Disqus - Latest Comments for Peeja

Re: Writing Om Next Reloadable Code — A Checklist

Peter Jaros — Fri, 09 Sep 2016 18:09:56 -0000

Also, you should be able to replace `(om/class->any reconciler RootComponent)` with `(om/app-root reconciler)`.

Re: Writing Om Next Reloadable Code — A Checklist

Peter Jaros — Fri, 09 Sep 2016 17:55:16 -0000

`.forceUpdate` doesn't appear to work recursively, so it only updates the RootComponent. Do you know of a way to get every component to update without remounting the app and losing local state? There's a project that relies of React internals to do it, which I may try: https://github.com/gaearon/...

Edit: My current solution, based on react-deep-force-update, which isn't pretty but appears to be working: https://gist.github.com/Pee...

Re: My Increasing Frustration With Clojure

Peter Jaros — Mon, 13 Jun 2016 10:37:43 -0000

It's hard to distinguish between entitled and caremad in a written medium. It's frustrating to invest a great deal of time and energy—and in the case of a business, money—in a language and then feel like that language is falling apart and that you're powerless to help.

I don't necessarily agree (or disagree) that this represents a breakdown and failure for Clojure and its maintainers, but when someone feels this way, a critical post like this, written respectfully, is an entirely appropriate way to start a conversation about it.

Re: Routing in Om Next — a Catalog of Approaches

Peter Jaros — Sun, 12 Jun 2016 22:11:13 -0000

What I mean is that there needn't be *any* mapping from the route keywords to the refs. The ref can always be the same constant, regardless of which route is active. Unless I'm missing something.

(BTW, Disqus's emails seem to think your blog is hosted at http://localhost:4000/.) :)

Re: Routing in Om Next — a Catalog of Approaches

Peter Jaros — Sun, 12 Jun 2016 14:05:06 -0000

Does the `:ref` have to be the name of the route, or even unique to the route? It seems to me that you could use a constant ref, such as `:page`, to refer to the routed component. It'll find the correct component—and the correct subquery—either way, no?

Re: Shut Up & Sit Down | SU&SD Play... Funemployed

Peter Jaros — Thu, 26 Nov 2015 10:35:58 -0000

> but it is the kind of game that people who generally hate party games will use to describe why they hate party games.

I think that depends on why the person hates party games. There are some people who hate them because coming up with creative ideas on the spot isn't their idea of a fun time, as you say. But the people I know who hate Apples to Apples and Cards Against Humanity hate them because they just involve Lego-ing two ideas together and hoping it's absurd enough to make someone laugh. For them, party games should involve *more* creativity. So for those people, Funemployed (and Snake Oil) give the party game hope.

Re: Vorthos 2015

Peter Jaros — Wed, 22 Apr 2015 11:45:58 -0000

That's pretty neat. Though I think I identify as all five Vorthoses *and* Melvin. :p

One nitpick, though: I wouldn't say that Timmy does things to have fun, or that Spike has something to prove. I'd say Timmy, Johnny, and Spike are each defined by *how* they have fun. Timmy has fun by being powerful. Johnny has fun by making creative choices. Spike has fun by honing the skill of winning the game. (It's easy to forget that Spike can lose and still have fun, but Spike's goal is to learn to win more over time.)

Re: 14 Weird Things New Yorkers Think Are Totally Normal

Peter Jaros — Tue, 30 Dec 2014 15:24:58 -0000

I have, in fact, said this, and this has been said to me.

Sometimes you just get a craving for street meat.

Re: Private: Why I’m a Narcissistic Asshole

Peter Jaros — Fri, 21 Nov 2014 12:18:00 -0000

But there's another way to use Instagram, and any social media: to stay in touch with family and friends. I use Instagram almost exclusively to share things from my life with people I actually know. Each like I get, when I get a like, is an individual message from that person acknowledging that they've witnessed something that's happening in my life. I have absolutely no use for the like *count*. It's meaningless to me. I tend to use Facebook the same way.

With Instagram in particular, I've never understood following strangers. At least popular Viners are doing interesting things. The top Instagrammers seems to be just…taking selfies? I don't get the appeal. But then, the wonderful thing about the Internet is that I don't have to. You may as well ask why kids love the taste of Apple Jacks.

Re: Shopping

Peter Jaros — Thu, 20 Nov 2014 09:57:17 -0000

I'm pretty sure Polymer is some sort of 3D printed shoe funded on Kickstarter.

Re: Review: Dog Eat Dog

Peter Jaros — Wed, 22 Oct 2014 09:23:36 -0000

I've been taking the Irish course, and it's great. Irish dialects vary quite a bit, though. Duolingo's course teaches "Standard Irish", but I'm not clear on exactly what that is or where it's spoken.

Re: Dropbox wasn’t hacked

Peter Jaros — Tue, 14 Oct 2014 14:53:29 -0000

You say your Gmail password was the same as your Dropbox password? What makes you think the password came from Dropbox's database? It's just as likely to have come from Gmail, or from a site that looks like Gmail that managed to phish you, or from a third service you used the same password with. There's nothing stopping the attackers from using that password to log into your Dropbox first once they've got it.

Re: Dropbox wasn’t hacked

Peter Jaros — Tue, 14 Oct 2014 14:34:41 -0000

How did they get into your Gmail account through your Dropbox? It sounds more like they got into your Dropbox though your Gmail by resetting your password.

Re: The Proper Pronunciation of Clojure's Assoc - Squid's Blog

Peter Jaros — Tue, 29 Jul 2014 10:02:04 -0000

I pronounce it "assok". But I can change. If I have to. I guess. (QOFM)

Re: Revisiting the ErgoDox

Peter Jaros — Tue, 29 Jul 2014 09:57:58 -0000

How is typing on the Clears over time vs. the Blues? I have arthritis, and I'm trying to be kind to my fingers. I have a Filco with Blues right now; I was hoping the auditory feedback would help me stop from bottoming out, but I'm not sure it's done much. But I understand the Clears have a bigger bump than the Blues?

I'm thinking of going Ergo Clear eventually, but taking the time to mod a keyboard's worth of switches is a big investment. I'm considering building an ErgoDox with Clears and modding them later if I feel like it.

Re: Verizon made an enemy tonight

Peter Jaros — Fri, 18 Jul 2014 08:32:13 -0000

That's not a stunning endorsement of Google Fiber.

Re: The FCC Was Hacked After John Oliver’s Net Neutrality Troll Call

Peter Jaros — Tue, 15 Jul 2014 15:37:12 -0000

Also, the article says this kind of "attack" consists of repeatedly submitting comments and then searching for them.

Isn't that exactly what you'd expect everyone to do? Submit a comment and then check to make sure it made it into the system? It doesn't take much for a popular show to DDoS a 17-year-old system by accident.

Re: Want to Get Out Alive? Follow the Ants - Issue 13: Symmetry - Nautilus

Peter Jaros — Sun, 11 May 2014 11:21:50 -0000

I think that's the point. There's always going to be one exit that's more appealing than the others on average, but the most efficient way to get everyone out is for some individuals to use what would be (in the absence of the rest of the people/ants) a less appealing exit. Instead, the ants almost all go to one exit, which slows them down as a group.

Re: The U.S. Government: Paying to Undermine Internet Security, Not to Fix It

Peter Jaros — Wed, 16 Apr 2014 10:02:59 -0000

The entropy is lower, but as it turns out, that doesn't really matter. A brute force attack on a password of that length is still going to take way too long. Using song lyrics lets you make a giant password that you'll actually remember.

Brute-forcing a password using a dictionary of English words is reasonable thing to attempt. Brute-forcing with a dictionary of song lyrics wouldn't work. On top of an absurd number of possible songs, there are too many variations on where to begin and end the quotation, not to mention whether to include spaces or punctuation, and if so how to punctuate it. Realistically, an attacker would have to brute-force the entire space of strings, which means that length is much more important than entropy.

For a substantially better password, throw a single digit into it. For more thoughts on that idea, see GRC's Password Haystacks page: https://www.grc.com/haystac...

The best password, of course, is completely random, and stored in a secure password manager. But sometimes you need a password you can remember.

Re: The U.S. Government: Paying to Undermine Internet Security, Not to Fix It

Peter Jaros — Tue, 15 Apr 2014 13:57:17 -0000

There's one other really important rule for passwords: never use a single dictionary word for a password. But feel free to use a dictionary word *in* a password. "elephant" is not a good password, but "elephanttoenail" is actually pretty decent.

Guessing words from the dictionary is sometimes feasible for an attacker, depending on the vulnerability they're exploiting. Guessing *combinations* of dictionary words takes a lot more guesses.

Using sentences or lines from song lyrics is a great way to make long, memorable passwords which are hard to guess.

Re: SADD – Redis

Peter Jaros — Tue, 25 Mar 2014 16:31:17 -0000

You're looking for a list. Sets are explicitly unordered, which allows them to be searched faster. A list is kept in the order you push elements onto it.

Re: SADD – Redis

Peter Jaros — Tue, 25 Mar 2014 16:23:03 -0000

Yes. At least, that's typically how a set is implemented in general. It's essentially a hash table with no values: you're only interested in whether a key is present. That also makes SISMEMBER O(1). The tradeoff (compared to a list) is the loss of ordering information (which you might not care about anyhow).

Re: Yes, You Should Write Controller Tests!

Peter Jaros — Fri, 07 Mar 2014 15:00:20 -0000

Speaking as someone who avoids fixtures like the plague, do you have any resources that describe a "right" way to do fixtures?

That's the part that always kills me: almost every test needs its own fixture object, and they come to depend on one another, and pretty soon you can barely touch the fixtures without breaking everything. Plus you've got data in the fixtures that's not actually important to any tests anymore, but it's hard to be sure, so it sits there and rots until the end of time.

Re: US Scientists Convert Algae to Crude Oil in Less Than an Hour

Peter Jaros — Thu, 26 Dec 2013 14:55:58 -0000

Oh, god, I have no idea. It's not a proposal. I'm just explaining how a fuel which releases carbon when it's burned can be carbon-neutral over the whole of its cycle.

If we only produce carbon from fuels at the same rate that plants fix carbon, our fuel use will be carbon neutral. Whether that's feasible or not is something I'm not nearly qualified to speculate about.

Re: US Scientists Convert Algae to Crude Oil in Less Than an Hour

Peter Jaros — Mon, 23 Dec 2013 14:37:56 -0000

I don't know of any particular studies, but the Crash Course Ecology episode on the Carbon Cycle explains things pretty well: https://www.youtube.com/wat...