred due to the 1000 octet maximum line length requirement.

Aloisius: If the period was the 999th character in the line, it would split it to the next line since the maximum line length in SMTP is 1000 characters including CRLF.

marcosdumay: Implementing basic SMTP functionality is challenging, let alone handling all the SMTP craziness that isn't covered in a short introduction to the protocol.

Dehyped title: OpenAI faces backlash over using voice similar to Scarlett Johansson's without her consent for its AI assistant 'Sky'.

Summary:

The article discusses how Sam Altman, the CEO of OpenAI, is being scrutinized for his actions and management of the company. OpenAI recently faced backlash from actress Scarlett Johansson over the voice assistant "Sky" in their ChatGPT 4.0 system, which she claims sounds too similar to her own voice. This incident, combined with reports of Altman's aggressive tactics, secrecy, and mistreatment of employees, is revealing his true character to the public. The article suggests Altman's public persona as a responsible AI leader does not align with how he privately runs OpenAI, marked by lawsuits, employee departures, and questionable business practices. Overall, the text portrays Altman as an ambitious and Machiavellian figure whose actions are undermining public trust in both him and OpenAI.

Comments:

panarky: The commenter believes Sam Altman is being sneaky and intentionally deceptive in his statements about the voice actor for Siri. They suggest we should not give such a deceptive person power over the future.

bbqfog: The commenter believes Sam Altman's sister accused him and his brother of raping her when she was a child, and that people are now shocked to learn he is "unethical" despite this history.

yieldcrv: The commenter suggests that unethical behavior, like running a company into the ground and getting a bailout, is common in the San Francisco/Peninsula area, unless you are "doing Steve Jobs correctly."

ur-whale: The commenter expresses shock that a "full blown psychopath" is now in charge of the most influential piece of technology developed since the internet.

Dehyped title: Designing user interfaces for optimal value density across time and space.

Summary:

The article explores the concept of UI (user interface) density, which refers to the amount of information and functionality an interface can provide over time. It discusses three main aspects of density: visual density (how much content is displayed on the screen), information density (how much data is conveyed), and design density (how design decisions communicate meaning). The article argues that true density goes beyond just visual elements and includes the temporal aspects of an interface, such as load times and response speeds. Ultimately, the author defines UI density as the value a user gets from an interface divided by the time and space it occupies. Designing for density involves optimizing for speed, usability, functionality and other factors to provide the most value in the smallest footprint.

Comments:

This Hacker News thread discusses the trend towards sparse, minimalist user interfaces, especially on mobile devices. Participants share their frustrations with poorly designed responsive interfaces that break basic functionality. They argue that dense, information-rich UIs are often more useful for professional users, while sparse designs cater to the lowest common denominator. The discussion covers topics like UI density, Fitts' law, and the challenges of designing for both casual and power users.

Insightful contributors and their summaries:

onemoresoop: I'm frustrated with mobile apps that have broken functionality due to overly sparse and minimalist designs, which seem optimized for larger screens at the expense of usability on smaller devices.

DrScientist: I've noticed that many React-based interfaces don't scale well and often don't allow you to scroll to content that's off-screen. What are the common mistakes developers make when building responsive UIs with React?

tracker1: I agree with the issues around sparse mobile interfaces. For smaller screens, modal dialogs and menus should be designed to take over the full screen rather than forcing users to scroll or hunt for content.

jakubmazanec: The problems with responsive design have nothing to do with the React framework itself. Poorly implemented responsive design is the fault of the developers and their managers, not the underlying technology.

BugsJustFindMe: Good interface design should balance information density with visual salience, using techniques like whitespace to clearly delineate different information sets. Density alone is not the only factor.

The_Colonel: Encapsulating information with visual frames, as was common in 90s-era GUIs, can help provide both higher density and clearer visual salience.

Spivak: What's considered "salient" can vary greatly depending on the user and their level of familiarity with the software. Designs optimized for occasional users may not work well for power users.

graypegg: JIRA is a visually dense application, but its speed and the number of screens required to navigate it can make it feel sparse despite the dense visuals.

abeisgreat: I'd love to see a "density renaissance" in mobile UIs, where designers question some of the assumptions around sparse, minimalist designs, especially as trends like large language models move towards slower, potentially inaccurate outputs.

david_allison: Mobile apps are constrained by accessibility requirements around minimum touch target sizes, so we're unlikely to see a major shift towards denser mobile UIs.

sebastiennight: Faking the appearance of faster response times, by slowly streaming output rather than showing it all at once, can make an application feel more responsive to users even if the actual processing time is the same.

tracker1: You can achieve dense UIs on the web just as easily as in desktop applications. The key is

Dehyped title: Wikimedia Enterprise offers API access to Wikimedia project data for search engines and social networks to improve user experience and knowledge applications.

Summary:

Wikimedia Enterprise offers APIs that allow organizations to access and leverage the vast knowledge contained in Wikimedia projects like Wikipedia. The APIs provide access to the most current and accurate information, with daily and hourly snapshots as well as real-time data streams. Wikimedia Enterprise offers features like written agreements, SLAs, and support to make the APIs safe and reliable to use at scale. The APIs are well-documented and supported, making it easier for teams to implement them. Wikimedia Enterprise also has tools to help detect inaccurate or biased information in the content, reducing risk for users.

Comments:

s1k3s: Wikipedia knowledge is not true knowledge, as it is based on what the majority of people decide is true, rather than objective facts. This "democracy" of knowledge is problematic.

The Hacker News discussion centers around the limitations of relying on Wikipedia as a source of knowledge, particularly for training large language models and AI systems. The main critique is that Wikipedia's content is determined by community consensus rather than rigorous fact-checking, which can lead to biases and inaccuracies being propagated.

This raises concerns about using Wikipedia as a primary data source for advanced AI applications that require high-quality, reliable information.

Dehyped title: CADmium: An open-source browser-based CAD program exploring new approaches to constraint solving, b-rep kernels, history tracking, and file formats.

Summary:

CADmium is an open-source, browser-based CAD program being developed. It aims to address limitations of current CAD software by using modern approaches like a physics-based 2D constraint solver, the Rust-based Truck B-rep kernel, a history-tracking system, and a 3D user interface built with web technologies. The goal is to create a CAD experience that is familiar to users but with fresh ideas, including the ability to easily version and collaborate on designs. The project is seeking help from the community in areas like Rust programming, computational geometry, and 3D user interface development. While the focus is on the core CAD functionality, the team is also interested in exploring related areas like toolpath generation and finite element analysis in the future.

Comments:

The discussion centers around the challenges of open-source CAD software and the need for more user-friendly options. Commenters praise the work on CADmium, a new browser-based CAD program, and discuss the pros and cons of existing open-source CAD tools like FreeCAD and Onshape. There is also a discussion of Ondsel, a company building commercial offerings around FreeCAD. Overall, the thread highlights the desire for high-quality, accessible open-source CAD solutions.

Summarized by commentator:

gen3: As someone who learned CAD on Autodesk Inventor, I found Onshape to be the only free modeling software I could grasp after a few hours. Usability seems to be a real problem in this space, with FreeCAD having too many ways to do the same thing. I'm excited to see the work on CADmium.

JeremyHerrman: The summary of the CAD kernel landscape is excellent, especially the quote about OpenCascade being the "Pontiac Aztek of b-rep kernels."

ecjhdnc2025: I am a mostly happy FreeCAD user, though I agree with the framing around the usability challenges. I'm excited to get involved with the CADmium project.

sliptonic: It's great to see others investing time and talent into open-source CAD. At Ondsel, we believe open-source CAD is incredibly important and we're thrilled to see innovation in this area.

krastanov: I'm trying to understand Ondsel's organization and offering. Is it to FreeCAD what Codeweavers is to Wine?

rjsw: We spend a lot of time working on the correct type system for STEP, and I don't think using an untyped format like JSON is a good idea, even though ISO 10303-243 uses JSON.

peppertree: Plasticity was open source before it abandoned C3D in favor of Parasolid. The repo can still be found here: https://github.com/nkallen/plasticity.

daemonologist: Onshape is great, but it could go away or have its free version cut down at any time. Having an open source alternative that's less of a slog than FreeCAD would be awesome.

Dehyped title: Startup seeks full-stack developers to build open-source AI tools for developers.

Summary:

Continue is a startup that is building open-source IDE extensions and other AI software development system components. They are looking for outstanding full-stack developers who are excited about learning from users, obsess over details, and want to work in San Francisco. The startup has raised $2.1M from Heavybit and angels, and believes that large language models present an opportunity to dramatically improve the developer experience. Their goal is to enable anyone to slip into a flow state when coding and never get stuck. The founders are Ty Dunn and Nate Sesti.

Comments: Unable to generate summary

Dehyped title: A new method for storing tabular knowledge in a plain text format with advantages over existing binary storage formats.

Summary:

The document proposes a new way to store all tabular knowledge in a single long plain text file using a simple syntax of spaces and newlines. The key concepts are measures (defined by parsers), concepts (defined by measurements), and unlimited comments attached to measurements. This allows all tabular knowledge to be stored in a computable form that can be read and written by humans, software, and AI. The system is called ScrollSets and is open source, with the PLDB.io website currently using it to store over 300 measures, 4,000 concepts, and 150,000 measurements. The document suggests this method may have implications for the weights and training of artificial neural networks.

Comments:

kjksf: I've built a web-based tool called Edna that allows storing knowledge in a single long plain text file. It supports dividing the file into virtual parts with different syntax highlighting. I've been using it as my primary note-taking and knowledge recording tool, and it works like a desktop app even though it's web-based.

FredPret: Edna's math block feature is very cool. It would be great if there was a way to reference variables or fetch data online to build a personal dashboard.

fellowniusmonk: I'm excited about tools like this that move towards structured collaborative text interfaces, as I think our current desktop UI paradigms will be consigned to history in favor of simpler, more integrated systems.

AtlasBarfed: This tool presents challenges around namespacing and indexing that are two hard problems in computer science.

m463: This approach of using a single big text file has a certain simplicity to it, similar to the "everything is a file" philosophy in Unix or Plan 9.

Dehyped title: Unable to generate summary

Summary:

We are sorry, we are not able to extract the source.

Comments:

This Hacker News thread discusses the ethical and legal issues with a service that hires big tech employees to appeal account suspensions for a fee. Many commenters agree this would likely result in the employees being fired for corruption and bribery, as it goes against company policies. Some see it as a necessary service due to the lack of proper customer support from these tech companies, while others view it as an unethical "black market" solution. Overall, the consensus is that this service poses significant risks for the employees involved.

Insightful contributor summaries:

constantcrying: This is essentially corruption and has serious legal implications that could get you fired.

MaxBarraclough: Paying employees to use company resources for personal gain is a form of bribery that companies would not tolerate.

JohnMakin: It's an open secret that having connections at companies like Meta/Facebook is the fastest way to get issues resolved, though the extent of this is concerning.

romanhn: While employees could help friends/family internally, assisting random strangers would be strictly against company policies.

8f2ab37a-ed6c: The lack of proper customer support from companies like Google has led to a "black market" for insider help to restore suspended accounts.

JohnMakin: Account suspensions by companies like Google are effectively theft, as there is no reasonable way for customers to appeal.

thih9: This service highlights the issues caused by both poor customer support and the gig economy.

blast: While unethical, I would be willing to pay for this service to fix an unjust account suspension issue.

aramova: The real solution would be to get a job on the team that oversees these appeals and then sell the ability to overlook cases.

bee_rider: Building a platform to bribe your way through opaque customer service systems is either low-level corruption or high-level performance art.

MrDarcy: The employees who participate in this service risk having their identities exposed through legal action against the service provider.

neilv: This could be used as a honeypot to weed out corrupt employees at these companies.

Dehyped title: Disable all AI-related features in iterm2 application

Summary:

This is a feature request on the GitLab page for the iTerm2 software. The user wants to know how to disable all AI-related features in iTerm2. The user expresses concerns about the potential privacy, performance, memory, complexity, and environmental impacts of AI features. The user suggests that this request covers a more general problem. Overall, the user is seeking a way to completely turn off any AI functionality within the iTerm2 application.

Comments:

mdeeks: I'm confused by the backlash against the new AI features in iTerm2, as they are opt-in and don't send anything to OpenAI unless you explicitly use the "Engage Artificial Intelligence" command. It's just a similar experience to using ChatGPT.

elicksaur: Some commenters are concerned that the mere existence of the AI features in iTerm2 means they can't use the app at work, which would be frustrating if it's their preferred terminal.

stalfosknight: I think people are just tired of the constant hype and push for AI to be included in everything, even when it doesn't really make sense or may not be safe. I'm wary of this new feature being added to iTerm2 as well.

tflol: You're missing that some people are worried this suggests all their terminal usage will now be used to train the AI, as the company has an incentive to leverage user data for that purpose.

Me1000: The AI features in iTerm2 are opt-in by default, and users can even supply their own OpenAI-compatible API URL to use the features without sending prompts to OpenAI, for privacy reasons.