Disqus - Latest Comments for akahn

Re: ActiveRecord is reinventing Sequel

Alex Kahn — Fri, 04 Mar 2016 11:04:26 -0000

Thanks, I missed that link.

Re: ActiveRecord is reinventing Sequel

Alex Kahn — Fri, 04 Mar 2016 10:44:19 -0000

You haven't really demonstrated that Sequel is better than ActiveRecord, just that ActiveRecord is playing catch up with Sequel. I agree with you that Sequel is better, and I prefer using it, but having more features sooner doesn't automatically make something better.

Re: A Primer on Protocol Buffers » Blake Smith

Alex Kahn — Tue, 30 Sep 2014 12:04:03 -0000

Check out beefcake: https://github.com/protobuf...

Re: Welcome Mat Marquis

Alex Kahn — Wed, 13 Aug 2014 17:24:18 -0000

Who *didn't* see this coming?

Re: Salmon Tofu Cakes Recipe - Viet World Kitchen

Alex Kahn — Tue, 24 Jun 2014 13:03:26 -0000

Purslane from the garden?

Re: Why mixed responsibilities resist refactoring

Alex Kahn — Wed, 27 Feb 2013 10:42:00 -0000

Another argument in favor of not using a mixin for fraud checking behavior is that in the class-based example, it's possible to test the fraud checking class in isolation, without loading Rails or the Receipt model. Instead of constructing a FraudChecker instance with a real Receipt instance, you can pass in a fake which responds to the methods FraudChecker calls. In addition to running fast, avoiding DB interacting, and avoiding Rails startup time, this has the benefit of making FraudChecker's dependency on Receipt's interface more explicit in your tests.

Re: On Rails, Sinatra, and picking the right tool for the job

Alex Kahn — Thu, 13 Sep 2012 11:52:25 -0000

Do the Sinatra API apps share model code with the main Rails app, since presumably the API carries out similar operations to what users do on the website?

Re: Kestrel in Production at Papertrail

Alex Kahn — Fri, 17 Aug 2012 10:52:50 -0000

Thanks for the post. Can you give a real world example of how Kestrel (as opposed to Resque) has allowed you to build a more flexible, more robust or just better architecture? I understand how the consumer of a message is less tightly coupled to the producer (the Rails app) in Kestrel, whereas in Resque the job itself is assumed to live in and be part of the same codebase as the app itself, but that section still isn't clicking for me.

Also, s/flexability/flexibility/.

Re: Don't make your code "More Testable"

Alex Kahn — Wed, 11 Jul 2012 12:35:57 -0000

I would love to see some example code written in a way that solves the problems you outline here. Especially #3, with events and listeners. I find myself writing code like the example code contained in this post and would like to go further.

Re: How much should I refactor?

Alex Kahn — Mon, 25 Jun 2012 11:32:00 -0000

I appreciate that you point out the balance between short term productivity and long term maintainability. Over the weekend, Matt Wynne expressed this same point in a talk at Gotham Ruby Conference.

However, I think you overstate the cost of jumping around codebases and projects. With adequate tools – Vim with ctags, bundler.vim, gem-ctags – this cost is nearly eliminated.

The joking criticism of SmallTalk was that "everything happens somewhere else". The opposite extreme, everything happening in the one place, is worse. The optimal place is somewhere between these two extremes. I lean toward the former, though, because I hope that once I find the relevant part of the code, it will be easier to understand.

Re: Hashes and encapsulation: Jon Leighton

Alex Kahn — Thu, 12 Jan 2012 10:27:34 -0000

I misunderstood. I thought the commenter was proposing DOMElement#[], so you would access attribute values like: el['href'].

Re: Hashes and encapsulation: Jon Leighton

Alex Kahn — Thu, 12 Jan 2012 10:15:34 -0000

How so? [] would just be a different name for the method you called attribute.

Re: Rails, Models and Business Logic

Alex Kahn — Sat, 07 Jan 2012 19:08:56 -0000

Just came across this post on twitter via Andrzej Krzywda (via Sean Allen).

I like where you're going. I think that you should go a step further with your mixin idea, though. Although you've separated the Braintree functionality into its own module, splitting it off into its own file outside of the ActiveRecord model, you haven't reduced coupling. By that I mean ActsAsSubscriber knows the intimate details of how the Customer class works. You can't test the former in isolation of the latter.

You'll get a more maintainable, extensible, and testable design if you establish a stricter separation between the concepts of Customer and Subscription. You can then define a narrow interface between the two objects, where they only pass as little information to each other as each needs. The dependencies of each class are reduced. This is composition. I can show you what I mean in code if you're interested.

Here's an older post that you may have seen before that argues for what I'm saying, and also argues that sometimes Rails features like callbacks and observers can actually be pretty problematic when it comes to maintaining a project: http://jamesgolick.com/2010... Also check out Avdi Grimm's and Gary Bernhardt's writings and videos on this subject.

Re: WEBSITE OF THE DAY: Paperless Post

Alex Kahn — Thu, 10 Nov 2011 10:18:44 -0000

I work for Paperless Post and would like to clarify what we do. We do not send physical stationery. Our cards are delivered over email. Hence the name "Paperless." ;)

Thanks for the writeup.

Re: Alex Kahn | Working Through Git Conflicts in Vim

Alex Kahn — Wed, 12 Oct 2011 21:11:18 -0000

Great suggestions, thanks.

Re: Alex Kahn | Working Through Git Conflicts in Vim

Alex Kahn — Wed, 12 Oct 2011 15:26:16 -0000

Most of my merge conflicts are fairly straightforward so I don't mind merging them "by hand". However, I do want to learn to learn to use fugitive.vim's more advanced merging tools for more challenging merges.

Re: Alex Kahn | Working Through Git Conflicts in Vim

Alex Kahn — Wed, 12 Oct 2011 15:11:43 -0000

Indeed, gc is the git subcommand for garbage collection. This is just a shell alias, though.

Re: What do you want to know about OOP and Rails?

Alex Kahn — Mon, 12 Sep 2011 13:37:15 -0000

I remember seeing a tweet positing that not encapsulating Rails associations is the cause of some very high percentage of slow tests/design problems/societal collapse, etc. I found that intriguing. Maybe you can write a bit on that topic.

Re: MongoDB as a Key-Value Solution for Logging Events Data

Alex Kahn — Tue, 26 Apr 2011 16:14:43 -0000

Sorry, that's written in Java -- gotta install JVM. Next!

Re: MongoDB as a Key-Value Solution for Logging Events Data

Alex Kahn — Tue, 26 Apr 2011 16:11:14 -0000

Riak is written in Erlang. Sorry, not popular enough.

Re: http://trymytrust.tumblr.com/post/4422223542

Alex Kahn — Sun, 10 Apr 2011 16:26:19 -0000

ERROR :(

Re: Jumping Long Distances in Vim

Alex Kahn — Sun, 27 Feb 2011 16:13:01 -0000

Very helpful. I tend to page with C-b and C-f. I also tend to use z. and z<enter> but zt and zz actually seem better since they don't change the cursor position.

Re: How-To: Conditional HTML tag with HAML

Alex Kahn — Sat, 26 Feb 2011 16:27:11 -0000

What do you think of this approach? https://groups.google.com/f...

Re: http://failingfast.tumblr.com/post/3430289830

Alex Kahn — Mon, 21 Feb 2011 19:25:55 -0000

dylan would love this

Re: http://trymytrust.tumblr.com/post/3405331269

Alex Kahn — Mon, 21 Feb 2011 09:29:28 -0000