My own career in technology has been inspired by several women in technology and engineering; my tech heroines include my mother, Linda Ziegler, my mother-in-law, Julia Russell, and two former work colleagues, the designer Kat Neville and the Java developer Magz McCarthy.
In some key ways, my own career path reflects my mother's. Linda started her undergraduate degree with the aim of becoming an historian, but as the limited opportunities in this area became more apparent to her, she determined to train as a chemist and worked as a manufacturing chemist in the plating industry. Similarly, I attained a PhD in British History and Economic and Social History and pursued an academic career for some years; at the end of my ESRC Post-Doctoral Fellowship, at which stage I hadn't successfully interviewed for one of the three positions in the nation in my field, I determined that I had to make a change and decided to do a 'conversion course', or a one year Masters in Information Technology. My mother's personal history of moving between a humanities and science discipline has always been part of my understanding of her, and was an important part of my inspiration to 'do what I had to do' in terms of getting the sort of training that would give me access to decent labour opportunities. I should, however, note here that my partner's example - he has a Bachelors in English Language Linguistics and a Masters in Cognitive Science and learned Java 'on the job' during the turn of the century Dot Com Boom - as well as his practical assistance as tutor was further integral to my attaining my MSc in IT with Distinction from Queen Mary, University of London.
Upon completing my MSc, I worked firstly as a Research Assistant in the Computer Science Department of Queen Mary and then as a Java web application developer at Xbridge/Simply Business and then its child company, BView, but it is only over the last year, working for myself building my own application that I've started to feel confident and passionate about working in Internet tech. It had been a difficult transition from a career path I had invested in so heavily and had been deeply passionate for to one I'd entered out of pragmatism.
My flourishing passion and joy for my work has been encouraged in no small way by Julia, Kat and Magz. Jules has a great love of gadgetry and Internet technology and stands out as a 'woman of a certain age' with such an appetite and ability to learn to use the latest tools and gizmos. She's been building web sites since the early days of the web and has recently been recognised as setting a gold standard of best practice for school websites (she works part time at her local primary school). Jules used Google Sites to create a user-friendly and professional site to communicate the school's happenings to the community. Indeed, Jules represents a quirky blend of 'good life hippy' and technologist and her home, nestled in the Scottish Highlands, is aptly called 'the electronic cottage'; much of her work is home-based and includes web administration for an international development organisation and desk top publishing for local authors. Her willingness to investigate new technologies and to understand and advocate the appropriate uses of different Internet medium such as the different uses of blogs versus forums for the benefit of less informed colleagues, is impressive and unique even without considering her demographic. The tenacity and fearlessness with which Julia approaches technology is entirely admirable and never ceases to amaze (and sometimes to shame) me.
At a different end of the women in IT spectrum are Kat and Magz, both of whom are in their twenties and so are my 'juniors'. Working with Kat and Magz helped to push me out of my 'funk'. I'd been indulging, without being fully aware of it, in a rather self-pitying attitude towards working as a technologist after splitting from my first love, history. These two women were full of fire, creativity, enthusiasm and tremendous smarts and helped me wake up to the fact that I was working in an exciting field, full of imaginative, clever and fun people. Kat's amazing adaptability - she was hired as a pure designer but has since developed an impressive set of UI coding skillz to complement her artistry - inspired and impressed the whole development team and taught me to enjoy the challenges of problem-solving, rather than getting so easily down and frustrated. My first memory of Magz was a story of a previous job where she'd worked some impressive Tomcat-clustering magic. I was immensely impressed by this tale and by the confidence, competency and intensity with which she approached every day programming problems. These two women were my first female-developer role models; while I was notoriously 'the first female developer' hired at Xbridge (this is how the CEO once introduced me to a potential hire, much to my horror) and though I worked with an absolutely fantastic group of guys, I was a bit alone in the wilderness until Kat and Magz arrived; they taught me to revel in my work and that, actually, Internet tech is a pretty kick ass place to be!
Tuesday, 24 March 2009
Monday, 23 March 2009
Pretty URLs in Google App Engine
Today I implemented so-called 'pretty urls' for my Google App Engine application. While my URL structure was reasonably clean before this work, I was inspired to make improvements after attending ThinkVisibility and particularly after talks by Rob Lee and Tom Smith. While it wasn't too painful to track down information on how I might achieve Pretty URLs in GAE, it did take a little trial and error to get running, so I thought it worth a post. Indeed, if readers have suggestions for a more elegant solution, I welcome comments.
What is a Pretty URL? A Pretty URL is a URL designed to be 'hackable', or human-readable. Ideally, users should be able to intuit a Pretty URL address, and likewise a Pretty URL might serve as a 'bread crumb' or 'sign post' indicating to the user where s/he is in the application. Compare, for example, a less than fetching URL (a) with a more pleasing example (b):
a) mysite.com/mycategory?myparam=an_entity_id
b) mysite.com/mycategory/an_entity_id
Example 'b' is just the sort of URL I desired; purdy, ain't she?
I chose to use my WSGIApplication's constructor to point Pretty URLs at appropriate Handlers:
The 'category' is associated with a particular Handler, while the '.*' regex matches the UID of the entity to be retrieved, which is passed on the URL. In the Handler's get function I retrieve the path instance variable include on the Request object, which returns a substring of the URL between the host name and query parameters. I then substring the path string in order to retrieve the key of the entity requested by the user:
This appears to be working well and as expected; I'm certainly blushing every time I lay eyes on one of my app's sweet little URLs.
(Readers may wish to view Google's doc on URL config; note if you try copying and pasting the syntax in their example you'll get exceptions, hence the trial and error aspect of my implementation.)
What is a Pretty URL? A Pretty URL is a URL designed to be 'hackable', or human-readable. Ideally, users should be able to intuit a Pretty URL address, and likewise a Pretty URL might serve as a 'bread crumb' or 'sign post' indicating to the user where s/he is in the application. Compare, for example, a less than fetching URL (a) with a more pleasing example (b):
a) mysite.com/mycategory?myparam=an_entity_id
b) mysite.com/mycategory/an_entity_id
Example 'b' is just the sort of URL I desired; purdy, ain't she?
I chose to use my WSGIApplication's constructor to point Pretty URLs at appropriate Handlers:
application = webapp.WSGIApplication([('/category/.*', CategoryHandler)], debug=True)The 'category' is associated with a particular Handler, while the '.*' regex matches the UID of the entity to be retrieved, which is passed on the URL. In the Handler's get function I retrieve the path instance variable include on the Request object, which returns a substring of the URL between the host name and query parameters. I then substring the path string in order to retrieve the key of the entity requested by the user:
path = str(self.request.path)
key = path[6:len(path)]
if key:
#Retrieve the entity and shoot back a responseThis appears to be working well and as expected; I'm certainly blushing every time I lay eyes on one of my app's sweet little URLs.
(Readers may wish to view Google's doc on URL config; note if you try copying and pasting the syntax in their example you'll get exceptions, hence the trial and error aspect of my implementation.)
Friday, 20 March 2009
Dynamically load YUI dependencies
I've been using YUI Grids and LayoutManager for the backbone of my app's interface since the end of last year. It's been a steep learning curve and I still consider myself very much novice and, indeed, only noticed this week the 'dynamic loading' tab on the YUI Configurator. Rather than statically including the required YUI CSS and JavaScript resources, it is possible to use YUILoader to dynamically import them on load. While I appreciate that YUI-experts will not be impressed by my YUILoader-epiphany, this approach has helped me to slim down my app's JS files while decreasing maintenance concerns and so I feel is worth mentioning for the benefit of other noobs.
As described on the Configurator page, it is possible to define the required resources in YUILoader's constructor and to tweak YUILoader's behaviour to suit your application. I found two constructor arguments particularly useful, 'base' and 'onSuccess':
YUILoader API documentation claims that the String value of base is the directory used by YUILoader when loading dependencies. However, I have noticed that no matter what directory location I pass as base, YUILoader fetches resources using the directory used to import YUILoader. That is, if I've imported src="http://yui.yahooapis.com/2.7.0/build/yuiloader/yuiloader-min.js" and set base to "http://yui.yahooapis.com/2.6.0/build/", YUILoader sucks down 2.7.0 versions of my required resources, whereas if I change the directory of the YUILoader import to 2.6.0 I get resources corresponding to this version. Perhaps I am missing something in how this all works, but I would expect to be able to override the YUILoader's build version by passing a value to base. On the upside, it is possible to control the resources gathered by YUILoader so, once you've gone to production, your app doesn't spontaneously upgrade with potentially dire consequences.
In contrast, I've found the onSuccess argument reliable for running the JavaScript used to render my app's display, i.e. to instantiate and run a LayoutManager. Previously I had included a script on every page to run LayoutManager, and have now been able to reduce the page-weight of my app by one file (before I've even started really polishing my frontend optimisation).
Finally, I've used YUILoader's insert() function to demand that only the JS, and not the CSS, dependencies are fetched:
I've overridden some of YUI's skin-sam CSS in my own CSS file which is imported after my third party resources. However, because the YUILoader inserts dependencies at the end of the page, my customisation of YUI's CSS was negated by a second and freshly imported batch of CSS dependencies. So, I've told YUILoader to just bring me the JS and I'll have to update the CSS version manually (unless of course some wise reader can suggest a more optimal solution).
Happy ui-coding.
As described on the Configurator page, it is possible to define the required resources in YUILoader's constructor and to tweak YUILoader's behaviour to suit your application. I found two constructor arguments particularly useful, 'base' and 'onSuccess':
var loader = new YAHOO.util.YUILoader({
base: "//YUI build dir",
require: ["dom","element","event","fonts","grids","layout","reset","selector"],
loadOptional: false,
combine: true,
filter: "MIN",
allowRollup: true,
onSuccess: function() {
//Stuff that needs doing on load, e.g. running LayoutManager
}
});
YUILoader API documentation claims that the String value of base is the directory used by YUILoader when loading dependencies. However, I have noticed that no matter what directory location I pass as base, YUILoader fetches resources using the directory used to import YUILoader. That is, if I've imported src="http://yui.yahooapis.com/2.7.0/build/yuiloader/yuiloader-min.js" and set base to "http://yui.yahooapis.com/2.6.0/build/", YUILoader sucks down 2.7.0 versions of my required resources, whereas if I change the directory of the YUILoader import to 2.6.0 I get resources corresponding to this version. Perhaps I am missing something in how this all works, but I would expect to be able to override the YUILoader's build version by passing a value to base. On the upside, it is possible to control the resources gathered by YUILoader so, once you've gone to production, your app doesn't spontaneously upgrade with potentially dire consequences.
In contrast, I've found the onSuccess argument reliable for running the JavaScript used to render my app's display, i.e. to instantiate and run a LayoutManager. Previously I had included a script on every page to run LayoutManager, and have now been able to reduce the page-weight of my app by one file (before I've even started really polishing my frontend optimisation).
Finally, I've used YUILoader's insert() function to demand that only the JS, and not the CSS, dependencies are fetched:
var options = [];
loader.insert(options, "js");I've overridden some of YUI's skin-sam CSS in my own CSS file which is imported after my third party resources. However, because the YUILoader inserts dependencies at the end of the page, my customisation of YUI's CSS was negated by a second and freshly imported batch of CSS dependencies. So, I've told YUILoader to just bring me the JS and I'll have to update the CSS version manually (unless of course some wise reader can suggest a more optimal solution).
Happy ui-coding.
Subscribe to:
Posts (Atom)