If you have not already done so, please have a quick look at the slides:

Or you can watch a webcast of a live talk by Eric on the same subject, “How to Build a Lean Startup, step-by-step”:

Earlier this year, a Google Group was started.

I try to force myself to use Things on a daily basis. I had put the library files on a local server, but with “Back To My Mac” enabled, I had troubles accessing it over the net sometimes.

When I read about Dropbox, I reminded myself that I’m spending quite some money every year on Apple’s .mac (now Mobile Me) service, but 99% of the time I am just using the email part of it.

So, I wanted to see if I could put the Things library on the iDisk, despite the fact that their FAQ says it’s not possible.

It is possible.

1. Quit Things.
2. Enable “iDisk Sync” in the Mobile Me preferences pane.
3. Move your Things library (if you have not changed anything, it sits in ~/Library/Application Support/Cultured Code) to your iDisk.
4. Create a symlink from ~/Library/Application Support/Cultured Code to /Volumes/iDisk/Documents/Cultured Code.
5. Launch Things – if a request comes up, browse to ~/Library/Application Support/Cultured Code/Things and select it

Voila. Life can be easy sometimes.

Atlassian has a new promotion for the upcoming release of Jira 4:

If you send in a screenshot of your current issue tracker, you can either choose a free 10 user license (complete with 12 months support and upgrades, renewable for $10/year), or a 20% discount on your order.

Cool, huh? As a very satisfied Atlassian customer, I highly recommend to check it out!

I’m quite happy with Google Analytics so far, but the fact that you don’t own the data coming from your users on your website (along with some remarks from our lawyer about exporting usage data into the US being not the preferred solution, especially in Germany) was enough to start looking around for other solutions.

Piwik is an downloadable web analytics PHP/MySQL application, that aims to be an open source alternative for Google Analytics. The feature set already is impressive, and can be extended through plugins.

A JavaScript tracking code snippet must be included on all relevant web pages, to enable tracking.

Speaking of plugins – there does not seem to be a whole lot of community provided plugins yet, and the list of available plugins is a bit cumbersome to browse and search (partly because its served by trac).

There is no out-of-the-box support for goals, but you get some extra reports like “Visits by local user time” that have no equivalent in Google Analytics (as far as I know – I’m no GA geek).

The GUI is informative and pretty to look at – but I guess the big thing is the API. You can access the data – your data – in JSON, XML, PHP or CSV format.

As I have no plans for the API so far, I’m most interested in comparing the numbers I get from GA with the ones from Piwik. I will report my findings here soon.

In the meantime, here are some slides giving some more detailed info about Piwik’s structure and features:

That’s right, a free license, no strings attached – for all 3 Jira editions.

I have not looked at the fine print, but it seems you only get official Atlassian support (which is excellent, btw) for the starter license.

Still, the personal license should be enough to get you hooked…

And, related to my previous post, here is a demo explaining how to setup BlazeDS and Flex in IntelliJ IDEA, including debugging both the client and the server side – nice!

• Implementing a new system from scratch
• Linking to an external hosted system
• Integrating an existing 3rd party survey tool into the web application

The survey features were important, but would not justify a big amount of workload, so implementing a new system from scratch (that would give enough flexibilty to support different types of question types and complex ways of analyzing results) was out of the question.

The downside of this decision was that the survey administration will not be integrated into the rest of the administration interface, but we could live with that.

Linking to an external hosted system (like SurveyMonkey or SurveyGizmo) sounded nice, but I needed to be able to control access to the survey via tokens – so that a single user could answer a certain survey only once by using a token issued to him by me.

Most of the external hosted survey tools have no problem using tokens, but I needed to dynamically generate this token within my web application right before I send the user the invitation to answer the survey, so I needed to be able to insert new tokens at runtime from my application – and could not figure out how to do that using a hosted system.

By integrating an existing solution, I would be able to have direct access to the (local) database, inserting tokens on the fly.

What I ended up doing was using the open source LimeSurvey. It’s a standard PHP/mySQL application, and seems like the defacto standard solution for surveys on the web.

LimeSurvey can serve all sorts of question types, and – most importantly – use tokens to control access to restricted, non-anonymous surveys.

These tokens are stored in DB tables named tokens_<SURVEYID> whereas <SURVEYID> is (you guessed it) the numerical ID of the survey.

Now, if you construct the SQL update statement manually, it’s pretty easy to concat the survey ID into the statement string, but since I was already using Hibernate, I wanted to refrain from hardcoding SQL in my application.

After some research, I implemented a custom Hibernate Interceptor, which allowed me to insert the surveyID into the SQL-string on the fly without any other hassle.

I will write more about how we use and trigger the survey invitations in a later post.