L'Alpiniste

The Web Column: Issue No.5

2011-06-06T07:13:35Z

Contents

Recent news
Front-End
Back-End
- The Secrets of Building Realtime Big Data Systems
The Python corner

Recent news

The latest news hot off the press.

Django | Weblog | Django 1.3 released

The latest iteration of the django web framework (one of the most popular ones) came out a couple of weeks ago. This release was delayed a few times but finally brings some much anticipated improvements: class-based views (cleaner view code FTW), better integrated logging, a standard way to handle static files and much more...

http://www.djangoproject.com/weblog/2011/mar/23/13/

Divide and Concur « Code as Craft

Interesting article on Etsy's use of Jenkins to split up their test suite to get it run inside of 11 minutes pre-deployments.

http://codeascraft.etsy.com/2011/04/20/divide-and-concur/

The road to being a kick-ass public speaker | Christian Heilmann's blog – Wait till I come!

Christian Heilmann provides some great insights and tips on what it takes to make a good presentation.

http://www.wait-till-i.com/2011/04/11/the-road-to-being-a-kick-ass-public-speaker/

byobu in Launchpad

Byobu is a fantastic set of enhancements to gnu screen. It provides a great deal of information and is on my list of base packages for any new machine.

https://launchpad.net/byobu

Puppet Labs - Style Guide

Puppet labs produce a style guide detailing exactly how your should layout your modules and manifests. Essential reading if you're using puppet and want to adhere to a common standard.

http://docs.puppetlabs.com/guides/style_guide.html

Fabric Gets a nice set of updates

Fabric, if you've not already heard of it is a python library that makes it really easy to handle deployments with a fabfile, which is essentially similar to a Makefile except that it it's just pure python.

Tav has built-out a bunch of additional features for Fabric: Running commands in parallel, API improvements, Fabric Shell. Staged environments. It's all there, the only downside is that it's not currently in a format that makes it easy for upstream to accept the patches. Still hopefully the work will be done to resolve that and we'll see these changes land in Fabric in due course. See Tav's post for more details (via Ross Lawley)

Iranian Hackers obtain SSL certs for major sites fraudulently

An fairly worrying report from eff.org on how Iranian hackers managed to trick a Certificate Authority into issuing fraudulent SSL certs for major sites such as google.com, login.yahoo.com.

Matt Haughey on 11 Years of Community at MetaFilter

You may have seen this already but if not it's well worth seeing. Matt Haughey talks about the tools created to manage and moderate the community on MetaFilter.

Front-End

Everything related to client-side development.

heatmap.js | HTML5 Canvas Heatmap Library

Heatmaps are a fantastic way to visualise geographical statistics and of course user behaviour on web pages. Heatmap.js is the easiest way I've seen to build such visualisations.

http://www.patrick-wied.at/static/heatmapjs/

jQuery: » jQuery 1.6 Released

jQuery 1.6 is out, with small but major changes which break backwards-compatibility (see .prop() and .attr()).

This is not really a problem though, as jQuery has mostly been bug-free for a while, so running on 1.5 until you've updated your codebase, however big it is, is perfectly reasonable.

http://blog.jquery.com/2011/05/03/jquery-16-released/

Web font services - An Overview - by sprungmarker.de

Sylvia Egger keeps a useful comparison table of the various webfonts hosting services out there, allowing you to compare how many fonts are available, how you can embed them, what you can use them for and which particular services are offered.

http://sprungmarker.de/wp-content/uploads/webfont-services/

Anatomy of a Mashup: Definitive Daft Punk visualised

Cameron Adams awesome visualisation of his own audio mashup of 23 different Daft Punk tracks using modern web technologies.

http://daftpunk.themaninblue.com/

Improve Browser Performance With the CSS Stress Test Tool » SitePoint

Bookmarklet that measures performance impact of CSS effects.

http://blogs.sitepoint.com/css-stress-test-tool/

An unconventional loading strategy for YUI 3 | Julien Lecomte's Blog

An interesting deck from Julian Lecomte on YUI3 loading strategies used on SRPs.

http://www.julienlecomte.net/blog/2011/03/583/

Back-End

Server-side technologies and services.

The Secrets of Building Realtime Big Data Systems

Nathan Marz's presentation on building scalable architecture. This is a very good introduction to the type of architecture needed for any "realtime" analytics platform. Nathan has also written a few tools/libraries which facilitate data processing on Hadoop (for example Cascalog: https://github.com/nathanmarz/cascalog).

http://www.slideshare.net/nathanmarz/the-secrets-of-building-realtime-big-data-systems

The Python corner

Python rocks. At Project Fondue, we love it and so we keep a close eye on what's happening in the Python realms. Here is an excerpt.

PyPy Status Blog: PyPy 1.5 Released: Catching Up

A new release of Pypy is out and now claims to be on par feature-wise with Python 2.7.1 "including the standard library".

Pypy is a major achievement and keeps getting better and better. Don't gorget to check their compatibility list before you switch to it over cPython though: https://bitbucket.org/pypy/compatibility/wiki/Home

http://morepypy.blogspot.com/2011/04/pypy-15-released-catching-up.html

pyke

Pyke is a knowledge-based inference engine (expert system) written in 100% Python.

Pyke Homepage
Blog post with a practical application
- In french
- English Translation via translate.google.com

PyCon

There's a huge amount of fantastic videos of the talks from PyCon 2011. Well worth checking out.

PyCon 2011 Videos

The Web Column: Issue No.4

2011-02-09T09:23:16Z

Contents

Recent news
Front-End
Back-End
- Elasticsearch
The Python corner
- datadiff
- Grin and Search it

Recent news

The latest news hot off the press.

Google Accuses Bing of Copying their Search Results

A fascinating article by Danny Sullivan that broke the news on Google's accusation that Bing are copying their search results. Google Accuses Bing of Copying their Search Results

Why You Should Never Search For Free WordPress Themes...

Want a free wordpress theme? You might be getting more than you bargained for. This article looks at various sites that provide free wordpress themes only to uncover a whole heap of obfuscated spam.. Talk about there being no such thing as a free lunch.

History Hack Day

YDN's Murray Rowan reports on the events at the recent History Hack Day. Sounds like it was a great success with many excellent hacks being produced over the course of the event.

"Hack Days get people engaged with and excited about a subject, and our subject is history. There's a lot of it about, but lots of it is stuck, on paper, in old HTML pages, in archives, on hard-to-reach databases. We want people to figure out how to set some of that free, and start bringing that history to life as... whatever crazy things people can come up with."

—Matt Patterson, Organiser of the History Hackday

Spread the Word on Facebook Scams and Phishing Techniques

Graham Cluley's Naked Security blog is a great resource for keeping abreast of the latest news relating to computer/web security. Recently a lot of the news has been related to scams on facebook.

Now for those of us who work in the industry it might be more easy for us to spot a scam from a mile away. However if you have friends and family that frequent facebook and similar sites it could be worth getting them to take a look at Graham's blog to understand how theses ploys work and how they generate scammers money.

Front-End

Everything related to client-side development.

Deferreds in JQuery 1.5

Using deferreds in Jquery provides a nice walk through of how to use the new feature in JQuery 1.5.

"With deferreds, multiple callbacks can be bound to a task’s outcome, and any of these callbacks can be bound even after the task is complete. The task in question may be asynchronous, but not necessarily."

—Eric Hynds

If you've ever used Twisted in Python this territory will be familiar to you.

HTML5 Forms

Chris Mills and Patrick H. Lauke provide walk through some of the shiny new HTML 5 features available for creating richer forms on Dev.Opera: New form features in HTML5

Breaking the Web with HashBangs

Mike Davies writes up an excellent critique on the trend of using fragment identifiers to provide state for Ajax based site navigation schemes. Breaking the Web with Hashbangs

Little big details

Quite often the work going into improving the user experience goes unnoticed even though it can turn on ordinary application or web site into something users actually enjoy. You know, that whole "delight your users" thing. Little big details is a blog pointing out those improvements and sometimes genuinely innovative ideas that make the software we like using.

Back-End

Server-side technologies and services.

Elasticsearch

As explained on the homepage, search is hard. That's exactly the problem elasticsearch is trying to tackle, and by the looks of things it might be successful at making search a lot more trivial than it currently. Many features one might expect from a robust search solution have been implemented in elasticsearch and quite often requiring little work or no work at all for the user: redundancy, failover, multiple indices, easy data format, simple configuration, automatically distributed architecture, choice of index storage etc.

We haven't tested it ourselves (yet), but this looks like an impressive project with a lot of potential. If you're looking at implementing search: give elasticsearch a go.

The Python corner

Python rocks. At Project Fondue, we love it and so we keep a close eye on what's happening in the Python realms. Here is an excerpt.

datadiff

This is an interesting idea. Datadiff provides a way to get a human readable diff from python data structures

>>> from datadiff import diff
>>> a = dict(foo=1, bar=2, baz=[0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
>>> b = dict(foo=1, bar=4, baz=[0, 1, 2, 3, 4, 5, 6, 7, 8])
>>> print diff(a, b)
--- a
+++ b
{
-'bar': 2,
+'bar': 4,
 'baz': [
 @@ -5,11 +5,8 @@
  6,
  7,
  8,
 -9,
 ],
 'foo': 1,
}

Grin and Search it

Brandon Craig Rhodes uses grin instead of grep. It's fairly similar to ack-grep but is written in python.

grin provides a way to search through files, colorising the matches in the output so you can see exactly what you were looking for and quickly.

Contribute to the Web Column

2011-02-08T10:27:43Z

We're always on the look-out for interesting information to add to our Web Column. If you've come across a useful blog-post or have discovered a great library that you couldn't live without or have a great tip to share then we'd love to hear about it.

Please note all content needs to be strictly web-development/tech/geek related. We welcome submissions from original content authors too, if it's well-written, informative and interesting then chances are we'll include it.

To send us a link, please drop us a line via our contact form. We're always interested in feedback too, so if there's something we should be covering more of please let us know.

As always, to keep up-to-date with all Project Fondue happenings follow us on twitter here.

The Web Column: Issue No.3

2011-01-18T10:10:11Z

Contents

Recent news

Google Drops Native support for H.264 from Chrome

There's been a strong reaction to the news that Google is dropping native support for the H.264 codec within Google Chrome in order to push forward their own open-source WebM codec. There's plenty of polarised opinons, and that's not suprising with Apple and Microsoft in one corner and Opera, Mozilla and Google in the other (see this chart from Bruce Lawson).

daringfireball.net - Google Dropping Support for H.264 in Chrome
wired.com - Google Reveals Plan To Remove H.264 Support From Chrome

Puppet Cookbook Launched

Former colleague Dean Wilson has been hard at work over the break putting together the Puppet Cookbook - It's full of sound tips and definitely a resource worth keeping tabs on if you're using Puppet for configuration management.

Why developers should care about system packages

In Why developers should care about system packages Gareth Rushgrove details why system packages are important and why packaging skills are important for developers to learn. He also mentions Launchpad's excellent PPA feature, which is something we use quite a bit here at Project Fondue for backporting newer package versions. Back-porting a newer version of a package is something that PPA's are really handy for. Particularly as with a lot of packages this is often a relatively simple task.

How Facebook Ships Code

Whether you like the site's purpose or not, there is no denying that Facebook is a high traffic, technically challenging site. These rumors on how Facebook handles development and deployment are of course to take with a grain of salt, but can provide a little insight into how different a company Facebook might be, should you care.

Bazaar Pipeline Plugin Demo

Bazaar Pipeline Plugin Demo provides a run-through of what can be achieved with the bzr-pipeline plugin. Which looks potentially very useful for breaking up large projects in to smaller bite-sized, reviewable chunks.

Google App Engine's new High Replication Datastore

Google has announced that it will now support a second datastore for App Engine as a remedy to availability problems encountered with the regular Master/Slave datastore. There are a few caveats though:

an application has to take particular precautions when using this new High Replication datastore, particularly regarding write latency, which can be much higher than before. Therefore data consistency is not guaranteed as much as it was with the Master/Slave datastore.
the new duplicates data across many data centers, which means more bandwidth used by Google, which in turn means the pricing of this new datastore is significantly higher than its little brother's. Google is starting with a three times more expensive scheme and will re-evaluate the price in June.

Despite its drawbacks, the new datastore seems like a pretty robust offer. If you value Google App Engine as your hosting platform of choice and should avoid downtimes at all cost, this might just be the solution you were waiting for.

Front-End

NoVNC

NoVNC - An exciting project which provides VNC with HTML5, websockets and canvas. via @sil

Modernizr 2.0 Beta Preview

Faruk Ate? and Paul Irish have joined forces with Alex Sexton to put out a 2.0 version of Modernizr, their feature detection Javascript library, which "allows you to target specific browser functionality in your stylesheet". This is a very handy tool when using a progressive enhancement approach, which you already do, of course...

Game development and Javascript

A very long list of game development resources in Javascript was compiled on Reddit. You can find there a plethora of graphics and audio engines to play with. With WebGL being implemented across most modern web browsers at the moment, game development will soon become a new avenue for your leet JS skillz :)

Framework activity

Symphony 2

Work on Symphony 2 is continuing with recent refactoring (and more) and a PR5 release which Fabien Potencier said is more or less the equivalent to a beta release. A release candidate is planned soon.

DjangoCon Europe 2011

DjangoCon Europe 2011 was recently announced and will take place in June in Amsterdam, in the old port district (a pretty cool and different area). All the relevant info can be found at the official site.

The Python corner

Python rocks. At Project Fondue, we love it and so we keep a close eye on what's happening in the Python realms. Here is an excerpt.

Ned Batchelder: Faked translations: poxx.py

poxx.py: A Nice utility for testing out gettext translations.

Pyro

Pyro stands for Python Remote Object library and does exactly that: let you call Python code running on other machines. The API is very simple and a new version, Pyro 4, is being written from scratch at the moment. RPC can be a tedious task, so there is no doubt this could come as a handy tool to quite a few people.

Focus on: WTForms

WTForms, by Thomas Johansson and James Crasta, is a forms library which is in appearance very close to Django's forms library, but is framework-agnostic.

WTForms is a great choice if you are using a lighter-weight framework with a standalone ORM. It integrates with result objects in a very elegant way.

Here's an example of a form taken straight from the docs. This is fairly similar to Django forms on the whole. As the forms are just python classes you can subclass forms quickly and easily to add an extra field if you need to.

from wtforms import Form, BooleanField, TextField, validators

class ProfileForm(Form):
    """User profile."""
    birthday  = DateTimeField('Your Birthday', format='%m/%d/%y')
    signature = TextAreaField('Forum Signature')

Getting Data into the form

The really nice feature is how you can easily use both formdata and an object to update your form.

def edit_profile(request):
    user = request.current_user
    form = ProfileForm(request.POST, user)
    if request.method == 'POST' and form.validate():
        form.populate_obj(user)
        user.save()
        redirect('edit_profile')
    return render_response('edit_profile.html', form=form)

There's two things to note here. See the form instance is created by passing in the POST and the user object. Forms get data using the following precdences:

Form Data - form data has the highest priority
Data Object - the data object is the next in line. Usually an object returned by your ORM - e.g. Storm, SQLAlchemy
Kwargs - kwargs take the lowest priority.

So anything that's not present in the form data (e.g POST) is filled in by the data in the user object in our example above.

This means it's possible to make create/read/update and delete forms (CRUD) in an arbitrary way.

Updating DB objects with validated data

To update the db object form.populate_obj(obj) is used. This simply updates obj with the validated data based on matching it up to the object attrs. It's very neat and easy to use. With a bit of thought it's possible to use your favorite simple framework + your preferred ORM + WTForms and create a nice set of re-usable CRUD forms.

The Web Column: Issue No.2

2011-01-14T09:15:34Z

Contents

Recent news
Design
- Retinart on page harmony
Front-End
- CSS3 page curls
Framework activity
- Django 1.3 getting there ...
- Smisk: A Web-service framework written in C and controlled by Python
The Python corner
- Configglue
- Why PHP Is Fun and Easy But Python Is Marriage Material
Focus on: Python YQL
End of Year!

Recent news

24 ways 2010 edition

24 ways is carrying on with great articles about workflow, organisation, APIs and shiny Javascript and CSS techniques.

This year, 24 Ways is also printing all articles, selling it for £8 and donating all sales profits to UNICEF. Very cool.

This year's 24 Ways is at http://24ways.org/2010/
Order an annual - hurry though; you can only order a copy during December!

Spacelog

James Aylett and (ex-colleague) Mark Norman Francis regularly organise /dev/fort, a week-long gathering of web geeks in an actual fort (yes!). Designers and developers alike collaborate without all the usual interference you can find in a "normal" office. Oh, and the Internet is a 30 minute walk away.

This year they launched Spacelog, a fantastic presentation of space mission transcripts from Nasa. Two missions are being featured so far: Apollo 13 and Mercury 6.

The UX, design and frontend work on this site are brilliant. It also offers very cool features like linkable transcript parts, for example http://apollo13.spacelog.org/05:03:12:07/05:03:12:35/#log-line-443527

Stay up all night reading Spacelog at http://spacelog.org
Don't miss Hannah Donovan's article on 24 Ways detailing the process used during this year's /dev/fort.

Chrome Cr-48

Google introduced a Pilot program for Chrome OS notebooks for US-based residents. Open until December 21st (Today!). It has been getting pretty good reviews all around. If they open up a program in Australia and the UK we'd be up for taking it for a spin :)

WebSockets availability delayed

You may have heard that Mozilla and Opera have announced that WebSockets won't be enabled by default in their future versions, due to a security issue with protocol. This is of course not very good news since WebSockets are pretty cool to play with and open up the door to a standardised way of implementing many new applications.

But do not fear, for this is not the end of WebSockets at all. Pusher, which develops a realtime communications platform based on WebSockets has all the details for you:

http://blog.pusherapp.com/2010/12/9/it-s-not-websockets-it-s-your-broken-proxy

Prey: Open Source, Cross Platform Theft Recovery software

Sadly Christmas time is well known to bring an increase in thefts of personal property. Prey is a nice bit of software which provides the means to activate tracking of your laptop or phone after it's been stolen (assuming you've installed it beforehand of course!).

It can do all-sorts of useful things such as reporting it's location based on wifi triangulation or gps if available as well as taking a photo of the thief if the device is equipped with a camera. Best yet Prey is cross platforma and is currently available for mac, windows, linux and android.

http://preyproject.com/

Design

Retinart on page harmony

Retinart, a beautiful design blog, has a great post of the Secret Law of Page Harmony, a pretty old technique to lay out elements on a page. They even give you the backstory behind the technique.

http://retinart.net/graphic-design/secret-law-of-page-harmony/

Front-End

CSS3 page curls

Page curls are quite shiny and Matt Hamm has come up with a nice example of how to make them using only CSS3.

http://matthamm.com/box-shadow-curl.html

Framework activity

Frameworks are built for The Greater Good ™. No matter which language or framework is your favourite, it's always good to be aware of what "the other guys" are doing. So here is what we picked up from the world of frameworks recently:

Django 1.3 getting there ...

... bit by bit. A few deadlines have been missed :( but things aren't that bad, and Django 1.3 is now scheduled for release at the end of January.

Django 1.3 will feature (amongst many other things) class-based views, improvements to logging and unit testing, and will deprecated mod_python support.

the latest release schedule blog post: http://www.djangoproject.com/weblog/2010/dec/13/django-1_3-release-schedule-update/

Smisk: A Web-service framework written in C and controlled by Python

Smisk is a framework for building scalable web-services using python. It's fully WSGI compliant and is currently being used in production by Spotify no-less.

The Python corner

Python rocks. At Project Fondue, we love it and so we keep a close eye on what's happening in the Python realms. Here is an excerpt.

Configglue

Configglue is a nice library that makes it really easy to combine both configuration files and command-line arguments. Fantastic for rolling utility scripts where the the amount of options can make the CLI-args seem a little unwieldy. There's also a separate project for using configglue to replace django settings.

Why PHP Is Fun and Easy But Python Is Marriage Material

An interesting article which takes a look at programming language choices and more specifically deals with the pros and cons of PHP and Python. It makes some good points and is well worth a read if you're a PHP programmer thinking about getting into Python.

http://onstartups.com/tabid/3339/bid/20493/Why-PHP-Is-Fun-and-Easy-But-Python-Is-Marriage-Material.aspx

Focus on: Python YQL

Python-yql has recently become the newest Open Source project maintained by the Project Fondue Team so we'd thought it would be a good opportunity to highlight it.

If you've not heard of YQL (Yahoo Query Language) (you've probably been living under a rock) it's a SQL-like way to query the internetz for data from a growing number of public APIs. The nice thing about YQL is that it takes a lot of the effort out of interfacing with other APIs.

Python-yql is a Python wrapper to YQL which looks to make using YQL from Python as easy as possible. Here's a quick example of what you can do:

>>> from yql import Public
>>> y = Public()
>>> res = y.execute("""USE 'http://www.datatables.org/
ukpostcode/ukpostcode.postcode.xml';
                    select * from ukpostcode.postcode
                        where postcode='SW1A1AA'
                            and format = 'xml'""")
>>> res.one()['geo']['lat']
u'51.501007'
>>> res.one()['geo']['lng']
u'-0.141588'
>>>

Python-yql documentation: http://python-yql.org
Python-yql Launchpad Project: https://launchpad.net/python-yql

Python-yql has been around for about a year, and there's been quite a few updates in that time. However we're always happy to hear from anyone with suggestions or feature requests. If you have found a bug or something doesn't work how you'd expect it to please file a bug on launchpad - https://launchpad.net/python-yql. Additionally patches are welcomed.

End of Year!

Ok, that's it! Join us again in the new year for the next edition of the web column. We're going to retreat to our Mountain hideaway over the Christmas holidays to eat fondue, drink and be merry. Here's wishing you all very happy Christmas and a happy new year!

The Web Column: Issue No.1

2011-01-14T09:13:37Z

Contents

Recent news

24 ways 2010 edition

For the 6th year in a row, Drew McLellan and friends are back to make December the best month of the year.

The 2010 edition is already packed with progressive enhancement, web fonts, Javascript, CSS, design goodness, and generally good advice, as usual.

Drop whatever you're doing and go read it at http://24ways.org/2010/. Now!

CSS3 Flexible Box Layout Module

In the same spirit, Robert Nyman has a great article and examples about the CSS3 Flexible Box Layout Module.

This is something many frontenders have been wishing for, and now it's just around the corner (the IE and Opera teams still need to work on it).

Robert's article: http://robertnyman.com/2010/12/02/css3-flexible-box-layout-module-aka-flex-box-introduction-and-demostest-cases/

JSMentors mailing list

Rey Bango and Asen Bozhilov decided to put together a list dedicated to all things Javascript, including reviewing code and getting by some of the best JS coders out there. If you don't feel up to scratch on JS, this might be one of the best places to find good advice.

The mailing list's site: http://jsmentors.com
The announcement at Ajaxian: http://ajaxian.com/archives/jsmentors-com-the-mailing-list-to-learn-javascript

New Web Messaging working draft

A new working draft has been published by Ian Hickson regarding communication between different browsing contexts (ie. documents) "regardless of their source domain".

Of course security measures are part of the spec, which still states "Warning! Use of this API requires extra care to protect users from hostile entities abusing a site for their own purposes."

Don't let that scare you too much and keep on eye on this spec though, as communication between different web pages would be a big deal if vendors start to implement it.

The draft is at http://www.w3.org/TR/2010/WD-webmessaging-20101118/

CSS railroad diagrams

Stoyan Stefanov has decided to "enjoy his unused vacation days" writing a CSS parser in Javascript. Yep.

After writing a lexer, he's now taking on the parsing task and details his work in a set of railroad diagrams which will give you a little bit of insight into what work browsers have to do upon page load.

The lexer code: https://github.com/stoyan/etc/tree/master/cssex
Post about parsing, with railroad diagrams: http://www.phpied.com/css-railroad-diagrams/

MobileSafari and HTML5

iOS 4.2 came out recently, with a new version of Mobile Safari and new features: accelerometer, Web sockets, HTML5 Forms, partial XHR2 support and better JS, DOM, SVG and Canvas support.

You can find a good list of the changes at http://www.mobilexweb.com/blog/safari-ios-accelerometer-websockets-html5

Apparently XHTML 2 is still around

Oh well...

http://www.w3.org/News/2010.html#entry-8962

Smashing Magazine roundup of project questionnaires

If you do freelance work, or even work for an agency, understanding what your clients actually want isn't always that simple.

I always found Clearleft's Client Ideas Sheet was great, but Smashing Magazine went around and did a round up of many questionnaires of this kind. From all these you can probably compile your own that fits your type and style of work. Very useful stuff.

http://www.smashingmagazine.com/2010/11/26/web-design-questionnaires-project-sheets-and-work-sheets/

Framework activity

OpenSocial 1.1 published

The OpenSocial spec is carrying on with version 1.1 published in November, with emphasis on pub/sub communication between gadgets.

http://blog.opensocial.org/2010/11/opensocial-11-published.html

Backbone.js and Django

Josh Bohde has a nice example of how to integrate Backbone.js with Django.

http://joshbohde.com/2010/11/25/backbonejs-and-django/

DjangoCon US 2011 announced

From the blog post: "DjangoCon US 2011 will be held in Portland, Oregon, from September 6-8 2011. This will be followed by a couple of days of sprints."

A couple of things to note: the venue has changed, and the organizers are calling for help via their mailing list.

Announcement: http://www.djangoproject.com/weblog/2010/nov/22/djangocon-us-2011/
Organizers' mailing list: http://groups.google.com/group/djangocon-organizers

Rails for Zombies

In order to make Rails fun to learn, the guys at Envylab came up with a pretty cool educational concept: "You'll watch five videos, each followed by exercises where you'll be programming Rails in your browser."

Oh, and the whole thing has a Zombie theme.

The site: http://railsforzombies.org

Symphony 2 is making good progress

A whole lot of work was checked in recently and apparently it could hit beta before the end of the year.

Summary of recent activity: http://www.symfony-project.org/blog/2010/11/28/a-week-of-symfony-204-22-28-november-2010

The Python corner

Python rocks. At Project Fondue, we love it and so we keep a close eye on what's happening in the Python realms. Here is an excerpt.

Textmate and Pyflakes

David Crammer wanted inline-error checking in Textmate and made a bundle for it, based on an existing one for Javascript.

Blog post: http://www.davidcramer.net/code/36327/integrating-pyflakes-into-textmate.html
Javascript Tools Bundle: https://github.com/johnmuhl/javascript-tools-tmbundle
David's Python Tools Bundle: https://github.com/dcramer/python-tools-tmbundle

Yahoo! Pipes, Python and Appengine

Greg Gaughan, inspired by others, has decided to compile Yahoo! Pipes into Python code and run them on AppEngine.

Blog post: http://www.wordloosed.com/running-yahoo-pipes-on-google-app-engine
The code on Github: https://github.com/ggaughan/pipe2py
The test site, on AppEngine: http://pipes-engine.appspot.com

Twisted 10.2.0 out

From the announcement: "Survivors of the release process - what few there were of them - have been heard to claim that this version is "awesome", "even more robust", "fun-sized" and "oven fresh"."

There's no reason not to use it then.

Announcement: http://labs.twistedmatrix.com/2010/11/twisted-1020-released.html
Twisted's site: http://twistedmatrix.com

Python redis graph library

Amir Salihefendic has implemented a small graph library on top of Redis.

Blog post: http://amix.dk/blog/post/19592#redis-graph-Graph-database-for-Python
The Package on Pypi: http://pypi.python.org/pypi/redis_graph/1.0

PySignals

Django implements signals as a nice way to achieve decoupling through your site.

Theo Julienne extracted that signal dispatcher into its own package and put it on Pypi. Good idea!

The Package on Pypi: http://pypi.python.org/pypi/pysignals/0.1.1

Mongrel and WSGI apps

A new WSGI handler for Mongrel was published by Daniel Holth. This adds to the existing "Mongrel2-WSGI-Handler" by Berry.

If you wanted to play with Mongrel2 and Python, you're all sorted now.

Scrapemark

Scrapemark is a nice little module that lets you easily scrape web pages.

The site: http://arshaw.com/scrapemark/

Focus on: Object Oriented CSS Framework

In most issues, we will spend a bit more time detailing a particular, not necessarily new, project which we deem particularly useful and might not be known from all.

This time: Nicole Sullivan's Object Oriented CSS Framework (OOCSS for short).

Through 2008 Nicole developed the idea of a simple, highly reusable and maintainable way of writing CSS code. At the beginning of 2009 she pushed the first public commit of the OOCSS project.

OOCSS emphasises the importance of separation between the structure and the styling (or "skin"). It also ensures that most "objects" can be embedded in most containers. You've certainly wanted to move a piece of functionality from one location to the other on a page before. Writing your CSS with OOCSS makes this process a lot simpler than it usually is.

Oh and of course, reusability for you as a developer, means reusability by the browser too. Less parsing is always good for performance.

OOCSS is not a massive framework of sorts. It is, essentially, a way of working that comes with examples. It advocates a change in mindset: what you usually see as a set of technical properties and attributes, coded in a linear way, element after element, you should try to see as functional, logical blocks and then organise your code as so.

Looking at the big picture is rarely a waste of time, and what you will gain is clarity both in your head and in your code.

You can certainly apply principles of OOCSS through your work without necessarily using the provided templates. But toying around with the code on Github doesn't hurt and will at least show you the way all this can be achieved.

Give it a go, if you don't adopt it, you'll probably learn from it.

Public site: http://oocss.org (apparently put together for the Velocity conference)
The code on Github: https://github.com/stubbornella/oocss
The wiki on Github: https://github.com/stubbornella/oocss/wiki
Nicole's post introducing OOCSS: http://www.stubbornella.org/content/2009/02/28/object-oriented-css-grids-on-github/

Security: Raw Roles in Docutils

2010-07-27T22:23:09Z

The flaw occurs via the roles feature which is designed to allow the author to bypass the parsing of rst. The only problem was that despite the options being set to disable the raw directive which similarly allows unfiltered text, the raw role was still permitted.

Here's an example of how this could be used:

.. role:: unsafe_raw(raw)
    :format: html

:unsafe_raw:`<p onclick="alert('hello')">Oh Hai (click me)</p>`

The good news is following the provision of a patch this was fixed in release 0.6 of docutils. If you're using reStructuredText on your site where third parties can use it to provide content, it's definitely recommended that you check to make sure you're using 0.6 or higher.

Using Apache in a Python Development Environment

2010-07-12T22:21:07Z

figure 1: Apache front-end and two web services.

Initially our development environment used the built-in WSGI server that comes with Werkzeug, but it occurred to me that ultimately this would be running on mod_wsgi under apache, so we decided to set-up the development environment to use apache instances with mod_wsgi running in daemon mode.

This actually proved to be fairly straight forward, all that's required is to generate a configuration for each instance at run-time, which points at each application.

By abstracting the start-up of each instance a single script can fire up the web-services, and additionally the front-end one after the other. The other advantage with this approach is that you can then develop any configurations that make use of apache modules with ease, so for example using something like mod_rewrite is no problem. In essence this means you don't have to write utility code that's just used to replicate something that your server provides.

Reloading on changes. Development servers such as Werkzeug's come with code to trigger a reload when code changes so that you can see changes instantly during development. Fortunately with mod_wsgi this isn't a problem as the code is provided for you see the mod_wsgi documentation.

Configuring this as part of the environment is simple and works very well.

All of this means you can have a development evnironment that brings you closer to a production environment with minimal hassle. So far this is working very well for us.

Building a Static Asset Cluster

2010-03-04T23:22:12Z

Initial Thoughts

As this configuration is going to be used by several sites it became clear that building some kind of redundancy into the setup would be a good idea. Otherwise if this ended up as a single point of failure, all our sites would have no images, JavaScript or CSS should something go wrong with the static server.

So the first piece in the puzzle was being able to share a single public ip address across 2 Slicehost boxes to provide high availability. We're using the excellent Slicehost as our provider and they're happy to support HA configurations. For a shared IP address all it takes is a mail to Slicehost support to say that's what you want. At the same time I added another public IP so that each box has a dedicated local network and public IP address as well as sharing a single public ip between them.

figure 1: Basic diagram showing simple static cluster across 2 boxes.

Setting up Heartbeat

Setting up heartbeat [1]. is an extremely straightforward process it's simply a case of installing the packages and then providing configs for each box as well as a centralised config which is the same on both boxes.

The way it works is that one box has the shared IP enabled on an alias of the primary network card eth0 in our case. If the non-live server detects the live box has gone down it switches on its alias for the shared IP address. This provides a quick and easy way to have failover. Slicehost provide detailed instructions on using Heartbeat with slices on their site [2].

Setting up HAProxy

The next piece is HAProxy [3] which we use as a load balancer. This handles all the traffic on port 80 and then balances it across our Nginx boxes which are serving from port 8080. HAProxy requires some fairly simple configuration where you define a farm and define the nodes you want to balance traffic over. If you are balancing traffic for an application it's possible to have HAProxy set cookies so that once a client has made one request all subsequent requests will go to the same box. In our case we are dealing with static content so it's not necessary.

server  Static1 xxx.xxx.xxx.xxx:8080 check port 8081 inter 2000
server  Static2 xxx.xxx.xxx.xxx:8080 check port 8081 inter 2000

So that HAProxy knows which of our Nginx boxes are up, it uses a HEAD request to each machine to check it's alive. This is carried out on a separate port e.g: 8081 so that it's independently controlled. Done this way we can switch off the Nginx config (via a script nginxdissite) in order to temporarily remove it from the farm. An alternative to this would be to turn off access to port 8081 through iptables. Either way; this makes it dead easy to be able to carry out maintenance without removing all static serving capabilities. Adding it back in is just a case of re-enabling the site and then HAProxy can start distributing traffic to that node again.

HAProxy also comes with a stats page which you can view as a way of being able to see what nodes are up and how much traffic is being served at any one time.

Configuring Nginx

The Nginx [4] configuration is pretty straightforward in that it's optimised for static asset serving and defines GZIP and Expires along with a couple of custom rewrite rules. The first is to allow Nginx to discard version numbers from requests. This is a cache-busting mechanism for when assets are set with long expiry date of a year in the future. The only way to get the client to download a new asset (when for example a graphic has been altered or CSS updated) is to provide a new URI. This is done by adding the revision to the asset file name. So rather than linking to /i/foo.png we can link to /i/foo.2348.png.

rewrite "^(.*)\.\d+\.(jpeg|jpg|gif|png|css|js|html)$"  $1.$2;

The second rewrite addition is to redirect subdomains to directories. This provides a nice easy way to be able to utilise new subdomains in the future depending without needing to update the configuration. This of course requires a wildcard A record to be set up in DNS as well as a wildcard in the server_name directive.

if ($host ~ '(.*)\.staticfil.es' ) {
    set $subdomain $1;
    rewrite  "^/(.*)$"  /$subdomain/$1;
}

Lastly I wrote my own nginxensite/nginxdissite for enabling and disabling configurations. This is based on the same Debian/Ubuntu convention used with Apache. The Apache convention is the following: you have sites-enabled and sites-disabled directories and the sites-enabled directory is set-up so that any configuration within it is automatically included. The a2ensite command symlinks a chosen configuration from sites-available into sites-enabled. The benefit of it is that it encourages good separation of a conf for each site and provides a quick way to drop a particular configuration or replace it with a new one.

Mirroring Static Assets with Glusterfs

The final issue with setting something like this up is how to cleanly share assets between servers. A basic solution would be to use rsync on cron which is fine, except that there would be a delay between getting files on one box before it is synced to the other one. To eliminate this delay I considered writing a script based on inotify which would run rsync as files were added but this would also create lots of problems to solve, such as ensuring adding lots of files doesn't result in too many rsync processes, but at the same time doing this without losing the almost realtime aspect of delivery which is the desirable result.

I then started looking into what was available from distributed filesystems. In the end after some testing I settled on using Glusterfs [5]. Glusterfs has a server and a client component and is very easy to set-up. In our case I made the two boxes run the server configuration which essentially makes each node a mirror of the data. Then each server also runs the client which is essentially a mount of the filesystem via FUSE [6]. Adding files through the client makes those files almost instantaneously available from both static servers which is perfect for what we are trying to do. The other cool thing with Gluster is that scaling this up through adding more boxes becomes very easy to do.

Glusterfs is also fault tolerant in that if we drop one of the servers out of the farm for maintenance and continue to add files to be served as soon as that box is back in rotation it will pick up all of the missing files.

So that's it's - feel free to add your thoughts, comments, experiences of this kind of set-up or similar in the comments.

Getting an Early Warning of Low Disk Space

2010-01-21T23:51:57Z

Keeping an eye on Disk Space

Running out of disk space is not a good situation to be in. Services that run out of disk space can cause your server to blow up in a very short space of time. As a result of that it's a good idea to get an early warning that your machine is on it's way to running out of disk space. This script is a noddy way to get that insight with minimum effort.

# -*- coding: utf-8 -*-

"""
Disk Space
Simple class for working out the free disk space on a system

by Stuart Colville http://muffinresearch.co.uk
License: http://www.opensource.org/licenses/mit-license.php

"""

import os
import math

class DiskSpace():

    """Free Disk Space"""

    def __init__(self, path="/"):
        """Init class and retrieves disk space info"""
        self.disk = os.statvfs(path)

    def has_free_space(self, limit):
        """Bool returns true if remaining space is above limit %"""
        if float(self.percent_free()) < float(limit):
            return False
        else:
            return True

    def percent_free(self):
        """Gets the amount of space left as a percentage"""
        return (math.ceil(float(100) / float(self.bytes_capacity())
                                                * self.bytes_free()))

    def bytes_capacity(self):
        """Returns the total capacity in bytes"""
        return self.disk.f_frsize * self.disk.f_blocks

    def bytes_free(self):
        """Returns the free space in bytes"""
        return self.disk.f_frsize * self.disk.f_bavail

    def bytes_used(self):
        """Returns the used space in bytes"""
        return self.disk.f_frsize * (self.disk.f_blocks -
                                                   self.disk.f_bavail)

    @staticmethod
    def humanize_bytes(bytes, kilo=1024):

        """Humanizes bytes

        See http://en.wikipedia.org/wiki/Kilobyte for info on the
        different ways to interpret whether a kilobyte is 1,024 bytes
        or 1,000 bytes

        """

        if kilo == 1024:
            label = ["KiB", "MiB" , "GiB", "TiB"]
        else:
            # fallback
            kilo = 1000
            label = ["KB", "MB" , "GB", "TB"]

        if bytes < kilo:
            return "%d" % bytes
        elif bytes > kilo and bytes < math.pow(kilo, 2):
            return "%.2F%s" % (float(bytes/kilo), label[0])
        elif bytes > math.pow(kilo, 2) and bytes < math.pow(kilo, 3):
            return "%.2F%s" % (float(bytes/math.pow(kilo, 2)), label[1])
        elif bytes > math.pow(kilo, 3) and bytes < math.pow(kilo, 4):
            return "%.2F%s" % (float(bytes/math.pow(kilo, 3)), label[2])
        elif bytes > math.pow(kilo, 4) and bytes < math.pow(kilo, 5):
            return "%.2F%s" % (float(bytes/math.pow(kilo, 4)), label[3])

Example

Here's a quick example of how this can be used in practice

#!/usr/bin/env python
# -*- coding: utf-8 -*-

"""Simple example command line usage:

Usage: ./check_disk_space.py <percentage free space>

e.g:

$ ./check_disk_space.py 30
Running low on disk space. 20.0% remaining (Warning Threshold: 30%)
Total: 115.13GB
Used:  92.35GB
Avail: 22.77GB

The idea would be to put this in a Cron Job and have it email you when
the free disk space is lower than "n" percent.

"""

import sys

sys.path.insert(0, '../')

from diskspace import DiskSpace

ds = DiskSpace()

perc_arg = len(sys.argv) > 1 and sys.argv[1] or None
percent_limit = perc_arg or 20

if not ds.has_free_space(percent_limit):
    print "Running low on disk space. %s%% remaining "\
       "(Warning Threshold: %s%%)" % (ds.percent_free(), percent_limit)
    print "Total: %s" % ds.humanize_bytes(ds.bytes_capacity(), 1000)
    print "Used:  %s" % ds.humanize_bytes(ds.bytes_used(), 1000)
    print "Avail: %s" % ds.humanize_bytes(ds.bytes_free(), 1000)

This last file can be saved as check_disk_space.py and made executable with:

chmod +x check_disk_space.py

All you then need to do is add a crontab entry and define the limit at which you want to be warned of impending disk space running out.

MAILTO=systems@yourdomain.com
0 */2 * * * /path/to/file/free_disk_space 15

This example will check the space every 2 hours and will email you when the remaining disk space drops below 15%.

A Branch of these files is available here: http://code.projectfondue.com/diskspace/trunk/files/head:/diskspace/

Suggestions for improvement are always welcome.

Loggerhead and mod_wsgi

2009-11-27T09:14:52Z

Currently loggerhead defaults to using pasteserver out of the box, which means you have to proxy this with Apache if you don't want to use pasteserver directly. Instead of doing that we went the route of setting up Loggerhead under apache with mod_wsgi.

Here's the wsgi file to set-up the application.

BZR_SERVE_DIR = 'THE_DIR_CONTAINING_YOUR_BRANCHES_HERE'

import sys
import os
sys.path.insert(0, os.path.dirname(__file__))

from paste.httpexceptions import HTTPExceptionHandler
from loggerhead.apps.transport import BranchesFromTransportRoot
from loggerhead.apps.error import ErrorHandlerApp
from loggerhead.config import LoggerheadConfig
from paste.deploy.config import PrefixMiddleware
from bzrlib.plugin import load_plugins

load_plugins()
config = LoggerheadConfig()
app = BranchesFromTransportRoot(BZR_SERVE_DIR, config)
app = PrefixMiddleware(app, prefix='/')
app = HTTPExceptionHandler(app)
application = ErrorHandlerApp(app)

I save this file as loggerhead.wsgi in the checkout of loggerhead.

Next all you need is a basic apache config and a WSGIAlias directive pointing at the wsgi file. Obviously it goes without saying this example is just for illustrative purposes.

<VirtualHost *:80>
    ServerName loggerhead

    WSGIScriptAlias / /PATH_TO_LOGGERHEAD/loggerhead.wsgi

    <Directory /PATH_TO_LOGGERHEAD>
        Order allow,deny
        Allow from all
    </Directory>

</VirtualHost>

Note

The following assumes you have a dns entry for whatever value you give ServerName - In my example I use loggerhead which I have set-up on my local machine to point to 127.0.0.1 so I have a local loggerhead running to view my checkouts.

Change the values of the paths and the ServerName as necessary
Save this file as loggerhead (or whatever you want to call it) in /etc/apache2/sites-available

Then all you need to do is enable it with

sudo a2ensite loggerhead
sudo /etc/init.d/apache2 force-reload

That's it!

You can view our code repository at http://code.projectfondue.com/ and a branch for these files is here: http://code.projectfondue.com/loggerhead-wsgi/trunk/files

Google Chrome Frame — good or bad for the web?

2009-10-02T20:35:38Z

Here is Google's description of what Chrome Frame is:

Google Chrome Frame is an early-stage open source plug-in that seamlessly brings Google Chrome's open web technologies and speedy JavaScript engine to Internet Explorer. With Google Chrome Frame, you can:

Start using open web technologies - like the HTML5 canvas tag - right away, even technologies that aren't yet supported in Internet Explorer 6, 7, or 8.

Take advantage of JavaScript performance improvements to make your apps faster and more responsive.

—http://code.google.com/chrome/chromeframe/

I'm intrigued by Google Frame; on one hand it's a pretty neat way of providing a more advanced experience to users of Internet Explorer, but I'm also concerned that it's providing a prop to browsers that really should be moving towards deprecation. If a browser isn't capable of displaying a site wouldn't it make more sense to point people in the direction of browsers that are more advanced?

Is this also saying that Google Wave will need Chrome Frame because it will only be meaningful to users with the most up to date technology. If you aren't up to date is there no base-line experience? If that's the case, a future that ignores progressive enhancement has a nasty aroma to it.

Update: 24th September 2009

According to this post on the Google Wave Blog, it seems IE 6,7 and 8 users will not be able to use Google Wave unless they install the Google Frame Plugin.

Google Frame and Accessibility

Early investigations by Steve Faulkner have shown that the inclusion of Google Frame locks out users of assisitive technologies completely.

Clearly this is an early release and assistive technology support could be improved in a future release. Whilst I'm happy to be proved wrong, I somehow doubt this will change given the epic technical challenges involved. That said, if anyone can do it, Google is a company that is best placed to make it happen.

Elsewhere

Apache: Moving from prefork to worker

2010-03-04T23:23:57Z

Differences between MPM Versions

So what are the differences between Apache MPM ("multi-processing module") types?

MPM Prefork doesn't use threads and instead spawns child processes to serve requests. MPM Worker on the other hand uses multi-threaded processes where each child process utilises a number of threads:

By using threads to serve requests, it is able to serve a large number of requests with less system resources than a process-based server. Yet it retains much of the stability of a process-based server by keeping multiple processes available, each with many threads.

—Apache 2 Documentation

Something to bear in mind is that you probably won't want to switch to worker if you are using PHP unless you are ok with using PHP as FastCGI and even then the PHP manual doesn't recommend using MPM worker.

For us that means we are now running our PHP apps using prefork and mod_php on a box dedicated to running PHP. This is better in a lot of ways because it means we can tune Apache for each specific application rather than running a mixture of PHP and Python apps on the same hardware.

Converting to mpm-worker

Actually carrying out the conversion on an Ubuntu box is as simple as running

sudo apt-get install apache2-mpm-worker

That will remove prefork if you have it installed and will install worker in its place and trigger a restart. That's all there is to it. (Though of course YMMV)

What Difference does it make?

In our case used memory levels dropped a reasonable amount as you can see in the last 6th of this graph:

CSS Sprite Generator v4.0 available on Launchpad

2009-09-01T11:03:28Z

This is useful for sprites which contain gradients that need to be repeated in a specific direction. There's also been some suggestion, particularly from the Yahoo! performance guys, that building sprites horizontally results in a smaller overall image size.

Arranging the images in the sprite horizontally as opposed to vertically usually results in a smaller file size.

—Yahoo! Exceptional Performance Team

Other Changes

Generated CSS

Displaying the generated CSS within a textarea field (rather than a scrolling div) to enable select all and easier copy and paste.

Component Spacing

A reduction in the default spacing (horizontal and vertical) between component images in the sprite. This makes the tool kinder to our servers and means, with a default generation, the browser has to handle a smaller uncompressed image in memory. This should be of particular benefit to the iphone which doesn't seem to handle large sprites too well.

"Be mobile-friendly" and don't leave big gaps between the images in a sprite. This doesn't affect the file size as much but requires less memory for the user agent to decompress the image into a pixel map. 100x100 image is 10 thousand pixels, where 1000x1000 is 1 million pixels

—Yahoo! Exceptional Performance Team

There's a balance to be had here, of course. The usual reason for including space between component images is to allow for text sizes increases which would otherwise result in partial display of the next image in the sprite. You can see how this would be a particular problem on something like the navigation of the Yahoo! home page.

Language Selection

Conversion of the language selection interface from a drop down box to a list of links to make it easier for someone to switch to their native language, particularly where they don't understand the currently displayed language. There's an interesting analysis of the use of drop downs versus links in the W3C i18n FAQ with the relative benefits and pitfalls of each approach.

Looking Forward

We've got a few other bugs that need fixing. We'll be working on those shortly and, in due course, we'll also be developing a command line version of the tool. This should make it easy to incorporate into an automated build script. Check back regularly for further details and do drop us a line if you've got any suggestions.

Get the Code

You can get the source code here on Launchpad. As always comments, suggestions, bugs and patches are welcomed.

Try Version 4 Online

CSS Sprite Generator version 4 is available online where you can dive right in and try out the latest features.

Raclette: Web App Utils for Python

2009-07-14T10:58:36Z

The goals for Raclette are to provide well tested middlewares and libraries that can be used to facilitate web development. We're also looking to keep the quality of code at a high standard as much as possible whilst balancing that with getting code out there so developers can try it out and provide feedback.

raclette.gbs: Graded Browser Support Middleware

The first app off the blocks is a WSGI middleware which provides a solution to implementing Yahoo's Graded Browser support strategy see http://developer.yahoo.com/yui/articles/gbs/ for more information.

Installation

sudo easy_install raclette.gbs

To use the middleware simply wrap the init of the middleware around your WSGI application

from raclette.gbs import GradedBrowser
...
application = GradedBrowser(application)

raclette.gbs works by analyzing the user-agent string and returning a grade stating that the browser is C, A or X grade. Yahoo's own strategy only provides details of which browser's are A grade so the C grade matrix is pretty much created based on guess-work around the notion that C-Grade browsers should be Antiquated and rare.

Ultimately for the purposes of this middleware only C-grade browsers really need to be identified as A and X grade are generally treated the same. However to be able to give the end user more scope to do different things with browser grades in development or production C and A grades are explicitly identified. It's up to the user to modify the browser matrix should they wish to add or drop support for any particular browser.

Known Issues

There's clearly some work to be done in terms of extending the data used to provide the decisions on which browsers are and aren't supported. With that in mind we're happy to get feedback on which browsers you think need to be indentified. Clearly it doesn't have to be something that's in Yahoo's matrix of support as it's clear that anyone using this may wish to define their own matrix of support.

Caching is not currently implemented. As the UA string is parsed on every request it's fairly crucial that caching is added - expect several caching solutions to be made available in subsequent releases.

Future Possibilities

Something that came up early on was an idea around extending support to include some notion of graded mobile browser support in the sense that you might have MA, MX and MC. At this point it's not clear if having any separate grades for mobile devices is a good thing or not, but certainly being able to detect common mobile browsers would be useful.

Getting the code

The code is available on launchpad.net at https://launchpad.net/raclette.gbs -- Contributions and patches gratefully received. All code is licensed under the modified BSD license.

L'Alpiniste

The Web Column: Issue No.5

The Web Column: Issue No.4

Contribute to the Web Column

The Web Column: Issue No.3

The Web Column: Issue No.2

The Web Column: Issue No.1

Security: Raw Roles in Docutils

Related Reading

Using Apache in a Python Development Environment

Building a Static Asset Cluster

Initial Thoughts

Setting up Heartbeat

Setting up HAProxy

Configuring Nginx

Mirroring Static Assets with Glusterfs

Related Links

Getting an Early Warning of Low Disk Space

Keeping an eye on Disk Space

Example

Loggerhead and mod_wsgi

Google Chrome Frame — good or bad for the web?

Google Frame and Accessibility

Elsewhere

Apache: Moving from prefork to worker

Differences between MPM Versions

Converting to mpm-worker

What Difference does it make?

Related Reading

CSS Sprite Generator v4.0 available on Launchpad

Other Changes

Generated CSS

Component Spacing

Language Selection

Looking Forward

Get the Code

Try Version 4 Online

Raclette: Web App Utils for Python

raclette.gbs: Graded Browser Support Middleware

Installation

Known Issues

Future Possibilities

Getting the code