HNHacker News
TopNewBestAskShowJobs

okennedy

393 karma · joined May 7, 2018

Hate notebooks... so do we. That's why we built one that doesn't suck. Check out Vizier, our spreadsheet-enabled, polyglot, notebook-flavored workflow system https://vizierdb.info
submissionscomments
okennedy··on Pangolin – Mobile AMD laptop with Ryzen CPU and Radeon graphics
+1 It might be a bit pricy, but the tech support is amazing.

I just spent a few weeks working with them to debug a transient kernel panic with my new Thelio. Throughout the entire process, I talked to support techs with actual agency, and not just automata reading a script. Serious kudos to System76.

(The problem ultimately turned out to be a clearly labeled experimental feature I'd turned on a few months earlier and forgotten about. Ooops...)

The Thelio is my daily driver, and PopOS just fades into the background.

okennedy··on Sixty Minutes Episode Is Pure Misleading Moral Panic About Section 230
Decouple discovery (e.g., suggesting posts from new users or paid ads) from the protections afforded to dumb carriers/hosting services (i.e., posts received from users you follow or are engaged in a discussion with).
okennedy··on Facebook: Free as in Bullshit
Worth noting: beer is encumbered by legal restrictions in many parts of the world. In these places, even free beer is still not legally available to anyone under 16, 18, 21, etc...
okennedy··on What If OpenDocument Used SQLite? (2014)
SQLite already has logic for this built in: transactions (https://www.sqlite.org/lang_transaction.html)

More broadly though, tech seems to be migrating away from explicit saves in general (e.g., when's the last time you needed to save something in a web-app). Maybe that's not such a bad thing? I learned to develop a reflexive cmd-/ctrl-s twitch as a youth, but I would have been just as happy to have not lost hours of work to power outages and/or cooperative multitasking.

okennedy··on Maybe we shouldn't want a fully decentralized web
I agree with your sentiment that this can be a slippery slope, but I also think that you (and TFA) are glossing over an important point: there's a huge difference between distribution infrastructure and discovery infrastructure.

The internet was full of conspiracy theories, crackpots, and silliness, well before the facebooks, youtubes, and so forth. The key difference is that these were hard to find. Early on, the most effective discovery infrastructure was based on hand-curated indexes (e.g., [1,2]). These moved slowly and were hard to game. Now, those same indexes are generated by algorithms, the gaming of which has spawned an entire cottage industry.

The parent's viewpoint and TFA are not incompatible, as long as we don't assume that content distribution and content discovery have to come from the same service. The open internet survived for over 2 decades without bringing down democracy. At the same time, it would be a good idea to see a bit more editorial responsibility from services driving discovery. Better still, if hosting and discovery are decoupled, there's no network effects or content archive creating a captive audience if editorial power is abused.

Tying things back to TFA, and perhaps the point that you're trying to make here, amadeuspagel: IPFS seems mostly aimed at providing hosting infrastructure rather than discovery. I'll emphasize the point again: censorship-free hosting has been largely unproblematic for the past two decades. What's been getting us into trouble now is discovery without editorial review (and forcibly tying discovery to hosting).

[1] https://en.wikipedia.org/wiki/Yahoo!#Founding [2] https://en.wikipedia.org/wiki/Whole_Internet_User%27s_Guide_...

okennedy··on Simple Password Management with GPG
There's one advantage to this approach that, as far as I've seen, no other password manager has been able to replicate. Pass is simple. It is literally a few hundred lines of clear, easy to read bash shell script (yeah, that's a phrase I never thought I'd hear myself saying, ever). It's simple because it's created by composing standard, simple off-the-shelf technologies (bash, gpg, git). Compared to more monolithic solutions, it's easy to compose those very same technologies in a different way. That, in turn, makes it easy to adapt to new settings, and super resilient to new use cases like different sync technologies.

(and yes, I do appreciate that this is not a feature that everyone wants/needs)

okennedy··on Maine Becomes First State to Use Ranked-Choice Voting in a Presidential Election
The key difference is that in (most) European electoral systems, voters vote for MPs, who then in turn vote among themselves to decide on a leader for the country. This creates a (limited) incentive for cooperation when no single party holds a majority, since a coalition government is required.

Simultaneously, voting for an MP has multiple impacts. Even if the MP isn't part of the coalition, they still have a say in how the government legislates, they can participate in debates, write bills, and take advantage of the resources of a government office to help their constituents.

Finally, voting for an MP is a local affair. The barrier to entry at this scale is high, but tractable. Alternative parties can offer credible competitions at the town or county level (e.g., the US congress has had at least some "independent" representation as far back as I can remember).

These three factors create a substantial difference in incentives in the US presidential election.

There are no shades of victory: Even if your elector votes for a 3rd party, there is no long-term benefit. After they cast their vote, they're done. This makes voting for a 3rd party extremely high risk: you're gambling that everyone else feels the same way you do. Even if you don't like either of the main two candidates, you're faced with what amounts to a prisoner's dilemma where the other 300 million players have repeatedly shown a policy of 'always betray'.

okennedy··on Facebook has sent a cease-and-desist letter to researchers
That makes quite a bit more sense. Thanks for clarifying.

To the grandparent: A researcher selling IRB-protected data would be effectively ending their academic career and opening themselves up to a mountain of legal trouble from the university and anyone who participated in the trial.

okennedy··on Facebook has sent a cease-and-desist letter to researchers
For a project like this to happen at a major US university (especially once outside funding is involved), it needs approval of the university's Institutional Review Board. Getting IRB approval entails researchers proposing a strict set of guidelines for how the data will be collected/used/stored, examining the potential for harm to participants, and convincing a room of very very risk averse individuals that the project is safe and bounded in scope.

This is in stark contrast to CA. "They're probably on the level" because they have entire systems in place to keep them there.

okennedy··on Moving your SSH port isn’t security by obscurity
I mentioned this in another response, but I'll refine my argument here.

Security by obscurity's negative connotation comes from the ratio of effort on the part of the system implementer to the effort on the part of the attacker. While it may take time to break into an "obscured" system, it also takes a lot of effort to build the system in the first place. Worse, if the security is broken, you have to re-implement from scratch.

Better approaches decouple the implementation/methods for enforcing security from the random components of the system that need to be kept secret. Whether or not this increases the effort for the attacker, it drastically reduces the effort for the defender. If an RSA key is cracked, a new one can be generated in less than a second, and once deployed the attacker is back to square one. Having SSH on a different port adds the cost of a full port-scan to any attempt to break in, vs the cost of changing one line in a config file.

That all being said, there is a major difference between changing your port number and changing your RSA key: The attacker/defender effort ratio scales linearly (vs exponentially). Every time you change your port number (a fixed cost), the attacker has to run another port scan (another fixed cost) to get past this layer. By contrast, the amount of effort needed to crack an RSA key is exponentially higher than the cost required to create a new one. However, adding a fixed port-scanning cost to each attack does actually correspond to a very high overhead for a non-targeted attack (their costs go up by ~2^16).

okennedy··on Moving your SSH port isn’t security by obscurity
This response is missing the author's point. There's a distinction between obscuring the methods and obscuring a parameter to those methods.

Security by obscurity, and in particular its negative connotation refers to obscuring methods. The reason it has a negative connotation is that the amount of work it takes to implement a secure system is high relative to the amount it takes to break it. By contrast, the amount of work needed to generate an RSA key, new port number, etc... is incredibly low relative to the amount of work it takes someone to break it.

Security by layers is important, but the article's point is that the ratio of effort to benefit for changing your port number is far far higher than the term "security by obscurity" generally implies.

okennedy··on Universal Paperclips
This is a fantastic game. If you haven't tried it already, don't be put-off by the somewhat slow start.
okennedy··on Pasco, Florida’s predictive policing system monitors and harasses residents
> The Sheriff’s Office said its program was designed to reduce bias in policing by using objective data.

Data is no more objective than the people collecting or analyzing it.

okennedy··on To head off regulators, Google makes certain words taboo
Google Reader?
okennedy··on Feds move to block California’s net neutrality law
I think that's the GP's point. Contradicting previously expressed principles shows how little the politician truly cares about those principles. Politicians lying or exaggerating is not surprising, but it's unfortunate that they don't get called on it by the media and/or public more often.
okennedy··on 27-inch iMac gets a major update
I agree that Apple has focused its attention exclusively on the consumer market... and that's a problem.

Developers write code for platforms that they're using. This is what drove adoption of Apple hardware in the early 2ks (a POSIX that runs MS Office!), which in turn set set the stage for the iPhone and iOS.

Unlike then, Apple now has a good grip on the consumer market: Not targetting iOS with a mobile/tablet release is a bad idea, regardless of whether or not devs are familiar with it. But OSX? There are a few areas where it's still has strong devotees: Color management in OSX is still fantastic, creating a lot of loyalty among artists, photographers, etc... It's still technically a POSIX, so it's still attractive to developers.

In short, OSX is not targetted at consumers. When someone needs a "computer" for their kid to do their homework, that's increasingly going to be an Android/iOS device.

For a platform to survive, it needs a healthy developer community. A developer community needs incentives (e.g., market-share or devs already using the platform). Right now, what are the incentives for OSX devs?

okennedy··on Google reportedly peeks into Android data to gain edge over third-party apps
The Librem 5's PureOS is very nearly stock Debian and the Purism devs have been very diligent about getting their stuff upstreamed into mainline Linux and Debian.

Pine doesn't employ software developers, but the Manjaro and Mobian communities are also doing their best to stay as close to stock desktop distributions as possible.

A key difference between the Pi and both the Librem 5 and Pinephone is that the latter two made an explicit design choice to use stock-standard (nearly) blob-free hardware. Unlike Raspbian, which relies on a custom kernel, PureOS, Mobian, Manjaro, etc... are very nearly standard desktop operating systems, with relatively minor tweaks to system defaults. Heck, PureOS and Manjaro are desktop operating systems just running on the phone with mobile-oriented shell .

There was a nice, related post on this from Purism a few days ago: https://puri.sm/posts/investing-in-real-convergence/

okennedy··on YouTube deleted an electronics repair channel [video]
This was exactly my intent. The problem is with YouTube trying to be a discovery service, while simultaneously requiring the use of its own infrastructure service to benefit from discovery. What is needed is discovery services that are not tied to any underlying infrastructure.
okennedy··on YouTube deleted an electronics repair channel [video]
This is the problem with combining services for infrastructure and services for discovery. Discovery, by definition, requires a carefully curated approach to filter based on interest, applicability. This forces discovery services to take for the material shown and promoted. Given the scale that services like YouTube operate at, this requires automation that is invariably going to make stupid mistakes.

Infrastructure services on the other hand are dumb pipes, giving them a lot more leeway. Historically (e.g., the safe harbor provisions of the DMCA, the recent debates around Cloudflare and the far-right), society has proven to be far more tolerant of services simply hosting a video or a stream, than services promoting it or deriving benefit from it (e.g., through ads).

Arguably, the latter is also the more critical. A lot of smaller-scale entities (like your church) can tolerate losing promotion for a day or a week far better than they can tolerate losing their infrastructure.

okennedy··on What Every Computer Scientist Should Know About Floating-Point Arithmetic (1991)
Even without conversions floats cause problems. Several years back I was working on a SQL query transpiler, and was getting driven up the wall trying to explain why compiling query 6 from the TPC-H benchmark on some of our backends would agree with Postgres, and disagree on others. Worse, it worked perfectly on the OCaml backend, we used for debugging, but broke on our performance-focused C++ backend.

I must have torn my hair out for nearly a week, but the culprit turned out to be TPC-H's defaults including a subtle and very hidden test for floating point numbers. Q6, with default values plugged in includes the bit:

  WHERE l_discount < 0.06+0.01  
In the C++ backend (with a standard `float` iirc), 0.06+0.01 != 0.07, and the query result ended up wrong.

Fun times. (I now occasionally toss Q6 in as a test query for my database class. Definitely teaches proper respect for floating point arithmetic :D)

okennedy··on The Atlantic lays off almost 20% of staff
It seems like this type of approach would incentivise clickbaity articles even more than the current financing structure does. Subscribing to a newspaper is a long-term arrangement, requiring the publisher to establish and maintain the reader's trust. Articles, on the other hand, are one-time interactions. Once you start seeing enough clickbait from a given source you can avoid the source in the future, but it would be significantly harder for publishers to leverage positive trust and reputation in such a setting.
okennedy··on SubEthaEdit 5 – native open-source text editor with live collaboration feature
I remember using it around the time they switched names. Although I barely ever used the collaborative features, SEE's multi-cursor editing capabilities were like magic. Yeah, other editors let you edit multiple lines, but SEE was the first editor I'd used that gave you multiple distinct cursors, rather than giving you a single vertically stacked cursor. Top that off with OSX's default emacs keybindings, and you had a super simple, but incredibly powerful editing experience.

I eventually needed something cross-platform and switched to Sublime, but I still have very fond memories of SEE.

okennedy··on Spore Prototypes (2008)
I've made several attempts at play-throughs over the years. Each and every time, the game crashed shortly after getting to the space stage. I'd curse myself for not remembering to save, forget about the game for a year or two, and the process would repeat.
okennedy··on Null Values in SQL Queries
The problem is a bit more subtle. NULL is often used with two distinct meanings:

- I know that Theodore Roosevelt has no middle name.

- I don't know what Theodore Roosevelt's middle name is

The semantics of each case are different in a subtle but important way. Consider the following two predicates

'Jim' = [no value] vs 'Jim' = [I don't know]

The former predicate is obviously false. 'Jim' is not equivalent to a non-existent value. The latter predicate however, can't be assigned a truth value. It's entirely possible that the unknown value could later turn out to be 'Jim'.

The key issue is that SQL treats NULL values according to the latter ('I don't know') rule and ternary (Kleene) logic, as TFA discusses. That, in turn means that using NULL to represent [no value] will lead to painful and hard to diagnose logic bugs (e.g., WHERE x = 1 OR x != 1 not returning all records).

[edit: clarifying why bugs will arise]

okennedy··on iOS 13's Music App Sucks
At the same time, the UX has gotten dramatically worse for people who do manage their own music libraries as well. The entire UI has been restructured around music discovery... which is appropriate for Apple Music, but really gets in the way when trying to find a specific album or track (e.g., I need to click through about three screens just to get to the album list, and there's no way to set the default). It's trying to do too many things at the same time.
okennedy··on What’s wrong with computational notebooks?
Check out our reproducibility focused notebook Vizier (https://vizierdb.info). In Vizier, inter-cell communication happens through spark dataframes (and we're working on other datatypes too. This makes it possible for Vizier to track inter-cell dependencies and automatically reschedules execution of dependent cells. (It also makes Vizier a polyglot notebook :) )
okennedy··on 0.30000000000000004
This specific issue nearly drove me insane trying to debug a SQL -> C++/Scala/OCaml transpiler years ago. We were using the TPC-H benchmark as part of our test suite, and (unbeknownst to me), the validation parameters for one of the queries (Q6) triggered this behavior (0.6+0.1 != 0.7), but only in the C/Scala targets. OCaml (around which we had built most of our debugging infrastructure) handled the math correctly...

Fun times.

okennedy··on Pocket Popcorn Computer: Linux in Your Pocket
> + settings for screen-lock-on-idle aren't saved - sometimes I want the screen to be always on, but it doesn't work

Depends on the window manager/desktop environment you’re using, but there’s dozens of Gnome3 extensions that do exactly this. “Caffeine”, for example.

okennedy··on Modern Data Practice and the SQL Tradition
The difficulty with the approach you propose, and by contrast one of the major strengths of SQL, is composition. Passing a list of column names and a list of conditions lets you express precisely two concepts: Filtering and Projection. You could also add more: A set of tables to join together, an "aggregation" version of the same operation, etc... Going down this path, however, leads to a monolithic function that becomes progressively harder to generalize.

What relational algebra (and by extension SQL) gets "right" is that each of these operations (Projection, Filtering, Join, Aggregation, Union) are composable: They take 1 or 2 collections as input and produce a collection as output. Moreover, each operation has simple and well-defined semantics from which you can build much more complex logic.

That's not to say that Relational Algebra can't be built in to an imperative language. Scala (and by extension Spark) collections are a great example of composable operators at work. Ruby's array methods, Python comprehension syntax, and Pandas/NumPy are similar examples of simple, composable primitives that combine to be much more powerful data transformations.

Apart from RA-based language primitives, there's also compiler support that allows you to use SQL directly, but avoid passing strings around at runtime. .NET's LINQ is a great example. I'll also pitch one of my own projects, DBToaster (https://dbtoaster.github.io/), which compiles view maintenance queries down to a C++ or Scala class.

In short, I agree that passing strings around leaves performance on the floor and leads to escaping and code injection nightmares. But SQL is the culmination of literally decades of use-based design, and any effort to replace it needs to take care to understand what it does well and why (like the efforts I reference above)

okennedy··on Ask HN: Feasible Alternative to the MacBook Pro?
I forget the full list, but these had been standard since I started with OS7.

cmd-s Save

cmd-n New folder

cmd-shift-s Save As

It's possible some have been added back.

← PreviousPage 3 of 4Next →