HNHacker News
TopNewBestAskShowJobs

alialkhatib

175 karma · joined March 14, 2012

computer scientist; previously anthropologist

https://ali-alkhatib.com

submissionscomments
alialkhatib··on Fry's Electronics owns and operates a rare 747SP
(Evidently) the San Jose SaberCats (football).
alialkhatib··on Starting Now at Netflix: Unlimited Maternity and Paternity Leave
Previous substantial discussion here: https://news.ycombinator.com/item?id=10006542
alialkhatib··on Netflix Announces Its New ‘Unlimited’ Maternity and Paternity Leave Program
A friend has 6 kids and I'm beginning to wonder what his company's pat/mat policy is...
alialkhatib··on No, I Am Not Lost – A Black Woman's Experience in the Stanford CS Major
Slightly tangential, but sometimes people I can only describe as "tourists" walk around in the publicly accessible areas, so it's not totally unprecedented.

For what it's worth, I tend to tell these people apart from the people with business in the building by noting where they're looking (tourists seem to be looking all over the place like they're taking it all in; people with business at Gates seem to be looking for numbers and other navigational indicators in line with tzs's comment).

alialkhatib··on Is Uber Making Taxi Service Better?
That's true. I think I just got annoyed at the emotional appeal of the Nash death and overreached. Sorry.
alialkhatib··on Is Uber Making Taxi Service Better?
That's true, but it's dwarfed by the amount of data representing traditional cabs. If we're not just selecting for cases involving deaths, a string of cases implicating Uber drivers in violence against passengers[0,1,2] suggests - if nothing else - that the notion that Uber/Lyft drivers are qualitatively safer than taxis is false. And in the last case, it's clear that shortcomings in a system (namely, the thoroughness of background checks Uber runs) might have prevented such a turn of events.

0: http://www.nbcbayarea.com/news/local/UberX-Driver-Accused-of...

1: http://laist.com/2015/04/10/uber_driver_attack_van_nuys.php

2: http://www.forbes.com/sites/ellenhuet/2014/06/03/uber-driver...

alialkhatib··on Is Uber Making Taxi Service Better?
I don't think this remark is appropriate, especially given that an UberX driver killed a 6-year-old about 2 and a half years ago, but mostly because Uber and Lyft have been around for only around 5 years at this point.
alialkhatib··on Show HN: The Most Influential Works on TvTropes According to PageRank
This is pretty cool. I suppose I'd like to see a breakdown of the influences some of the more popular sources had on other things, or maybe to know who are the most prolific borrowers of references and themes (influencees? Chrome tells me that's not a word). Nevertheless, it's interesting to think about cultural references within a very closed environment like TV shows and movies as a network, with all the reasoning affordances as such.

As an aside, TV Tropes doesn't use HTTPS as far as I know (and according to this too [0]), so none of the show/movie links resolve unless you go in and remove the prefix and let the browser figure it out. Looking at your source, it seems you could just change the "https://" to "http://" or even "//" (although this would break when(/if) you start using HTTPS down the road).

[0]: http://tvtropes.org/pmwiki/posts.php?discussion=13461707110A...

alialkhatib··on Numi: Beautiful calculator app for Mac you will love to use every day
I also use Alfred for a quick definition if I'm pretty sure the system dictionary can handle the word (otherwise I ask Google), but I agree that a vanishingly small number of additions to Spotlight would render Alfred defunct for me. For 90+% of the things Alfred does, I find myself thinking "oh that's pretty neat", but I simply can't make it go into my workflow.
alialkhatib··on Numi: Beautiful calculator app for Mac you will love to use every day
Agreed. I barely use Alfred that much and that's mapped to the similar-but-different option+space (Spotlight is mapped to command+space by default). If this kind of functionality could extend Alfred (or Quicksilver?), there would be no question that this would see relatively regular use from me.

If I need to open something, I'm much more likely to open another tab of my browser and dump queries (especially things like exchange/conversion rates) into Google. It's a shame, too, because the demo here is really persuasively showing that this can automatically do stuff I previously assumed necessarily takes an extra brain cycle or two.

alialkhatib··on Paul Graham's Amazon reviews
"Paul Graham mainly wants to know how he can opt out of this 'feature'."
alialkhatib··on Yet Another Language Speed Test – Counting Primes
I'm glad you brought this up, because I was thinking about it but felt like my relative unfamiliarity with C(++) and a handful of other languages in this "shoot-out" made my understanding here flawed (or at least partially naive). I'm also going to piggyback a bit because another thought I had seems related.

There's also the issue that among languages like Python, standard libraries (like multiprocessing[0]) make parallelizing code trivial (even with/despite the dreaded global interpreter lock). I rewrote some of the code the author uses and on a smaller number (6710886 instead of 67108864) I'm seeing ~1/3 the time to run compared to his baseline with 4 worker processes (27.469517s vs 1:25.373176). I haven't looked into what might be keeping it from being a full 4x faster, but my hunch is that it's a combination of other things going on in my laptop and the undeniable difference in execution speed in cPython (specifically, cPython3.4).

One might make the argument that this is a wildly unfair comparison, but my bigger-picture argument is that a language that abstracts things like parallelization to the point that one can invoke such features in a line or two of added & changed code is probably worth factoring into an analysis, if we're even interested in analyzing languages in this way.

If we're going to write C-like code and run it as Python, should we be surprised when it doesn't live up to those expectations?

[0]: https://docs.python.org/3.4/library/multiprocessing.html

alialkhatib··on I broke Uber
previous discussion with author's responses here: https://news.ycombinator.com/item?id=9715730
alialkhatib··on Google Calendar to end SMS notification on June 27th
Ah, maybe I should have clarified. While I've taken some care to curate the notifications I get on my phone, there are nevertheless times that Twitter, Facebook, Gmail, etc... would just be a nuisance. When I'm meeting with someone, the buzz of my phone is distracting to me and strikes me as a little rude to the other person (we both know I'm thinking about that buzz, and I want to be completely present in that conversation), which is why I tend to enable Do Not Disturb mode when I'm supposed to be engaged. The problem is that some notifications are valid to let through - in particular, a reminder that I have an event in 15 minutes (particularly a one-on-one meeting) is absolutely worth letting through.

My question was whether there's a policy suitably fine-grained to let me whitelist an app (or a certain type of notification) so that it can get around Do Not Disturb mode and reach me without having to completely shut down notifications from other apps in the settings.

alialkhatib··on Google Calendar to end SMS notification on June 27th
This is disheartening. I had Google Calendar's number whitelisted so I'd get notifications (ie event reminders) from Google Calendar (and important people) when I was in meetings and whatnot. Now that I can't simplistically use iOS's contact whitelist to trump Do Not Disturb, do I have to get more involved about which apps have permission to send me notifications? Do I even have that fine-grained control in iOS?
alialkhatib··on Show HN: Filld – Gas delivered to your car, on demand
This seems like it would make more sense for Filld and consumers on a recurring delivery model.
alialkhatib··on Why I Don’t Date Stamp My Blog Articles
See, I take the opposite view for the same reason. I see the contextual information like date as informative in my own posts and in others. An outrageous post looks less outrageous posted in a time when our knowledge of the world was different. Insight that's on the mark is that much more prescient if someone made it well before anyone else saw that same insight.

I think there are probably good reasons not to provide context for your posts, just like there are times that it makes more sense to post anonymously (or pseudonymously), but I don't think this post makes a good case for removing context like post date by default.

alialkhatib··on Keep your application settings in sync (OS X/Linux)
This is pretty cool. For a while I've used a dotfiles repo with various settings and config files in there, but I never got around to writing a script to automate symlinking things on new machines (and which things not to bother with - eg Sublime Text preferences on some remote machines).

Can you explain how sync differs if you use Git as your sync arrangement vs Dropbox? I know that Dropbox will propagate changes automatically, but git shouldn't, so would I have to go in and commit/push/pull changes if I opted for git, or does this make a decision about when to make these changes?

alialkhatib··on CrowdFlower Launches Open Data Project
This is interesting, but a few concerns are nagging me:

1) There's been some research about the demographics of crowdworkers [1] that questions whether it's a representative sample of any particular population, and I wouldn't immediately concede that conclusions drawn from these survey results are externally valid.

2) There's at least a little controversy around how well (or rather, poorly) crowdworkers are paid and treated, with workers on Amazon's Mechanical Turk going as far as crowdsourcing a set of guidelines for academic requesters to right some pretty egregious wrongs that have happened in the past few years (disclosure: I was involved in the research [2] that facilitated those guidelines).

I wouldn't necessarily forsake this data or anything learned from it, but these concerns make me wary to use it myself very readily.

Does anyone else have concerns? Am I just being hyper-vigilant because this is a personal research interest? I'm open to accusations that I'm biased, at least in this case :P

[1]: Ross, Joel, et al. "Who are the crowdworkers?: shifting demographics in mechanical turk." CHI'10 extended abstracts on Human factors in computing systems. ACM, 2010.

[2]: publication/citation forthcoming; for now, a link is easier -> https://ali-alkhatib.com/media/papers/pn2032-paper.pdf

alialkhatib··on A female computer science major at Stanford: “Floored” by the sexism
As a PhD student here at Stanford, I find this really distressing.

For what little it may be worth to anyone reading this who is(/was) thinking about coming to Stanford, I have heard professors and grad students talk about ways to be more inclusive in their actions and language. I've heard my advisor stumble over "you guys" only to stop and correct himself with "you all". I can't speak for every lab (there are numerous, each with different cultures and norms), but hopefully I speak for everyone when I say that we grad students and the professors are at least open to learning about how we can make the CS program (and the field as a whole) a safer space for everyone, if not actively working toward that goal. If nothing else, we can promote an inclusive culture in how we speak, behave, and in what behavior we (don't) tolerate.

For Lea (and for countless others), I worry her perception of CS has been irreparably damaged. Hopefully that won't be the case going forward.

alialkhatib··on Jawbone: The trials of a 16-year-old can’t-miss startup
Uncanny that I had the exact same timeframe for getting a replacement; in my undergrad I was doing an ethnography on QS culture and spent about 2 weeks waiting for my UP band to get replaced when it broke, only to have the replacement break a few weeks later. I ended up buying a second band and at one point used each in rotations while the other was in the process of getting replaced. After my fieldwork I stopped bothering to replace them despite them still having some time left on their warranties.
alialkhatib··on Place-hacking nocturnal London
Interesting that it's called "place-hacking" here. For a long time (or just in the US, or both) this is known as "urban exploration". I wonder if this culture differs from UrbEx in ways other than name.
alialkhatib··on You cannot have more than 10 years of experience in Rails
It's possible that the email meant 10+ years of collective experience between those 2 developers, but then it would be a bizarre detail to include (especially given the brevity of the email). Still, it sounds based on the interview details like they didn't bother to put someone technical on the phone interview, which makes me reluctant to give them the benefit of the doubt.
alialkhatib··on Show HN Browser extension Snooze lets you put off tabs for later
I find myself comparing this to Pocket, which I tend to use to queue content for later in much the same way that it seems Snooze is made to support. What Pocket offers over this is a cleaned up page of text to read, but what Snooze seems to offer is bringing things back to the forefront automatically.

I've debated where to go from here in my thoughts, and I think the issue I'm struggling with is that the use cases seem different. When I think about "snoozing" something (e.g. Google Inbox, a conventional alarm clock, etc.) I find that the common denominator is that someone or something else is on the other side waiting on me. If nobody was expecting me in lab, I'd come in whenever I woke up naturally (and I do). Similarly, I snooze an email in Inbox (when I do) because someone's on the other end waiting (with some variable amount of urgency) for my reply. A silent "deal with these later" folder would become a black hole, or at least would have a molasses-slow turnaround time.

Given this preconception, I'm not entirely convinced that "snoozing" is the right solution to the problem I'm assuming you're trying to solve (ie trying to help people keep track of and get through all the cool stuff they see on the Internet, right?). Instead, you kind of want to find the gaps in people's time and fill it with the right article for that stretch of time (which is difficult to do for myriad reasons).

I think something along these lines that would be interesting would be an extension that prompted you to catch up on your queue after a few minutes of browsing unproductive sites (e.g. Facebook, HN, Youtube). Even better, if the user could give it a list of sites and say "if I'm browsing any of these sites for more than 5 minutes, please remind me that I have a queue of webpages I wanted to read."

These are just initial thoughts, and I apologize if it's overly critical. I like the idea, and there's a lot of potential here. It's also entirely possible I'm just flat out wrong (and if someone has different experiences or insights that conflict with mine, I'd love to hear it).

alialkhatib··on ChatGrape's Google Drive Integration Presented in Under 2 Seconds
I'm assuming the intended link is http://blog.chatgrape.com/2014/11/12/the-incredible-google-d...

The original link was to the animated gif within that post

alialkhatib··on Badassify your terminal and shell
One thing I like about zsh that doesn't seem to get that much attention (and seems to be overlooked here) is path expansion.

If you're typing a path (let's say to ~/Dropbox/Public/...) you can say "cd ~/d/p" and hit tab to fill in even the intermediate parts. Given that the only path that matches "~/d./p." is ~/Dropbox/Public/, zsh will fill that in. If it hits an ambiguity down the proverbial road, it'll expand until it hits the ambiguity and list potential matches.

It's really convenient knowing that I can say something like "cd ~/d/c/s/c" and it'll fill in the rest.

edit: I tried to indicate a basic regex pattern but I think HN is reading it as trying to italicize "/p." instead. I later tried escaping it with backslash, but no luck. Sorry if that's really confusing.

alialkhatib··on Ask HN: What's the best 4K computer display available on the market?
Very true. You've summed a lot of the compromises that one has to make to enjoy the UP2414Q (or indeed pretty much any 4k monitor these days), and I really think that someday soon we'll get this right. Until then, these are the pains of being an early adopter. But I don't regret it :)
alialkhatib··on Ask HN: What's the best 4K computer display available on the market?
I'll state at the outset that, provided you have the ability to output using DisplayPort 1.2, you should consider the Dell UP2414Q, which I'm using and really like. Otherwise (and even in this case), you might be happier waiting it out.

There were a few caveats that prevented The Wire Cutter from recommending any monitor, but some of those don't seem to apply as readily now as they did at the time of last edit.

I got the Dell UP2414Q - a 24" ~4k monitor, making it around 183 ppi. The primary issues The Wire Cutter brought up in their rundown (of this display and generally) seemed to be the following:

- price (in this case $1100)

- the resolution being too high for the screen space

- poor refresh rates (30Hz in this case)

On the first point, you can now get it for around $750, maybe less, in the US. The second point was somewhat platform-specific, but on OS X you can tell it to display in HiDPI mode and it'll deal with the resolution in pretty much the same way that it deals with a retina Macbook Pro's 2880x1800 display. Namely, it'll scale everything at 2x. This is idyllic. If you have an rMBP, imagine that kind of sharpness across a 24" display and you're in the right ballpark.

The last point, about refresh rates, is trickier. My understanding is that this display has no trouble displaying at 60Hz once you enable it, but I have this display hooked up to a hackintosh, so myriad driver quirks make it impossible for me to say definitively whether the video card, the monitor, or the whole universe is conspiring against my success in trying to get it to display at 4k@60Hz. I think it's the buggy nature of the solution I came upon, which I could only describe as "not working in varying degrees". I'm still looking around, but I'm not optimistic for a real fix to get 60Hz for another month or two, if I'm lucky.

If I had a 2013 or later rMBP, I understand I could test it since that supports the DisplayPort 1.2 standard, but mine is from 2012. So I'm stuck testing with a desktop that shouldn't really even exist, heh.

So bringing it full circle, I recommend that you give waiting a chance. The Dell monitor I have has been fantastic, but not without its frustrations. If you can wait 6 months (or better yet, a year) I emphatically believe that the long term solutions to these kinds of displays will have shaken out by then. DP 1.2 and HDMI 2.0 will almost certainly be ubiquitous in video cards and probably in most laptops. It'll just be a lot easier to get things up and running.

That being said, if you're willing to put in some work and get frustrated a couple of times along the way, the reward is quite nice.

alialkhatib··on Real-time visualization of DDoS attacks
This is really cool. In the same space (ie real-time visualization of DDoS attacks) is DAEDALUS-VIZ, although I don't know if/where it was implemented for people to watch.
alialkhatib··on How to surprise your website visitors this Halloween?
This is unrelated, but if people using the HTTPS Everywhere extension are seeing a broken page, it's the extension's fault; the page's CSS and JS files are all being called by the HTTP protocol, which browsers tend to frown upon when the page is HTTPS (or vice versa).

To the author, you can avoid this by using a "protocol relative URL" (instead of calling for http[s]://... you would call for //...).

More details here: https://stackoverflow.com/questions/4978235/absolute-urls-om...

← PreviousPage 2 of 4Next →