Disqus - Latest Comments for tswicegood

Re: Python Patterns: kwargs helper method - TravisSwicegood.com

Travis Swicegood — Sun, 08 Feb 2015 20:01:26 -0000

@tanenn, that would work in this example's case, but not fails to address a few things things: 1) objects that are have properties that are not accepted as kwargs by __init__; 2) values where the kwarg and the property name do not match. In the case of Werkzeug's Rule object, both of these present problem. Additionally, it does not allow a hook in to third-parties to actually change much. They have to change how their object is actually represented at the __dict__ level to have full control which just seems like a major source of potential trouble.

Using vars(self) is a simple solution, but errs on the side of being too simplistic.

Re: Rethinking Web Frameworks in Python - TravisSwicegood.com

Travis Swicegood — Sat, 26 Jul 2014 14:54:57 -0000

I moved this code out of a gist and into a repo with a bit of cleanup here: https://github.com/tswicego...

Re: Timeless Way of Coding - TravisSwicegood.com

Travis Swicegood — Mon, 23 Jun 2014 19:37:56 -0000

Doh! You're correct, it's flake8 that complains for me, not pyflakes.

False positives on pep8? I haven't run into that.

Re: Timeless Way of Coding - TravisSwicegood.com

Travis Swicegood — Mon, 23 Jun 2014 11:02:30 -0000

You wrote it correctly -- the code in my post has a space between the keys string and the colon.

Re: Understanding the range of programming and data skills a journalist needs

Travis Swicegood — Thu, 20 Feb 2014 15:56:40 -0000

This sounds great. I think there's a lot of potential here. One thing I would caution against is trying to find the 1:1 CS to practical programming approach. It sounds like you've taken the right courses at Stanford, but a lot of university are teaching a very theory heavy CS that isn't immediately applicable to people trying to learn how to do stuff. At some point, knowing sort algorithms is a good thing, but not when you're trying to show a list of restaurants and their health ratings.

Can't wait to see what you guys accomplish at Texas State when you get back!

Re: Tools vs Materials - TravisSwicegood.com

Travis Swicegood — Tue, 21 May 2013 10:59:22 -0000

Totally agree! Use whatever tools work for you, but at the end of the day implementation needs to be HTML/CSS/JS. Without those, all you're doing is abstractions.

Re: Java WAR-like files for Python

Travis Swicegood — Thu, 08 Mar 2012 12:34:48 -0000

zeeg: the only thing they don't do is handle system dependencies

Re: Java WAR-like files for Python

Travis Swicegood — Thu, 08 Mar 2012 12:31:14 -0000

fwiw, PHP has this as of >= 5.3 -- it's called phar

Re: Java WAR-like files for Python

Travis Swicegood — Thu, 08 Mar 2012 12:27:21 -0000

this seems like something that could be done with self-contained virtualenv instances that were executed from the outside. the only problem is having it self-contained so you could do `python --some-flag application.pyapp`

Re: Java WAR-like files for Python

Travis Swicegood — Thu, 08 Mar 2012 12:25:13 -0000

zeeg: that's where my mind went to with this immediately...

Re: Java WAR-like files for Python

Travis Swicegood — Thu, 08 Mar 2012 12:23:06 -0000

I always thought eggs were really close from the outside

Re: Where Iphone App Dashboard Breaks Down

Travis Swicegood — Tue, 06 Mar 2012 00:38:28 -0000

I've found myself thinking the same thing. I spend most of my time in the "recent apps" mode (or that's what I call it) or switching back to search. Seems like some combination of the two would much better. Maybe the iPhone will take a queue Mission Control in Lion and enable an auto-layout based on usage. Keep the first list of apps something you can curate, then all of the other's are sorted based on last use/frequency of use.

Re: The ol' bait and switch: Twister Group is a fraud - TravisSwicegood.com

Travis Swicegood — Sat, 18 Feb 2012 14:59:04 -0000

Ya know, it's amazing that you still feel the need to monitor this thread and answer every single problem you have. You're a horrible business person who threatens people, then when exposed blames it on other people. Please quit using my blog as a way to try to play damage control when you screw people over.

Note to readers: the above post is the last post that is going to be tolerated from the Twister group. Future comments will be removed as spam.

Update: As a footnote to this, no fewer than 5 new comments came through from the Twister group yesterday afternoon after posting this.

Re: http://blog.mattwaite.com/post/16942178252

Travis Swicegood — Fri, 03 Feb 2012 17:49:35 -0000

Matt, said this on Twitter, but I'll say it here too so people reading the comments can see it. For the love of all things holy, don't start them out using python-twitter. It's horrible. You go over to the dev.twitter.com and find all of these cool things you want to do, and then have to figure out how in the world to do it using python-twitter which has invented their own API.

A much better solution would be using [requests][] from Kenneth Reitz. It's a minimal wrapper around HTTP calls that allows you to use the actual API rather than having to learn two things.

[request]: https://github.com/kennethr...

Re: 30 second apache/nginx/gunicorn timeout — Reinout van Rees' website

Travis Swicegood — Sun, 27 Nov 2011 17:14:43 -0000

Celery is dead simple to set up. The only area I've had issues with it is setting up RabbitMQ. Erlang uses the hostname and on most consumer machines the hostname isn't actually pingable, so you get a connection timeout. I ran into that before it was documented though---it's one of the "gotchas" in the Celery docs now.

I'd definitely set it up and make sure it makes it into your book. No Django book is complete without covering it.

Re: 30 second apache/nginx/gunicorn timeout — Reinout van Rees' website

Travis Swicegood — Sun, 27 Nov 2011 10:27:36 -0000

+1 on caching plus a Celery PeriodicTask to update the cached version every N seconds. Unless there's an easy way to know when the data is updated (a model save signal, for example).

Remember, that 60 seconds is time that a gunicorn worker isn't able to respond to anything since its blocking. If it's blocking while waiting on the DB or some such, you might want to consider the async workers so at least it can deal with other requests while waiting on the service to response.

Re: MattDeBoard.net

Travis Swicegood — Sat, 18 Jun 2011 13:21:27 -0000

Congrats! Would love to see your work with Django + Fabric if you go that route. Be sure to post it to the mailing list or Convore if you do!

Re: http://hacktyler.tumblr.com/post/5322426478

Travis Swicegood — Mon, 09 May 2011 14:09:49 -0000

Welcome to Texas. Though, you will have to start getting used to saying "I work at the Chicago Tribune" instead of just "the Tribune" or people will think you're working for us. :-)

Keep an eye on the Austin PUG and AWPUG on Meetup -- we have a few weekend hack days throughout the year that you could come down for if you want to geek out down here.

Re: http://localhost:4000/2011/04/24/twitter-timeout/

Travis Swicegood — Tue, 26 Apr 2011 14:54:49 -0000

Very well played. :-) I actually considered yanking them and forcing people to use trackbacks when I went to a static site, but decided this was the better option.

Re: Packaging reuseabe & testable Django apps with virtualenv, pip, and Fabric

Travis Swicegood — Mon, 04 Oct 2010 09:34:48 -0000

I've attempted to use virtualenvwrapper, and personally, I think it's overkill to the Nth degree. It may have served a purpose at some point in the past, but now I don't see any reason to use it other than you don't want to full with the various ignore files for your VCS of choice.

Re: Buying an Mahattan Co-op

Travis Swicegood — Fri, 27 Aug 2010 01:30:54 -0000

Catching up on some old blog posts… Congrats!

It's amazing the hoops you have to jump through to buy a place. My wife and I went back and forth with the bank at least a half dozen times. "We need this piece of paperwork and we're set." The next day: "Now just this one more piece..." and on and on. Thankfully we didn't have to contend with a co-op or HOA at least. :-)

Re: http://www.travisswicegood.com/2010/08/24/razors-and-development/

Travis Swicegood — Tue, 24 Aug 2010 23:46:23 -0000

This is actually the reason for using the Data Mapper pattern. All of the location logic is in the mapper class with the model free to be a dumb container of data and the logic to manipulate it. DM support backed in as part of Rails 3 is one of the most interesting parts of that release, in my opinion.

Re: Our Git Workflow: Private Development, Public Releases - Braintree

Travis Swicegood — Tue, 17 Aug 2010 14:51:40 -0000

Squashing commits is definitely a useful tool at reducing noise, but for most projects its not necessary and the way Braintree uses it effectively hides all development behind a curtain. Having a bunch of commits that show the evolution of a design can be helpful to someone working as a "code archaeologist" trying to understand the process of the code. If all you show is the pristine, ivory tower you hinder that process.

That's not to say that every dead end should be preserved, but I've found that once --squashed is introduced, everything gets squashed and the true history is lost. This is especially important for third-party developers working with your code. Having a bunch of single version commits that contain everything in it means that developers that are using your code have to wade through large diffs in order to figure out what has changed and being able to track it as it evolves.

I've worked for an open-source company in that did this via its releases. Everything got pushed out at a zip/tarball once it was ready for beta, much to the chagrin of the large community of developers around the project.

Re: http://www.travisswicegood.com/2010/08/09/telecommuting-culture/

Travis Swicegood — Tue, 10 Aug 2010 10:35:32 -0000

Fun side note: if you search for "telecommute" on GitHub's new job site, only one of the six positions actually allows telecommuting, the other 5 are saying they don't.

Re: http://www.travisswicegood.com/2010/08/09/telecommuting-culture/

Travis Swicegood — Tue, 10 Aug 2010 09:59:22 -0000

My guess is that you've never actually done this. It's nearly impossible for you both to run the computer doing different things. As one person's typing, you're sitting there waiting on them. Just as if they had the keyboard.

There are some interesting comments along these lines on Hacker News from people with experience in both. Definitely worth reading.