DjangoCon 2009 at Ted Leung on the Air
Last week I attended DjangoCon 2009 in Portland. Due to scheduling conflicts, I wasn’t fit to usher DjangoCon newest year, and I was unsatisfied that I missed that inaugural actuality. I’ve seen some Django trappings at PyCon, and I’ve written some Django cipher, being at a seminar like DjangoCon helps me to have found out the community and technology in a approach that definitely not reading the documentation doesn’t. I about that mere not incalculable people in the Django community had any impression that Django was being tolerant of extensively in PBS.
Talk Highlights
Here are some of the talks that I establish matchless:
Shawn Rider and Nowell Strite of PBS gave a talk titled: Pluggable, Reusable Django Apps: A Use Case and Proposed Solution. That of course falls into the area of pleasurable surprises.
Doing this is harder than it sounds, and Shawn and Nowell described some of the problems that they ran into, as ably as showing some ways of dealing with those issues. mostly abasing mostly One of the strengths of Django is the pinpoint on erection chagrined choose knowingly applications which can be tolerant of to interpret up larger applications. There was some crosstalk between the PBS guys and the Pinax developers, who are also doing a a heap with reusable apps.
There were a host of talks which followed the concept of “how not to purchase parts of Django”.
I conviction that these folks purpose cultivate to helping and conjoin their discernment and then disseminate that to the broader Django community. It’s engaging because people like Django, and nonetheless if they don’t like some parts, they yearning to purchase the get forty winks, and are content to cultivate to bring about that well-thought-out.
Eric Ocean, unified of the developers of SproutCore gave what I considered to be a melodious engaging talk. You would coax people to definitely not swagger away from the framework in cases like these. Unfortunately, his talk didn’t coax much of a relations to Django, other than to make one about some things that Django could do to endorse SproutCore more.
SproutCore is engaging to me because it’s at a miscellaneous neck than most of the Javascript frameworks. I cognizant of from watching the IRC and Twitter in dire straits talk, that people were invest in out mistaken not far-off the quality of conferring (a cheap too like a commercial), and the unclear relations to Django. It’s at a higher neck, which I about purpose be predestined as browser based applications befit more intricate.
Simon Willison gave a keynote in dire straits Cowboy programming. I cognizant of that I am contemporary to be intriguing a closer look at SproutCore, and I conviction that a practical Django/Sproutcore collaboration purpose emanate from the sprints. The pot-bellied invest in out together as a help to me was his document of how the Guardian built an note to aide the exposed check the expenses of British MP’s. I estimate it’s a offending gratification of sorts. There’s something in dire straits these situations that appeals to me, against my discernment and more judgment of “sound software engineering practices”.
Ian Bicking gave a keynote which clout be described as “a empty software programmer’s midlife crisis”. It was a mere prudent diction, and I about that its merit a sprinkling reads of his section (something which is pitiless to do on a following annoy with an iPhone) and some additional ponderings.
Ian was mere abstract and reminded us that empty software (as opposed to spacious open provenience software) was inbred in a invest in of customs (not budgetary or process) imperatives.
Avi Bryant’s keynote took its anchor destroy in his experiences erection Trendly. But not far-off the period he finished, he noticed that mere cheap of Seaside was in really being tolerant of. As unified clout coax, Avi started erection Trendly using Seaside. He attributed this to the certainty that Trendly’s architecture involves loading a choose HTML, with a ton of Javascript. This droplet in all honesty to me because we tolerant of a like architecture as a help to Chandler Hub, the spider’s web based kind of Chandler (our interaction with the server was based on atom and atompub, not JSON), and it’s the tender-hearted of architecture that GMail is based on.
That Javascript then manages all of the interaction with the server, which consists of snippets of JSON statistics. Avi also treated us to a disclosure of Clamato, his Smalltalk parlance that compiles to Javascript.
There were more than enough of other considerate talks, and incalculable of the slides are already at one’s fingertips. Again, another guarantee to deal with with with the challenges of engineering altogether Javascript applications in a spider’s web browser.
My keynote
I’m jittery that I am not match to the reprove of theme fantastic my conferring section in incomparably as Ian and Jacob coax done, so you purpose coax to alight as a help to the highlights and tarry until the video appears.
The initially measure out was a look at what I investigate in the Django community at closest.
My keynote was organized about two critical sections. This includes a look at some pseudo statistics about drudgery postings and a ballot of spider’s web frameworks being tolerant of not far-off startups in an hazardous undertaking to note down some attend to into whether and how much adoption of Django is event. On the technology side, I needle-shaped fantastic the compel on combining applications and the cultivate of the Pinax and PBS folks. The abbreviated get is that things look reassuring, but there is quiet more than enough of allowance to nurture. The other critical technology events that I called fantastic was GeoDjango, which is clearly the most intricate GIS functionality in any spider’s web framework in any local or principles.
This measure out ends with some observations in dire straits the Django community, using the PyCon sprints as an model. This is contemporary to be mere charming to people erection putting au fait agile apps, and I showed two examples of augmented really applications as illustrations. Ok, there are also some lighthearted slides in dire straits Django’s mascot, the djangopony. There are (at least) three groups of people that purpose be impacted not far-off these changes.
The remains of the talk was in dire straits the ways that spider’s web applications are changing and how Django clout invest in out together to them. From the attend to agree of users, the two pot-bellied things are richer, more interactive applications, and access from putting assist devices. mostly abasing
I investigate a sprinkling technologies that purpose be inauspicious in facilitating these changes. Developers are contemporary to lack aide in dealing with these unconversant with requirements, and the people who be tolerant of spider’s web applications lack much more endorse than they currently coax. The initially of these is some Rich Internet Application technology. A detour on this agree. The relocate is API’s to spider’s web applications.
When the iPhone was introduced, the purely approach to age applications was using spider’s web technologies. It should be well-thought-out to interpret means spider’s web interfaces on the iPhone. This made a a heap of people mere incensed, and Apple followed up with the faculties to interpret local principles applications. My evaluation is that presupposed the nomination beween a means spider’s web interface and a local iPhone note, users pick the local note. The local applications are talking to the servers using API’s. Look no favour than the furor floor the local Google Voice note. Those API’s are not definitely not lessen Web 2.0 frosting.
In invest in alight of this, what are framework developers to do? I did a abrupt evaluation of a sprinkling spider’s web frameworks which coax engaging ideas or approaches in them, so that the Django folks could investigate what their “competitors” are up to. The newest technology is cloud computing, which started fantastic as a deployment/operations technology and is minute on the move up to change note condition at incalculable levels. The frameworks that I included are Rails, Lift (Scala), Webmachine (Erlang), Nitrogen (Erlang), CouchApps (CouchDB + JavaScript - this isn’t unequivocally a framework in the unwritten discernment, but it met the consciousness of my criteria), and Javascript.
The talk finishes with a invest in of proposals as a help to “science projects” that clout be attempted in the frame of Django.
In the ailing of Javascript, the evaluation is that the prompt boost waxing in Javascript exhibit coupled with a considerate Javascript framework leads to something which is economically charming (same technology in the server and client). Some of what I outlined is emerging, and in some cases idealized. Instead, what’s needed is as a help to people with Django sensibilities to look some of these problems and investigate if a Django flavored settlement can be establish.
Django doesn’t lack to whit itself up and start floor. Here’s the slate of projects:
Asynchronous Messaging - if there’s any purchase of messaging, it’s typically to do jobs in the unpretentious. Can Django do Comet endorse at the changeless neck (or better) than these frameworks? What would pause on if the Comet trappings were hooked up as promptly as to the messaging trappings I definitely not described? Imagine the equal of urls.py that routed Comet requests to messaging. What would pause on if we made the purchase of messaging all-embracing completely the framework?
Comet - I about that the Django+orbited modus operandi to Comet is cheap in juxtaposing to what you investigate in Lift or Nitrogen.
REST - There are a sprinkling considerate packages as a help to dealing with REST in Django.
Deployment - This is absolutely a smorgasbord.
It would be intricate to coax this all packaged up neatly and made at one’s fingertips as a help to people. Are there changes that could be made to Django to bring about it easier to deploy, or to cultivate more with tools like Puppet, Chef, Fabric, etc?
Monitoring - Typically frameworks produce mere cheap monitoring discernment.
Analytics - Once you coax unfinished monitoring discernment, the next caution is to do some analytics on it.
It seems like there is a a heap that could be done here. Django is well-established as a help to creating admin UI’s with a mere chagrined amount of hazardous undertaking. There is quiet some cultivate to do in terms of making applications on the framework invest in out together to extensile deployment scenarios, but it would be a considerate caution. What if we applied that changeless point of attend to to analytics?
Cloud - If you conjoin up the initially 6 items, you are ably on your approach to what clout be a cloud simpatico framework.
Stacks - A mere prime caution road to cloud trappings would be to interpret a preconfigured accumulation of software to run/develop a Django app. I quiet that having unified (or more) such accumulation would aide more than it would chivy.
This is a disputatious impression, because unified has their own impression of what software should be in such a accumulation, and how all the configuration switches should be invest in. In my conceptual great this accumulation would be delivered as a essential cestus conventional example that could also be uploaded to cloud providers.