Disqus - Latest Comments for peteonrails

Re: Alpina's Startimer Pilot is Simple, Yet Elegant

peteonrails — Fri, 30 Mar 2018 11:00:47 -0000

I saw this advertisement disguised as an article in the magazine a couple of days ago. You describe a watch "designed with aviators firmly in mind" and then fail to mention whether it has a GMT hand or a timer, or really any other feature that aviators find important.

Yet you are really excited that it is water resistant to 330 feet (you know, the depth at which we fly our airplanes).

It makes me wonder whether the Flying Staff who claims the byline really did write the copy for this advert, or whether the manufacturer provided the words you printed. If it's the former, you've done a bad job of evaluating a new product, and I am surprised. If the latter, you ought to at least mark the page as an advertisement.

Either way, I'd prefer to see an honest, fair evaluation of a new and expensive aviation watch. Do you really think this thing is great for pilots? Why or why not?

Re: Exclusive: FAA probes Air Canada near-miss at SFO

peteonrails — Tue, 11 Jul 2017 15:37:12 -0000

28-R is not the runway on the left. It's the runway on the right. The taxiway is to its right. Coming from the other direction, that runway is Runway 10-L.

Re: Exclusive: FAA probes Air Canada near-miss at SFO

peteonrails — Tue, 11 Jul 2017 15:34:31 -0000

Taxiway lights are blue. Runway lights are white and have REIL indicators. Things happen, but it is not hard to tell the difference.

Re: Supervisor threatens to hang worker for drinking from 'white people' fountain

peteonrails — Thu, 05 Jun 2014 08:01:12 -0000

"So they are to blame since they didn't know about it?"

Yes.

Re: Supervisor threatens to hang worker for drinking from 'white people' fountain

peteonrails — Thu, 05 Jun 2014 07:14:51 -0000

"Federal Compress very much regrets that the allegations were not reported to it when the first incident is claimed to have occurred."

It's Federal Compress' fault. Nobody else's. It's not hard to see how a company that fails to accept responsibility and apologize directly, choosing instead to deflect blame onto their customers, might harbor a culture that allows this kind of behavior to develop in the first place.

Re: Open WhisperSystems >> Blog >> Open Whisper Systems is coming to iPhone!

peteonrails — Thu, 06 Feb 2014 11:43:51 -0000

I look forward to an iOS release. I also get it that you're building a free, open source product and snarky nagging about the delivery date only serves to make you less excited about building it. :)

Good work, keep it up!

Re: Lessons Learned: The First Two Years of Running a Software Consultancy

peteonrails — Wed, 25 Dec 2013 12:24:38 -0000

You wouldn't have gotten away with leaving it out! And it's an important part of company lore now.

I enjoyed last year's retrospective and this one. I look forward to the next one.

Re: Littlelines | Matt Sears | A Guide for Writing Maintainable Rails Tests

peteonrails — Mon, 23 Dec 2013 11:46:15 -0000

Heya Dan: I agree that you often want to test that a user has orders: see my comment above. What do you think about scoping down the tested interface to only those associative methods that you actually use? eg:

expect (user).to have_many(:orders) => expect(user).to respond_to?(:orders)

For most cases I think that's a good test that allows for refactoring, though for more tightly coupled relationships the has_many() matcher may be appropriate. I think there's a tipping point in there somewhere.

Re: Littlelines | Matt Sears | A Guide for Writing Maintainable Rails Tests

peteonrails — Mon, 23 Dec 2013 11:43:07 -0000

Good points, all.

I agree that testing it { should have_many(:orders) } may be the wrong test. What you're really interested in is testing that you can ask a User for orders, and it gives you back a set of them. To Dan's point, { should have_many(:orders) } is a shorthand way to check for all of the build_* and other associative methods. Usually, though, you only want to test a subset of them, which would allow a refactor later. 90% of the time, what you really want is: { should respond_to?(:orders) }

As far as mocks and stubs go: I have found little utility in the user.any_instance.should_receive(:orders).and_return(fake_orders) expectation syntax. Every time I have written one of those, it has broken later for no good reason. Those kinds of false negatives discourage good testing and you are right to call that out. FactoryGirl has been a big win for many developers recovering from over-mocking and over-stubbing.

In my work at FasterAgile, we tell our clients that 100% test coverage is not an unreachable or unreasonable expectation. I have noticed that developers who are new to TDD often have trouble grokking what to test in the first place. Some will structure unit tests like acceptance tests. Some can't do an acceptance test without tons of setup and mocking. Many will give up and just test manually. In all cases, writing our requirements as pending tests has been the key to drive behavior change.

Thanks for the article!

Re: Andrew Palmer Sends Jaws II (5.15a)

peteonrails — Tue, 22 Oct 2013 16:24:31 -0000

That's quite an accomplishment, congratulations! I'm kicking myself for being to lazy to hike up the hill on Sunday. :)

Re: What A Time To Be 17 Years Old

peteonrails — Mon, 21 Oct 2013 09:38:47 -0000

Guide him into a field where he can start a business that he will be successful with. Plumbing is fine, but not if you're going to be a plumber for 40 years. Own the plumbing company. Working your ass off so that someone else can get all the money is still a kind of slavery.

Re: Goodbye TextMate | Pete On Rails

peteonrails — Mon, 02 Sep 2013 10:28:11 -0000

Hi Kevin. This is a really really old article. Today, I use Sublime Text on my linux machine. Give it a go!

Re: Goodbye TextMate | Pete On Rails

peteonrails — Sun, 02 Jun 2013 16:05:19 -0000

It's a pretty old article, so I'd argue that it's no longer timely. However, the folks who read my RSS feed were interested at the time. I'm sorry that you didn't like the article.

In the future, please remember Rule #1: Don't Be A Jerk.

Thanks,

--Pete

Re: Herb Conn Dies at 91

peteonrails — Fri, 10 Feb 2012 20:10:12 -0000

Herb and Jan put up many of the classic routes that I started climbing on about 20 years ago. Though it's of an easy grade, Conn's East at Seneca Rocks is one of the most wildly exposed pitches I've ever set foot on. I'm already planning my trip back to WV to climb it again in his memory. You'll be missed, Herb. Thank you for a lifetime of exploration, and for sharing it with the rest of us.

Re: The Revenge of Inigo Montoya

peteonrails — Thu, 21 Jul 2011 09:57:42 -0000

I don't usually feed the trolls Derek, especially not on something I
wrote over a year ago. But in this case, I think a response is
warranted.

You're conflating a bunch of issues in your argument, but here's my best go:

1. Yes, I read all of the code behind ci_auth. The :false symbol may not
have been evaluated anywhere for its truthiness in the library itself,
but I assure you, it caused a bug that I spent a long time tracking
down.

2. Yes. I understood the requirements of authentication at said company.
I stand by my assertion that rolling your own security / authentication
library was a mistake.

3. There's a reason there are no tests in The Mod. You'll have to ask
David why, though. I'm really sorry that you're stuck maintaining that
code base for more reasons than the fact that it's got poor unit test
coverage.

In my original post, Derek, I didn't rag on your code. I didn't mention
your name, your employer, the project in question, or even say it was a
huge steaming pile of shit. If you think The Mod codebase sucks, then
you should make the case to your bosses for a rewrite, rather than rag
on me for making the best of a shitty situation with your stakeholders.

Just don't put "if case == :false" in there anywhere and I'll be happy.

Re: Goodbye TextMate

peteonrails — Tue, 29 Mar 2011 09:11:25 -0000

It's not THAT outrageous. :)

Re: Goodbye TextMate

peteonrails — Tue, 29 Mar 2011 09:11:04 -0000

No doubt! I ended up using RedCar. I found it to be a nice happy medium.

Re: Goodbye TextMate

peteonrails — Tue, 29 Mar 2011 09:10:26 -0000

I used to LOVE emacs. Over time, I migrated away from it. I can totally understand the cult-of-emacs crowd's love of the most long-running successful editor in the universe, but I never stay with it for long. I'm the same way with vi, vim, and macvim. I get it, but I don't get it. :)

Re: Goodbye TextMate

peteonrails — Tue, 29 Mar 2011 09:08:19 -0000

Thanks for writing, Adrian. In all candor, my switch to NetBeans was fairly short-lived. I didn't last the whole year. Based on Tom's feedback 6 months ago, I started using Redcar, which is everything I like about TextMate, and nothing that I hate about TextMate, plus it's free.

Re: Goodbye TextMate

peteonrails — Tue, 28 Sep 2010 09:19:09 -0000

Thanks Tom! Haven't tried it, but I'm giving it a whirl now. I appreciate the feedback.

Re: Why I Don't Like Stock Buybacks

peteonrails — Fri, 25 Jun 2010 11:15:42 -0000

I always have viewed a stock buyback as a signal that the company it under-performing in the innovation department.

That said, I think it is a responsible thing to do if management can't adequately invest my capital in innovative projects. Giving me my money back is a good thing.

The different take I have is that I'd take the money and run: to invest it in a company that needed the cash so that they could change the world.

Re: VoteFu is now available on GemCutter

peteonrails — Thu, 03 Jun 2010 20:38:28 -0000

I am glad you got it working, but your original point about an update to the docs being required still stands. Thanks for the analysis of what went wrong: it's either a bug I need to fix, or an incompatibility that I need to document.

Take care!

Re: VoteFu is now available on GemCutter

peteonrails — Tue, 01 Jun 2010 15:29:35 -0000

Yes, I concur. The documentation is currently very bad. It got outdated over time, and I've put off updating it until I can get the new version out the door.

If you send me a gist or snippet with your migration and your model I can help you get it working.

Re: The Revenge of Inigo Montoya

peteonrails — Tue, 25 May 2010 16:05:09 -0000

Other implementations of The Inigo Montoya Attack include cargo-culting things like the bang-bang boolean trick when you don't really need it:

val = !!false

Why would you do that? Why?

Re: The Revenge of Inigo Montoya

peteonrails — Tue, 25 May 2010 16:00:26 -0000

I have seen this in exactly one codebase. To protect the innocent, I will not say whose codebase it was, but I will tell you that our friend in common Kate probably knows why the symbol :false is used in place of 'false'.