Disqus - Latest Comments for aaronoliver

Re: We need to make our conference presentations more accessible

Aaron Oliver — Fri, 22 Feb 2013 15:33:50 -0000

I like the idea of making code examples available on some publicly-accessible repository, like github, which can then be referenced in the paper, on slides, etc. That might also allow for the revisions you mention in the last bullet.

Re: Did I Just Spot CSS In a Tron:Uprising Episode?

Aaron Oliver — Mon, 11 Feb 2013 12:09:15 -0000

Thanks, Les! That's still weird though, right? I mean, why chuck Google's CSS in there?

Re: Breaking It All Down: Simplify Solutions

Aaron Oliver — Wed, 02 Jan 2013 01:17:56 -0000

Har har. I can't breathe, I'm laughing so hard.

Re: Tool Thrash

Aaron Oliver — Sat, 26 Nov 2011 20:53:39 -0000

If *only* something like that existed....

Re: Tool Thrash

Aaron Oliver — Sat, 26 Nov 2011 20:51:20 -0000

We've been using Projecturf. It might be good for some, but for us it wasn't a great fit. Some things we found problematic:

* No permalinks to individual tasks, tickets, etc.
* "Tasks", "Tickets", "Discussions", and "Designs" were all different things, and hard to connect and navigate
* There was no sense of "status" on items. So no way to tell if something was "in progress", "complete" or "won't fix".

There's a few more. It just doesn't feel like tool good at tracking development.

Re: Multi-purpose Manifesto - Oliver Architecture Blog

Aaron Oliver — Tue, 15 Nov 2011 14:25:13 -0000

That's BS. Rails is the only framework you'll ever need. It does EVERYTHING. Guaranteed to keep your actual effort real close to zero.

Re: Dear Facebook, That Was Smart

Aaron Oliver — Sun, 25 Sep 2011 23:20:00 -0000

I think they would have announced F8 way earlier in that case. I think you'd agree that 70-hour weeks aren't a valid medium or long-term strategy for ending up with nice things.

Re: The Tourniquet Pattern Explained

Aaron Oliver — Thu, 05 May 2011 14:53:57 -0000

*sigh*. Ok, I'll update it to be accurately arcane

Re: The Tourniquet Pattern Explained

Aaron Oliver — Wed, 04 May 2011 10:48:25 -0000

It's pretty much the same. Kind of a "facade after the fact". The biggest difference is it's purpose and one's mindset when using it.

Re: The Tourniquet Pattern Explained

Aaron Oliver — Wed, 04 May 2011 06:58:32 -0000

Fixed! Thanks for the heads-up Jordan.

Re: Are Queues Un-Monitorable?

Aaron Oliver — Sat, 19 Mar 2011 14:22:57 -0000

Thanks Dave. That's what I was going for. Ted doesn't do comments on his blog, so I was wondering what some of the counter-points were.

Re: Don't Forget the Meeting AFTER the Standup Meeting

Aaron Oliver — Sat, 03 Jul 2010 13:42:57 -0000

Ahh, that sounds like heaven Abinesh....

Re: Recipe for a Death March

Aaron Oliver — Mon, 14 Jun 2010 13:05:07 -0000

Eelco, you're right. If minimal/no specs are delivered, then you have the same problem in reverse. I think the trick is to eliminate the "requirements are done" milestone, so the business-folk AND developers are collaborating every day to build the product.

Re: We're Gonna Need a Bigger Marker

Aaron Oliver — Tue, 08 Jun 2010 09:57:23 -0000

Awesome. Now I've got THAT song stuck in my head...

Re: I'm Waterfalling SO HARD This Week

Aaron Oliver — Fri, 16 Apr 2010 09:17:37 -0000

If you're suggesting what I think you're suggesting, meet me at the agreed-upon location at the agreed-upon time. Viva la résistance!

Re: Twin Ego Peaks

Aaron Oliver — Sun, 04 Apr 2010 08:42:46 -0000

Yep. I read a post somewhere that said each time you add a layer of abstraction, you introduce an infinite number of possible defects. Maybe a tad melodramatic, but helpful when figuring out the cost/benefit of adding a layer.

Re: The "Found Some" Data Access Problem

Aaron Oliver — Wed, 31 Mar 2010 12:25:05 -0000

Thanks Christian!

Re: The "Found Some" Data Access Problem

Aaron Oliver — Wed, 31 Mar 2010 12:24:53 -0000

My initial version of the exception didn't have the "found_bags" attribute, just the "bad_ids" list. My thought was that the caller could decide to do the set arithmetic and ask again for the good ones. That seemed a little wasteful though, and in the interest of being nice, I added the "found_bags" in there.

I think it works just as well if you don't include that, and just raise the error, plus some information (the bad ids) that can help the caller try to recover.

Re: Repent! Automated Web Testing is Inevitable

Aaron Oliver — Fri, 12 Feb 2010 20:31:32 -0000

Selenium RC (the remote control one) runs on Windows just fine, and should also work on Linux. Also, it's worth noting that the browser tests don't need to be written in the same language as the site. However, it can be handy if it is. That way you can do cool things like check the database for records after executing a test.

Re: Scrum's Devil's Advocate: Uncle Bob

Aaron Oliver — Tue, 09 Feb 2010 22:26:54 -0000

I think that's where we need to respect the authori-tah of "Individuals and interactions over processes and tools" There's an implied "just do it" or "don't get bogged down" in that line of the agile manifesto.

Re: You might “like” this.

Aaron Oliver — Tue, 09 Feb 2010 15:25:31 -0000

Shiny. Watch out Google Friend Connect/Buzz/etc...

Re: An Ideal Number of "Idea People", Please

Aaron Oliver — Tue, 11 Aug 2009 22:16:59 -0000

Excellent points. Too few people have read it. I was amazed on one project to learn that NONE of the managers had even HEARD of The Mythical Man Month.

Re: An Ideal Number of "Idea People", Please

Aaron Oliver — Mon, 10 Aug 2009 10:29:14 -0000

Thanks for the excellent comment, Matt.

In this case, feel free to replace "idea person" with "product manager" or "site designer"; basically the person or people that dictates the customer experience.

Looking forward to your response post.

Re: Commit EVERYTHING

Aaron Oliver — Fri, 07 Aug 2009 10:48:20 -0000

I'm the same way. As soon as I have something that's not broken, I commit it, then go back and make it pretty or optimize.

Re: Bikeshed: Where Does the Expected Value Go?

Aaron Oliver — Thu, 16 Jul 2009 08:29:46 -0000

I like to use the assert_equals() method from nose.tools, since it shows me what the values being compared were. Does that change matters? assert_equals(expected, actual) or assert_equals(actual, expected)?