Thursday, March 12, 2009

Camptocamp at Toronto Code Sprint

Olivier Courtin and Thomas Bonfort were at the Toronto Code Sprint. I look foward to use the great developments made over there in MapFish.

The plot starts with a Toronto hotel on the banks of the Ontario lake booked from the 7th to the 10th of March, 4 whole days initiated by Paul Ramsey gathering the C tribe of Open Source GIS.

Paul successfuly managed to obtain sponsorship and to create the impulsion to unite more than 20 developpers essentially from the MapServer, PostGIS and GDAL communities, but also TinyOWS, libLAS and OpenLayers (a couple of developpers from closed-source MrSid were also present).

A big thanks to him and to Tom Kralidis for looking after the logistics of the events, and to the sponsors for financially participating to allow this to happen.

At first sounding North-American only, the event finally took an international accent with the presence of some european developers, Camptocamp having sent Thomas Bonfort and Olivier Courtin to take part in the sprint.

The event was “back to the roots” for the MapServer tribe, probably more resemblant to the Mapserver User Meetings, these having been superceeded in 2006 by the FOSS4G conferences where highly technical meetings between core developers are greatly reduced.

For the PostGIS tribe, this was trully a new thing, the major actual contributors having beforehand never met physically !

One of the major advantages of this kind of event is to allow a very rapid pace of development for each project, ensured by the physical (or IRC for some) presence of the major contributors who can straight away provide their expertise and visions, take decisions, and be sollicited on specific questions.

There was a strong interaction between members of each project, but also between members of different projects, learning on the way how to make their code best interact
with one another.

The second advantage of these meetings is that they help build up and keep a team dynamic, which is somewhat harder to obtain solely through online discussion and exchanges.

And the building of a dynamic for 20 people spanning a 4 day period was clearly addressed, beginning right at the first evening the day before the sprint at the local bar, a beer in one hand, a burger in the other, with geek Tshirts being the rallying apparel.

Day #1 was mainly passed on discussion and planification of the objectives and actions to take in the following days, as well the individual roadmaps for each project. This lively, noisy, and rich in propositions first day finished on a virile and sportive event, namely
a Hockey game between a Canadian and American team. For us europeans not used to icerings, the show was permanent, passing from intense game periods, to manly fights between players, to entertainment between the game periods (with a musical band, cheerleaders, and superhero contests)

The following days were much more “to the point”, alterning intense studious periods where only the keyboards clicking could be heard aside from the AC fan and the cans of pop being opened (the choice of said pops having been longly discussed on the sprint list beforehand, of course !), and animated technical discussions that necessitated or led to the intervention of the concerned gurus.

The daily ending call around 6PM was the sign that our physical needs should take the pace over our intellectual ones (in other words, that it was time for beer and food!).

Nevertheless, these moments weren’t solely fun and games, as many technical and fructuous discussions took place late into the night.

The last day (#4 for the initiated) was somewhat slower, as folks were getting tired and slowly leaving depending on the time of their flight back home.

All in all, it was a real pleasure to meet all these great people, and getting to share some well-spent time together. As you can imagine, the public was essentially masculin, Regina Obe of PostGIS fame being the sole feminine presence on this weekend of the international women’s day!

The theme of the session had been clearly set by Paul beforehand, namely “The Need for Speed”, as a beneficial emulation has existed since 2007 between the C and JAVA tribes on the grounds of pure performance.

And you have to know that for the C tribe, being caught up on the grounds of pure performance is in the very least a direct affront, so a proportionate response had to be taken! We do have some margin before we get to throw our gloves on the ground and start a fight like what we saw on the ice ring though!

This post gives the general tone of the event, we’ll elaborate in the next couple of posts on the more technical details of what was accomplished, namely on the performance front, the respect of OGC/ISO standards, and the creation of high quality and high resolution maps.

4 comments:

Anonymous said...

"I look foward to use the great developments made over there in MapFish."

Funny, I thought it was a C sprint. Is MapFish the new brand name for the whole OpenSource GeoStack ?

Anonymous said...
This comment has been removed by a blog administrator.
Tom Kralidis said...

Wouldn't the improvements made in MapServer, Proj.4, etc. be indirect improvements in MapFish anyway? I think everyone wins.

Cédric Moullet said...

Yes Tom, everyone wins ! MapFish can work with any OGC compliant cartographic server, like MapServer. MapFish needs some data and OGR /GDAL is often used to manage them and for their storage, one address: PostGIS. I could continue this long list ;-)