HNHacker News
TopNewBestAskShowJobs

drewcrawford

4,834 karma · joined June 13, 2008

AI tooling and systems/SDK engineer · Rust, WebAssembly, Swift/Apple platforms · wasm-bindgen contributor · building AI Bot Mafia (botmafia.games) · sealedabstract.com · github.com/drewcrawford
submissionscomments
drewcrawford··on DivX Now Installs Malware
Vuze for Mac does.
drewcrawford··on Dissent: accountable anonymous group communication
If you're correct, I don't really see what is novel here. If I already had a group of 50 people I could trust, what would I need this software for? How is this better than just handing out credentials to my IRC Tor hidden service?
drewcrawford··on Dissent: accountable anonymous group communication
So the problem I see with that solution is twofold:

* If there is a central authority collecting the information, then as far as Jeff the User knows, his information could be stored in plain text on a subpoenable hard disk by the central authority.

* If the information is verified by a set of moderators, then the moderators must be numerous enough that no single moderator could capture a substantial portion of the identity data. But to achieve that, the mod requirements would be low enough that an attacker could infiltrate their ranks and uncover the identities of users.

drewcrawford··on Dissent: accountable anonymous group communication
There are something like 20-ish papers here. I have skimmed several of them without any clear answer to this question:

How exactly do you stop sockpuppetry while maintaining anonymity?

In order to ensure that a person cannot create more than 1 account, it would be necessary to observe some property of the person that is difficult to alter (such as their appearance, or their social security number) for uniqueness on the network.

Now if this network is truly anonymous, then it must be disjoint from every other network. And there is at least one other network ("real life"), so whatever this property is, must be something that an adversary in the real-life network would not be able to compare to the anonymous network.

And so certainly, it must not be some property that a person in the real-life network could observe by following you around or ruffling through your drawers and then compare to the answer you gave to sign up for the anonymous network. So it must be something in your head (like a password).

But if it is in your head, isn't it easily changed, and this property would be exploitable to create sockpuppets?

I must have missed something fundamental here, possibly in terminology. Can somebody who is current on the research here enlighten me?

drewcrawford··on CTL-C: Re-imagined IaaS with Docker Containers
Hello,

I am probably as close as you will get to an early adopter for this product. Make absolutely certain that I show up on your "customer interview" list. E-mail in profile.

drewcrawford··on This site is down because the owner stiffed the web designer
I'm continually fascinated by these comments (there's one on every story like this) that try to second-guess the decision based on no actual information. Doesn't it seem likely that the person who did this is in a better position to judge the potential consequences than you?

Unless OP is right on the threshold of being hard up for work, it makes no real difference to his dealflow, and it has a nonzero probability that it will shame somebody into paying.

drewcrawford··on Bitcoin is a Startup
> The only problem here is that coffee is a commodity, not a startup. Anyone can grow it and sell it.

The better comparison is to oil. Oil is a commodity, in that anyone can buy it and sell it: but there is a fixed amount of it, and one can only acquire oil by taking oil from the fixed supply. Further, the earlier you got into oil, the more money you made, and the same is true of Bitcoin: the authors have made out to the tune of [100M or more](http://bitslog.wordpress.com/2013/04/17/the-well-deserved-fo...) simply by being the first miners.

So I don't know exactly what kind of definition you want to use for "startup", but bitcoin is certainly software, that was written by a small team of people, that solves a problem, that has made the authors wealthy in a very short amount of time. If someone like Rockefeller started out today, I think we would probably characterize his operation as a "startup", and the founders of Bitcoin are doing much the same thing that he did.

drewcrawford··on Facebook vulnerability 2013
Well according to Facebook, "this is not a bug". Which means the feature works as intended. If he is using Facebook as it is intended, then how can he be breaking the TOS?

When an employee whose job it is to evaluate security issues says "this is not a bug", that determination carries the force of law the same way as if it appeared in the TOS. You cannot rely on people to follow some nebulous "spirit of the TOS" when meanwhile your employees have already made a contrary specific determination for how it applies to this particular bug.

drewcrawford··on Greenwald: Low-Level NSA Analysts Have Powerful and Invasive Search Tool [video]
> How _dumb_ must one be to continue to trust these agencies

Taking a deep breath for a moment, if you look at TFA it contains two contradicting allegations, from sources that appear to me to be roughly similarly credible.

On what evidentiary basis have you decided one source is more reliable than the other? And why are the people who believe the other source "lobotomized"?

I am not really sure what is going on here, but I do not feel that this article alone presents a sufficient basis for a reasonable person to form an opinion.

drewcrawford··on Licensing in a Post Copyright World
This is a great article.

One thing I would add to this is that nobody really knows what licenses like the GPL actually do. Jacob Kaplan-Moss published a fantastic piece called "20 questions about the GPL"[0] , and my attempts to clarify a few of them with the FSF and SFLC have so far revealed that not even two developer advocate organizations can agree on even fundamentally basic concepts about the most popular license, so isn't that a problem?

The reason that I contribute to the "projects on GitHub with no license" problem is that I cannot find a license that unambiguously does what I want, after spending many, many weekends on the problem. I could write one, or get my lawyer to, but license proliferation is grounds for expulsion from the cool kids club. Since I have no good options, I take no action.

Fundamentally, I believe we need a new license steward that isn't afraid to update regularly to address 1990s questions like dynamic linking and optional dependencies, and 2010s questions like app stores. Something like Creative Commons for code, where you have some options to choose from.

[0] http://jacobian.org/writing/gpl-questions/

drewcrawford··on Show HN: I'm a single founder. Please review my bootstrapped startup
Source: I run a few production backend services in Python. Mostly to Debian boxes.

No matter what your product anticipates, it doesn't anticipate all the things I do, in fairly boring applications. I compile my own Python. I install postfix for some applications. I install a kernel module that increases the Linux entropy pool. You have not anticipated my set of problems.

What this means is that no matter how much great stuff your product brings to the table, I bring just as much, if not more, configuring your product and designing a deployment solution around it for my application.

If your product isn't open-source, that's a non-starter for me. Because now I'm worried that I could spend a lot of time configuring the product, and not being able to add that one feature that I need. I'm not sure if you can run a business with an open-source deployment product, but I know that I won't ever use a closed-source one.

Do a better job of explaining how these apps run. I took a look at your sample manifest files, and it wasn't clear to me if you are installing nginx or something? Or if you intend the apps to be daemon-like? So create a page that explains, in words, how to deploy a daemon-like app and how to deploy a webapp, and who is responsible for installing the web server and editing its configuration file(s) (do you clobber them?) and what web server it is supposed to be.

Finally, add a docker.io example. It's not production ready, but it looks like it's designed for people like me who have deployments involving heterogeneous payloads. I'm going to move to it eventually, so if I was to switch deployment stacks I would want good support for it.

drewcrawford··on Why mobile web apps are slow
No. But Apple has a shockingly small number of employees relative to other big software names, and they have had historically had much fewer third-party developers than other platforms.

As a result, they have chosen to consolidate the mindshare onto one memory management technology rather than dilute the talent pool with two competing proposals. Given that iOS is much, much more popular than OS X, they have chosen to standardize on the technology that works best for the popular platform.

drewcrawford··on How Driverless Cars Could Reshape Cities
> Are bikes somehow magically different than cars?

It's not so much that bikes are magically different than cars, but that NYC has many more viable transportation alternatives than ordinary American cities. If you find yourself missing a bike for the return trip, public transport is a reliable backup in NYC. This is not the case in most other cities in the US.

drewcrawford··on Pirated Apps in the App Store
I have actually sent a DMCA notice to Apple.

While I would also suggest the DMCA route, I can tell you that you will experience the exact same effect (no obvious response) observed by the OP.

drewcrawford··on Stop selling your Software for Peanuts
> I don't see any reason that the marginal value to the consumer of software would be near-zero.

If you look at the top 10 paid apps in the iOS app store, 9/10 are games. In a sufficiently large market, games are interchangeable; I might equally prefer hundreds of games. As long as at least one of those games is free, the marginal value of some particular game to me is zero.

Now this may not be true for example when considering applications used for business, that provide perhaps a lot of value. But these represent just a tiny fraction of the overall market, which is mostly games and other mass-market, interchangeable software. It is large in absolute terms of course; that is how a great many developers make their living. But it is very small relative to the overall ecosystem, so it is fair to characterize most purchases as having essentially no marginal value.

drewcrawford··on OS X is holding back the 2013 MacBook Air’s 802.11ac Wi-Fi speeds
Probably because afp or cifs are going to be the situation that ordinary people experience a WLAN bottleneck?

The speed of RTSP over 802.11ac would be relevant if you are giving a keynote address in your living room that needs to be streamed to the hundreds of interested participants in the kitchen.

drewcrawford··on Packagers don't know best
> If the Riak guys want to use leveldb in a way Google won't support, they should rally with the package managers and get Google to stop being "pretend open source."

They don't have any control over Google, by "rallying" or otherwise.

Consider the recent case of the WebKit/blink split. Here you have two sets of some of the world's smartest engineers, who cannot agree about how to render a webpage! And this is a well-defined problem. There are actual standards about how to render webpages! In theory, everybody agrees about what's going on here, and yet it's fork time.

As for what Library X does, there are no standards, and not necessarily does anybody agree about what they are building. And let's be honest here, you probably do not have WebKit-caliber developers hacking on Library X. So the chance that you can arrive at consensus for Library X is much lower than for WebKit/blink.

Meanwhile, instead of dicking around with the will-they-wont-they-merge-upstream committee, you can just ship software that works in practice to people that want to use it. If you are a software developer, and you have the choice between writing software and arguing about it, it is usually a good bet on average to write software.

drewcrawford··on Ask HN: Passive income ideas?
> The thing is maybe those groups aren't willing to pay for such a solution?

Actually I think they want to pay more. At least I do. HipChat charges $2/user; I would pay $5 or $10.

The real problem is that tools designed for companies consolidate the billing. If I start a chat for my local user group then I am on the hook socially to upgrade my plan to let new people join. If I am successful in cultivating a useful chat then essentially I am punished for it, and the punishment grows without bound. One of the groups I participate in has over a thousand members.

Meanwhile I would be happy to pay for a friend or three, with the understanding that the other 997 people who may want to join need to pay their own way. And with mostly individual accounts you can probably ask for significantly more per user than HipChat can.

And the effort to get people to sign up might be effectively free if you funnel part of the license fee back to the organization who no longer needs to pass the hat around for pizza money or whatever. This may create regulatory problems, but if we're just talking about the customer acquisition equation, there are a lot of awkward situations where somebody forgot to bring cash that I would very happily roll into a monthly debit to forget about.

drewcrawford··on Ask HN: Passive income ideas?
> The only problem with this is that outside clients wont have access to you exchange server.

My company consists of: me and the guy sitting a few feet from me. If I wanted to schedule a meeting with that person, Exchange would be a poor improvement on the existing solution of turning my head.

drewcrawford··on Ask HN: Passive income ideas?
Two ideas.

1. I desperately want to be able to send a link to somebody and have them be able to book a N-minute meeting on my calendar (CalDav) without any further interaction. I have spent probably several hours trying out products, all of which suck. I currently spend $30/month on a tool that really doesn't even work.

2. The enterprise IM (CampFire, HipChat, etc.) is pretty well done at this point, but the "group chat for the somewhat public" space isn't. If I am supporting an open-source project the only game in town is IRC, which is hard to use for people who need support. If I want to have a permanent chatroom with 15 software developers that I know socially and allow people to drop in, there's no system for that.

drewcrawford··on Shakespeare.txt.jpg: A JPEG compression experiment
it's comomn kdownlege in inoitrmfaon thoery ciecrls taht you can reoerdr lttres in wrods and the relsut is pecfretly undeastnrdlabe as lnog as you keep the fsirt and lsat letrtes the smae.

http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=4126...

drewcrawford··on Usability vs. Security in the Context of Apple iOS Mobile Hotspots
Per [1], it seems that the method in question is some sort of Markov generator that Apple uses to generate the passwords in the first place.

[1] https://www1.cs.fau.de/filepool/projects/hotspot/hotspot.pdf

drewcrawford··on Apple’s Commitment to Customer Privacy
I think you may be misreading Apple's statement. The security of iMessage does not derive (solely) from PKI. Instead, the spokesman claims:

> There are certain categories of information which we do not provide to law enforcement or any other group because we choose not to retain it

Now it is up to you whether or not you take this statement at face value. But it's consistent with the other things I know about iMessage, as a person who has studied it in far more detail than your average HN reader, and so I believe it to be true.

If it is true, then Apple would be limited by technology to only provide iMessage dumps that started the date of the tap. If they did at all.

And I would be very surprised if they do at all. You obviously have to comply with a warrant to provide data if you have that data, but it becomes much more difficult, legally speaking, to compel somebody to collect information that they do not have to begin with. The CALEA was conceived to "correct" this "problem" but it only applies to "telecommunications infrastructure" organizations like ISPs, and courts have been reluctant to expand its application to actual internet services. It's something of an open secret that it was easier to simply acquire Skype for $8.5 bn and modify their technology than it was to obtain the necessary ruling under the CALEA that would compel them to modify it themselves.

Now if you are plotting revolutions, by all means, don't use iMessage. But my assessment is that the NSA threat level is lower than any other type of communication an ordinary person would be able to use, such as phone, e-mail, or even SSL.

Just as a point of comparison, DJB et al recently (2013) published a repeated plaintext attack on SSL/TLS that is viable at 2^24 TLS sessions. That sounds high enough to be safe, but if you consider the server case or the cron job case of rapidly opening SSL sessions, not so much. And attacks only get better. History suggests that the NSA is often a good ten or fifteen years ahead of the independent cryptographers, and meanwhile we are at plausible SSL attacks today.

drewcrawford··on The lack of proper “alter table” support in SQLite
Let me jump in as somebody very familiar with two of the tools mentioned in the OP--sqlachemy and alembic.

> posts about why they are evil

There are legitimate flame wars to be had about ORMs being "evil", but they all have to be had in the context of using them to solve a particular thing. It is true that ORMs do not solve all problems for all people, but it is also true that they solve some problems for some people. There is no point in hashing it out without knowing what OP or me or you or some particular person is doing.

> being able to do arbitrary schema migrations is considerably harder

Being able to do arbitrary schema migrations is not a feature anybody is seriously suggesting, either for SQLite, sqlalchemy, alembic, or any other tool. What is being suggested is merely that SQLite support ALTER TABLE, a very particular, well-defined feature that has been standardized since at least 1992. In the intervening two decades I am not aware of any serious proposals to the effect that the standardized behavior is flawed, although there are various vendor-specific extensions. If you would like to raise an objection to the standard, I would be greatly intrigued.

> I couldn't find an explanation as to why they think the copy and adjust approach would be so hard

All snark aside, four out of five SQL vendors support alter table. They would be doing this logic solely for SQLite. Why would they do that? The path of least resistance is just to tell people to migrate to MySQL. That's what people who need migration support end up doing.

> alter table is very infrequently mentioned

You seem to anticipate that users will discover that SQLite does not support alter table correctly, and that following this discovery they will continue to use SQLite. My experience is at variance with your anticipation.

Specifically, I don't post to SQLite-users, first and foremost, because I am not a SQLite user, because it does not support alter table. The sampling bias would actually preclude you from hearing about any really critical feature requests.

> you would have to wait years before it is present on SQLite versions used by stable/enterprise distributions

Literally an argument against doing anything to SQLite ever. Is this a common objection to feature requests on the mailing list?

> You'll also need to establish what you want to happen

OP needs to establish nothing, this has literally been decided a full eight years before the first line of SQLite was ever written.

I know you're thinking "SQLite is fine, go use MySQL if you need this alter table nonsense". And that's precisely what people who do a lot of migrations end up doing. I've migrated every SQLite codebase I've committed to this year to either MySQL or Postgres or NoSQL. Every last one.

drewcrawford··on NSA leaker is a patriot, not a traitor
This is surprisingly easy to fix.

What you have to do is

1. Any agency can designate a document as classified, a designation which remains in play for four years. This dovetails nicely with the election cycle. Also, the majority of police, etc. undercover operations are shorter than this period. Above this you are in "major case" / RICO / terrorism / etc. territory.

2. Extending the classification beyond four years requires a motion to show cause before a bona fide judiciary, where the agency must show by clear and convincing evidence that the classification falls into an enumerated exemption criteria. The classification can be extended indefinitely, but only in increments of 12 months and only upon the approval of the judicial branch.

A proposal like this addresses most people's objections. The real problem with the current classification scheme is A) like Mickey Mouse's copyright, it expires on paper but not in real life, and B) the agencies are all essentially self-policing and nobody really checks in on them to see if the classification decisions are correct.

The thing about this proposal is that it requires all three branches to sign off and say "yes, this should stay classified" and it also limits the number of secrets you can have, because there are only so many lawyers and so many judges, and now you have to decide if you want to allocate time in the budget to make motions for the next 50 years. There are legitimate secrets that are worth that kind of hassle, and then there are the other 90% of stuff that the alphabet soup of agencies does.

drewcrawford··on Pearl Boy WebGL Demo
I see a lot of jitter on a very fast Mac. (Safari 6.0.4).

Somebody who understands Web GL tell me -- developer's fault or browser's fault? It looks a little like numerical instability to me rather than a framerate issue per se, but I'm no 3D expert.

http://jsbench.s3.amazonaws.com/jitter.m4v

drewcrawford··on Microsoft responds to YouTube demands, 'more than happy' to include ads
> Once upon a time people screamed murder when a platform company started weaving its own interests directly into its platform. For example, Microsoft

When you say "screamed murder", do you mean that people got huffy about it on the Internet? That's still the case. Exhibit A: your comment.

Did you mean "antitrust lawsuits used to be filed about it?" A random sampling of any paragraph in the actual court decision should convince you that the situation is not comparable to today:

> Microsoft possesses a dominant, persistent, and increasing share of the world- wide market for Intel-compatible PC operating systems. Every year for the last decade, Microsoft's share of the market for Intel-compatible PC operating systems has stood above ninety percent. For the last couple of years the figure has been at least ninety-five percent, and analysts project that the share will climb even higher over the next few years. - U.S. v. Microsoft, Findings of Fact

If at some point Apple's share of the phone market, or even of the ARM-based phone market, exceeds 90 percent, over a period of a decade--by all means, let's file a lawsuit about it.

drewcrawford··on Our Failure In the App Store May Help You
> Ironically, App Store distribution is becoming nearly as difficult and cost prohibitive as pre-internet game retailing.

The only reason that this is true is that game development is essentially a game of madlibs. As soon as I say "rail shooter" or "turn-based strategy" you instantly know how to play, because it is exactly the same as the last four games of that type you played with different artwork.

I don't play a lot of games these days, but the last game I played that was mildly innovative was Portal. Big-box success, right? No. It was a student design project that had actual dead-tree mainstream press coverage[1] before it got bought by Valve.

Imagine for a moment if all the spreadsheet software that was written were Excel clones with slightly different menus. Success would all come down to the marketing lottery, no? But take a look at the world that we actually live in [2] [3] [4] [5]. It is easy to stand out if you are doing something legitimately different than the other entrants in the space.

And there are plenty of people making innovative games. I've compiled this list in five minutes: LiquidSketch, Fez, Amnesia, Monaco, Gish, Darwinia... I don't see these guys worried about standing out in the App Store.

I played the game in TFA and it is literally 1 drag and tap per game. During the two games I played, I got four different upsells and confused the ads with the user interface at least twice. If there is any innovation in this title, it revolves around delivering the least amount of game per ad/upsell than anything I have ever played before.

[1] http://www.jeepbarnett.com/gallery2/v/scrapbook/narbaculardr...

[2] http://getpermanent.com

[3] http://grid.binarythumb.com

[4] http://www.lumina.com/uploads/technology/Whats%20wrong%20wit...

[5] https://itunes.apple.com/us/app/glimpse/id423562777?mt=8

drewcrawford··on Why I Develop For The Mac
> What's wrong in this case with the very good performance across many platforms (even tablets) of the SVG-backed RaphaelJS?

So I attempted to determine the accuracy of this claim. I ran a benchmark from Kevin Roast [1] who seems to author a lot of Canvas demos. For each of the 8 tests in his benchmark, I recorded the FPS reading that I saw that was the lowest (e.g., framerate dropped to X at some point during the 5-second test).

Mac iPad

20fps 16fps

29fps 18fps

30fps 21fps

30fps 30fps

28fps 6fps

29fps 18fps

29fps 7fps

12fps 30fps

mean_mac = 25.875, median_mac = 29 mean_ipad = 18.25, median_ipad = 18

On the Mac side, I can see both sides of the issue. Arguably 26-29fps in a wide variety of situations is good enough for a wide variety of applications. At the same time, I can understand the author really wanting to blow past 29fps.

On the iPad side, the issue is more clear. I think most people would say that 18fps is unacceptable for a drawing application.

(If these numbers are contributing to the "pseudo-myth" of slow Canvas performance, please point me to a reasonable benchmark. This one is just the most comprehensive one that I found.)

> At the risk of being a bit rude, it's worth noting that he doesn't bother with all Desktops, just Mac, which causes further suspicion that this is really just a rationalization piece and not about performance.

He addresses the choice of the Mac platform in some detail in the "Drawbacks" and "Conclusions" section of the post. At the risk of being a bit rude, it's poor form to dismiss someone's reasoning as a "rationalization" without addressing the reasoning on the merits. To the extent that his choice of Mac over Windows et al is specious, it is not a claim that is supported by your comment.

[1] http://www.kevs3d.co.uk/dev/canvasmark/

drewcrawford··on Stay away from rebase
Suppose you want to know, not just the fact that dev is failing, but why it's failing. The particular commit that caused it to fail.
← PreviousPage 11 of 29Next →