Revolution Systems Blog

Plat_Forms Results Announced

The results from the Plat_forms contest have finally been announced.  You can find a brief overview and the full results here .

While the Revolution Systems team did not come out the winner, overall I'm happy with how the contest turned out.  They seemed to have focused in on some areas that I didn't expect, but they definitely show everyone in a fair light.

As I believe I mentioned in a previous blog post, our two biggest mistakes were strategic and not really code related.  We should have opted to completely ignore the SOAP aspects of the application.  We knew that SOAP support in Perl is pretty poor going in and Phil even took some time to code up some libraries to make it a bit easier for us. 

The other mistake we made was to do "big bang" integration at the end.  While many developers loath big bang integration, in our day-to-day environment it rarely if ever causes us any trouble.  But leaving the integration until the end of a 30 hour work session when we were exhausted was a mistake.  I know of at least a couple of requirements we missed simply because we forgot to link them into the site properly.  Code was all there, we were just tired!

Plat_Forms Recap

So we survived Plat_Forms. Overall I think the contest was fun, interesting, and very well run.  But the real test of this is going to be how the results come out.

I think we fared well, compared to the other Perl teams, but we certainly did make some mistakes. We didn't prepare our VMWare server image as well as we should have.  We also took the approach of building major pieces of the functionality and then linking them all up towards the end.  In retrospect, we should have taken a more (ugh I hate saying this) Agile Programming approach and done more small iterations.  Normally this isn't at all an issue, but with the entire development happening in 30 hours it would have been a better approach.

What I really underestimated was how much of a role fatigue played in this contest. I guess I'm not as young as I used to be! We had a few big mistakes that can only by explained by being way too tired to be coding. If the contest had been three 10 hour sessions or even 2 15 hour sessions we would have done at least 50% better than we ultimately did. 

Progress at Plat_forms

Well the contest began roughly 21 hours ago and I think we are making pretty good progress.  Just got back from a quick 4 hour nap in the hotel, my team makes are more hard core and only took 3 hour naps. 

It's difficult to tell how we are stacking up to the other teams as there is very little interaction between the teams as we are all too busy building our applications.  Check out the contest blog for more regular updates from yesterday and today, including some photo's of all of the various team members. Luckily those photos were taken around 11pm last night so we don't look too tired!

Arrival in Germany

After an interesting flight where two people got sick over the Atlantic ( luckily there were 2 doctors and 4 nurses traveling on the flight ) and a ground crew who added too much oil to the engines we finally made it into Germany yesterday morning.

One of the cultural differences I forgot about when traveling in Europe is the lack of free wireless Internet access.  Turns out our hotel only really has wifi from 6pm to 9am for a reasonable price. We just finished registering at the Open Source Meets Business conference so I'm taking advantage of their free wifi to download the 3 metric tons of E-mail that is waiting for me.

Plat_forms Web Framework Contest

As some of you may have heard, along with some help from the German Perl Workshop Revolution Systems is sponsoring a team for the upcoming Plat_forms Web Development Platform Comparison.

I will be blogging about the trip and contest here and my team member Phil Crow will be also over at use.perl.org.

Personally, I think the results of this contest are going to be facinating. We'll finally have some in depth analysis of some of the more popular technologies in as close to a real world scenario as you can create artificially.

My only worry is that the results will be dismissed by the community at large due to the lack of Python, Ruby, and .NET being involved in the contest. I'm not sure, but I bet the fact that it is in Germany ( and in the middle of winter no less ) is part of it!

We're obviously a Perl team, but here is a list of the technologies we'll be using during the contest:

• mod_perl
• Gantry
• PostgreSQL