Some rights reserved by SteveGarfield

About 5 years ago I was asked to teach a 4th year undergrad software engineering course at the University of Toronto. The course had been previously cancelled due to low enrollment; in an era dubbed the “Software Gold Rush” a cancelled course indicated something was wrong…

Software engineering is difficult to teach
Students are expected to learn how to avoid mistakes they never made. A great divide results from the instructor talking about concepts suitable for a mature organization when students are all about working their ass off and getting things done the night before. We borrowed several lessons from startups, having been personally involved with two startups over a dozen years. The way startups work are much closer to students ways of doing things. Since launch, course enrolment has tripled and two Y Combinator applications have been submitted based on class projects. Here is what we have learned so far:

1. Use a startup software process
Students are all about getting things done the night before; similar to how startups work. Teaching a heavyweight process feels foreign because students haven’t made the mistakes to understand reasons for the overhead!

2. Change the project every year
There is nothing more of a turnoff than a make-work project with antiquated technology. Instructors that use the same project over and over are sleepwalking. A new project each year puts the instructor and students on equal footing, solving problems together. Make the class goal to have someone apply to Y Combinator. Discuss the non-technical issues of software such has how people are going to use the product, how are you going to sell it, what is the competition like, what is the business plan. One big class project brings issues into the classroom better resembling the real world. This also allows non-trivial projects to be developed and students to test-out roles (e.g. project management) that would not otherwise exist.

3. Allow controlled crashes
Let the students make mistakes. For example, let them avoid source control. A student who looses code because of clobbered checkins will be a lesson learned for the entire class. However, when crashes occur, it is the instructor’s responsibility to manage and fix it. After the mistakes have been made, teach them about process. Keep things light and give them references for their future travels. During lectures on process, tie them into the mistakes that were made. Make process real.

4. Demo early and often
Create a culture where the principal deliverable is working software rather than documentation. Use early demos to correct mistakes and give guidance rather than having them worry about their grades.

5. Instructors should code
The instructor-student relationship changes dramatically if the instructor contributes code. Everyone becomes a peer instantly. This improve communication and follows the startup philosophy that even managers should write code.

Next steps
The course has been well received by the students at UofT. I have much more regular contract with students from this class than the other courses I have taught at UofT and UofW. I am interested in hearing from anyone who is interested in providing continuity to the students; a partner that would provide input on the project at the beginning, stay involved with it during the course, and offer a path forward for interested students ready to commit to a startup.

I was doing some evening reading I came across and interesting  thought on the intersection of social networks and content:

Media is fast-becoming an important area for the social network. The news feed, long a place for friends to share personal photos and thoughts, is becoming more of a content discovery engine

I’ve posited elsewhere on the role of content in the evolution of the Internet. In a nutshell, if Web 2.0 is the social web, then I respectfully submit for your consideration that Web 3.0 is the content web. What I mean by this isn’t that the Web 3.0 is about taking traditional content that has been available in traditional media (literature in books; movies on film; music on CDs) and putting them on the web, obviously we are well into that trend. What I’m referring to is entirely new forms of content that have yet to be created for the social web

Then my thoughts drift to Canada and the potential behind Canadian start-ups. Canadians have always been innovators in content, even if that innovation didn’t happen at home. Note:

• Canada is becoming a leader in video games
• Canadians are funny
• NFB remains one of the preeminent film associations and archive (and is surprisingly innovative)
• Rush is still alive
• Canada has some of the best animation schools around

So does all this bode well for Canadian content companies? As we think about national competitive differentiators, aspects of a country’s economy that sets it apart from the rest of the world- should we be thinking content?

If Canadian start-ups could combine technology smarts and content smarts, could the result be a whole new class of content, the likes of which we haven’t seen yet?

I had an opportunity to catch up recently with John Green, cofounder of Savvica, and over an early morning breakfast got schooled on what’s new at LearnHub and JumboTests.

StartupNorth: What is JumboTests.com? What’s the relationship to LearnHub.com?

John Green: JumboTests is an extraction of core test preparation technology and content from LearnHub, and factored for a different audience and user acquisition model. Both sites are run by the same content, engineering, and community management teams here at Savvica Inc.

SN: So what is the change in target audience with JumboTests?

JG: LearnHub, although useful to English speakers everywhere, is targeted at the Indian student market. More than half of Savvica’s employees are in fact in our marketing team, which is based in Delhi, India.

JumboTests, on the other hand, is not focused on any particularly geography; it is equally useful to anyone studying for the standardized tests covered on the site (including GMAT, GRE, SAT, TOEFL, and others). This makes JumboTests especially relevant in the US and other Western countries which have the majority of the test takers every year.

SN: How are the user acquisition models different?

JG: Visitors to LearnHub mostly come through search engines. LearnHub has several hundred thousand pages indexed by Google and other search engines, and we rank at the top for hundreds of popular search terms. The site has a mix of user generated content and content made by our expert staff. For instance, LearnHub has the world’s largest free GMAT question bank, which has thousands of questions developed by us and our community. It is very popular.

Search engine traffic is less of a factor for JumboTests. Instead we are growing through strategic partnerships. Sites with existing, large, and relevant user bases (such as job sites, education sites, or portals) essentially embed JumboTests into their site using our partner platform. Our partners get hundreds of high quality practice tests that drive engagement, page views, and a split of the revenue.

Since the JumboTests launch 3 months ago, we have entered into long term relationships with 3 partners, all top in their categories: TalentEgg, India.com, and The Globe & Mail. All of these integrations are already deployed and online.

SN: So what’s next?

JG: We feel both properties have a bright future. Between the 2 sites, we help over 350,000 students a month with test prep, university applications, and career advice. That’s a lot, but there are many more students that aren’t on our sites but should be.

LearnHub is the largest education website in India, excluding reference sites like Wikipedia. But India has a long way to go in terms of Internet penetration. It has about 2x the Internet users as Canada, but it also has over 1 billion people. The number of Internet users in India doubled last year to 50-60 million. We are pegging our growth to outpace the Indian Internet penetration rate over the next 5 years.

JumboTests is tackling a more mature market. That is why we are growing it primarily through partnerships. There are a lot of established online channels whose audiences would benefit from our unique content and delivery technology.

When a critical server crashes in most businesses, an alert email is sent out, often to a group of people, in the hopes that someone will look into it. But who should handle the problem? What if they are unavailable? Or what if no one was checking their emails? A crashed site can result in lost revenue for a company but the problem of alerting the right people in the right way was traditionally only available to large enterprises.

PagerDuty is a new Toronto startup that aims to solve this problem through their newly launched web application at PagerDuty.com. Started by former Amazon employees, the newly launched (and currently free) beta service is a well-executed, easy to use to solution to a common and costly problem.

I love how it works. To use it, you sign up using a one-page form, and you’re then set up with your own account (accounts are free during their beta, but it’s not clear how long they’ll be in beta). You then set up your own alert software to send emails to PagerDuty whenever you want a notification (this is the most time-consuming part because you’ll need to set up each monitoring software individually). PagerDuty doesn’t do the actual monitoring; rather, it receives alert messages via email and routes them to the right person on your team based on rules you build. Alerts can be in the form of SMS, phone calls, or email.

One of my favourite parts of the software is the great UI on the built-in calendar that allows you to set up which person on your team is on-call at any time.

Here’s how PagerDuty’s Alex Solomon describes the service:

PagerDuty prevents a sysadmin’s worst nightmare: coming in to work on a Monday morning and finding out that their site has been down for the entire weekend. […]  When something goes wrong, PagerDuty springs into action by calling, SMSing, or emailing the engineer on duty.  If for some reason, PagerDuty can’t get ahold of the person on-call, it automatically escalates the alert to someone else.”

Everyone on the PagerDuty team spent some time working at Amazon which uses sophisticated internal tools for handling problem alerts. According to Solomon, they created PagerDuty because, after leaving Amazon, they “spent some time looking for something similar to what Amazon has, but the solutions we found were either lacking key features, or were priced well beyond what we were willing to pay.”

The team had the option to possibly start up PagerDuty somewhere in Silicon Valley, but in the end, they chose to start in Toronto because of reduced costs and the fact that they had a better network here. “the Valley wasn’t really an option for us. We chose Toronto because all of us grew up here — most of our network is based here, so it really made sense to try not to stray too far from home,” says Solomon.

PagerDuty has done many things right. The website looks slick and the UI is smooth. (Some have commented that the front page may look a lot like some other Web 2.0 startup websites but that’s probably a good way to start, in my opinion). The problem they deal with is a real one that their team has experienced in the past, and they’ve solved it elegantly. Some of the challenges I expect they’ll have is that the market will be limited to a specific segment of companies that already use server management utilities that send emails – marketing to the right people in these companies and gathering enough adoption will be a challenge at least in the beginning.

PagerDuty founding team is Alex Solomon, Andrew Milkas, and Baskar Puvanathasan.

You can learn more about PagerDuty’s service and sign up for a free account on their website at PagerDuty.com.

I’m going to try using the service at my company.

Lets say that you run a highly successful online community of blond vegans who have a penchant for Prosciutto. You would like to manage local reviews for that community, but you do not have the technical ability or the data you need to get it kick started.

Montreal based Praized is an innovative solution in the heavily contested local listings and review space. Praized is designed a white-label platform that integrates seamlessly with editorial content by using either an API or plug-ins that are compatible with SixApart?s MovableType and WordPress. Bloggers and site editors can embed snippets of merchant information within posts or news articles to drive traffic to their Praized-powered local section. Praized also designed its platform to be available to Facebook application developers and others through an API.

Praized communities enable users to search, discover and discuss places with like-minded people. Users benefit from discovering the ?long tail? of places via discussions on lesser known local merchants that struggle to be found through Web search. End-users also get real value from social tools that allow them to tag, comment, bookmark, share and vote on places that matter to them.

Praized is bringing a really novel and sensible approach to local listings. Realizing that you can bridge the gap between the hype-local and centralized business models can bring opportunities in a lot of markets, and that is what Praized is doing here.

The first Praized-Powered community is now active at Mocolocal and they also recently announced distribution agreements with Yellowbook in the US and Yellow Pages Group Co. in Canada.

Other Canadian local-search and review companies include iBegin, who have moved in to the data wholesale business, and ZipLocal (TSX:ZIP) who recently launched ZipDating. Praized does go beyond just reviews and listings, they also have a recommendation system that allows users to suggest places and things to friends in their social network, including Facebook.

Well.ca, Canada?s favorite online drugstore, has just landed what is the largest investment the Maple Leaf Angels have made to date. Angels participating in the round include: Colin Webster, former CEO of Truition a JLA Ventures backed company and Jim Estill, a Research In Motion board member and CEO of Synnex. Well.ca has come a long way in just one year.

I first met Ali Asaria, Well.ca’s Founder & CEO, at last year’s Mesh Conference. Ali was at the time equipped with a less memorable domain (CanadaHomeHealth.ca), a plan (Well.ca), and TONS of enthusiasm. Our conversation touched on financing a startup and Ali cheerfully admitted he hadn’t a clue how to do it… yet.

What they have created in one year is a sight to behold. The company is running smoothly, sales are growing briskly, and the first angel financing is secured. Well.ca has over 10,000 registered customers, ships thousands of products per month, and is seeing 35% growth month over month. Why do customers keep coming back? Just in time delivery of over 10,000 products sure helps, but I am convinced it is Well.ca’s cheerful attitude and super customer friendly service.

In our first days of operation, we literally hand-filled-out postage labels to ship to our customers, and, carrying towers of boxes, I walked them over to the local post office. It?s exactly one year later ? the mesh conference is about to take place again, and we?ve grown from 2 employees to what will be 9 by the end of this month. We no longer hand-write postage labels, but we still write hand-written notes to thank every one of the thousands of new customers that have bought from our site.

Jevon and I are brimming with pride. The angel financing marks a rite of passage for Ali as an entrepreneur and for Well.ca as a company. And we can hardly wait to keep on celebrating their many future successes. Ask Ali what the best part of the adventure has been and I guarantee he’ll have this to say: “The greatest part is the group of people we have working here.” Here’s a team that is doing well and having fun:

Lip Dub – Young Folks from ali on Vimeo.

Two months ago we received an email from TalentEgg’s founder, Lauren Friese, brimming with excitement about her latest venture into the world of online recruitment. TalentEgg, which just hatched, is a website that connects high quality Canadian employers with students and recent grads that are looking for meaningful work.

Job seekers using the site will be able to build TalentCards (Resumes), ask for advice using Grad Q&A (Forum), and read up on how to land a job on the TalentEgg Insider (Blog). Employers can create a free profile, but have to pay to advertise specific job openings, create awareness with site sponsorship, search through TalentCards, and send out targeted emails to job seekers. For the month of April the site is free on a trial basis to employers. TalentEgg hopes to cater to small and medium size employers who can’t make it to every campus recruitment day.

This is a tough market to crack, TorStar’s Workopolis dominates; even the venerable Monster has had trouble getting traction in Canada. And TalentEgg faces a classic chicken or the egg problem: job seekers are interested in sites with lots of jobs and rational employers will only pay to advertise once the site’s user base reaches a certain threshold. Despite all this, something tells me it is only a matter of time before Lauren Friese figures out how to make this site lay some golden eggs. Congrats on the launch!

OmniDate, based in Toronto, has been hard at work building a virtual date technology and it’s likely we’ll see their avatars coming to big name dating sites soon. We’ve all heard about sex in Second Life, the truth however is a lot less steamy; most people have a hard time getting Second Life installed and running on their machine.

Unlike Second Life, OmniDate intends for its avatar dating system to be used by real people who want to set up real life dates. Going on a virtual date is less time consuming, less expensive, and more secure than meeting at a local Starbucks. Once you’ve found someone you enjoy chatting with online, it is more likely you’ll enjoy meeting that person in the real world. My guess is that people will also use OmniDate to flirt, etc.

There are quite a few things to like about OmniDate’s approach:

• It is entirely flash based, this means there is no download, installation, or PhD required. This increases adoption.

• Rather than putting all their eggs in one basket and building a brand around virtual dating, OmniDate is starting out licensing its technology to established players with large audiences and strong brands. This increases their likelihood of success, sidesteps the cost of acquiring initial users, and removes the burden of building brands for each market segment.

Don’t think dating sites would be interested? Guess again. Technologies like this increase the entertainment value of dating sites (read: ad impressions), keep users subscribed longer (read: recurring revenue), and get users comfortable interacting with each other (read: higher conversion).

OmniDate is already working with some large dating sites who plan to use the virtual date technology on their sites. Is now the time for avatars to go mainstream? I can confidently say increased interaction makes sense and it is safe bet online dating will evolve past profiles to entertainment experiences. Give OmniDate’s recently launched demonstration site a try for yourself and leave your thoughts in the comments.

Taking off my rose colored glasses for a second, I think OmniDate has a few things it could improve:
- I read all about the challenges Pixar designers had with Ratatouille. Test audiences noticed if the color of lettuce wasn’t just right, Pixar spent an inordinate amount of time on the color green. Likewise 3D human avatars can go from cool to creepy very quickly. I was thrown off by the laugh an avatar makes when you type ‘LOL’, touching another person’s avatar is also a touchy matter. The avatar experience is still a little rough around the edges; that said, I think the team will iterate quickly and continue to improve the already good design.
- OmniDate has a room builder in the works, I would like to see an avatar builder also. You wouldn’t want your date thinking you are a super model would you…

OmniDate was founded by Igor Kotlyar, a serial entrepreneur who has already successfully built and sold a startup. OmniDate is a 6 person team and growing fast; they are interested in meeting with avatar designers and licensing partners.

Contact: Igor Kotlyar, Founder

By day, Anthony Carbone is an engineer at DuPont Canada. But from 6pm to 2am, he’s his own man. Not only does he travel back to his old university town, Guelph, to grow his property management firm, but he also moonlights as a web designer for hire with his partner Vinay Menon.

It’s in his soul. He has to be busy. “I?ve always been an entrepreneur, ever since high school; cutting lawns, doing landscaping, selling my time as a web programmer and developer back in university.”

Sure, there’s the extra money. But it’s more than that. Rattling around the back of his head are a huge number of unexpressed ideas that he feels compelled to act on. In fact, there is one idea in particular Anthony and Vinay have been driving towards since they met in undergrad.

Mad Whips

As Anthony tells it, “I met my partner outside of the engineering building at the University of Guelph in my second year and the topic was cars, money and the Internet. It was just at the time when everything was peaking and the Internet bubble was at its prime.” They decided since to moonlight as web designers to raise enough cash to launch their true passion, a car spotting online community called MadWhips.

Taking photos of whips? Well, I had to ask too…
Anthony: Obviously the ‘whips’ is referring to the new slang term for your ride.
Sunir: Is it really?
Anthony: Yeah, well, like your crib is your house, your whip is your pimped-out ride, right?
Sunir: I feel old now. Thank you very much.

Hitting the road

But isn’t moonlighting a problem for DuPont? Anthony says, “I have a really good relationship with my boss and he knows that I?m not really interested in going anywhere for the next two, three, four years and I still enjoy that corporate education that I?m getting by being at DuPont and interacting with all the different business units. That kind of corporate experience to me is more important than venturing off on my own right now.”

They’ve been striving to achieve their dream for years, working hard on the side. But it’s on the side, and their day jobs rule their schedule. The question Anthony left me with was: “When do we take on that certain level of clientele and when can we afford to say, ‘Okay, one of us can quit our jobs’?”

Contact: Anthony Carbone

This is part of a series of entrepreneur trading cards by Sunir Shah of FreshBooks.

If you are like me, your blog aggregator is getting a little out of hand. Once you start climbing over 150 feeds, and well in to the 200s, you are starting to get overloaded. I have, on a few occasions, deleted all the feeds from my feedreader and have started from scratch.

So far in it’s life, RSS has been kept pretty simple, and that has been a big reason for it’s success. Things are changing however. Every major browser now incorporates RSS in some way, and it is becoming more and more of a mainstream tool.

It is time for RSS to come of age, and to do that we have to get smarter about how we manage feeds. Right now, early adopters are up to having 200, 300, 400 or more feeds and the design of the aggregator hasn’t changed much in 3-4 years. when I ran Blogtrack.com almost 6 years ago, we were trying to create the aggregator. AideRSS is now reinventing how we use RSS feeds.

To help cut down on the noise coming in through your aggregator, the AideRSS guys have come up with what they are calling PostRank.

Postrank is a combination of how many links, mentions and conversations there are about a particular post. If you look at the screengrab you can see that AideRSS gets information about each post from places like Bloglines, Technorati, the blog itself (number of comments), and del.icio.us amongst others.

“PostRank? is a scoring system that we have developed to rank each article on relevance and reaction. It is a core part of the AideRSS engine that works to ensure that this digital assistant is helping you to tame the RSS beast and keep your news stream manageable.” – FAQ

The issue of currency vs. relevancy
The biggest tradeoff in moving from a normal all-you-can-eat feedreader to something like AideRSS that filters posts based on their popularity is that you are now relying on other people to participate to help you filter your posts. That is ok, and it works, but it also means that you aren’t going to be on to the latest meme right away. My solution is to put many of the less frequently updated and less interesting blogs in AideRSS while keeping a lot of my daily favorites in my regular RSS reader. Because you can import your AideRSS feeds into your aggregator, this is really easy. Cut down on the noisy junk and still get all your Valleywag and Scobleizer up to the second.

Will it Grow?(tm)
It’s easy to misunderstand RSS plays. Very few people really understand the RSS market, or the vision for how RSS will grow in the future. Even those who “get” and use RSS day to day have very little understanding of the business opportunity. I was not alone in wondering about Union Square’s investment in Feedburner until I started using Feedburner. Feedburner saw a real pain for publishers (understanding the use and reach of their RSS feed) and they delivered solutions for it incredibly well. AideRSS is doing the same, but they are bringing the same sort of value to both the publisher and the reader. We have added the AideRSS sidebar to Startupnorth, you can see it in the right-hand column.

Overnight hits such as mybloglog have shows that if you provide a few tools that are just interesting enough to both publishers and readers, then you can really hit a home-run.

One of my favorite things about using AideRSS so far is how snappy it is. My only complaint is that it creates some uncertainty about how often the feeds are being updated. I’d like to know the last time each feed was updated somehow, and have the ability to manually request that it be updated.

The core AideRSS services will always be free, with optional premium services available later on at a cost. I could see a service such as a customized newsletter for busy individuals (ie: “send me the top PostRank posts about the Real Estate industry once a day”). AideRSS will be the authority on what the most relevant content in the blogosphere is, and there will be many ways to capitalize on that.

AideRSS is a Waterloo, Ontario company, and they have taken a small amount of funding so far, but they are on the lookout for investors who understand their space, and what they want to accomplish.

For me, it’s an obvious one. Without trying to sound like too much of a cheerleader, I love AideRSS and I want them to succeed only so that I can keep using their service.

If you want an invitation to their beta, I suggest you ask in the comments below, I am sure they will let as many in as possible, and perhaps Rob can relax a bit now, help is on the way it seems.

Update: AideRSS has launched for public consumption, and Read/Write Web has a great rundown as well.

Contact Ilya Grigorik