[mapserver-dev] Fwd: [OSGeo-Discuss] Google Summer of Code 2008

[Steve Lime]

I'd co-mentor that one too since it's likely, at least in part, the GML output path.

Steve

>>> On 2/21/2008 at 1:11 PM, in message <47BDCCC9.5000006 at dmsolutions.ca>,
Yewondwossen Assefa <yassefa at dmsolutions.ca> wrote:

> If the KML output is chosen by a student, I would like to be able to 
> mentor or co-mentor the project.
> 
> Assefa
> 
> 
> 
> Daniel Morissette wrote:
>> Steve Lime wrote:
>>> FYI folks. Would be nice to participate this year. If others agree we 
>>> should start getting
>>> some ideas together. A couple of ideas off the top of my head:
>>>
>>>   - KML driver for output, draw and query
>>>   - implementation of Steve Woodbridge's label collision test idea
>>>
>> 
>> Yes, I think we should try to do a better job with GSoC this year. Last 
>> year if I remember correctly almost all participating projects got 2-3 
>> students, so it's a shame that MapServer wasn't there.
>> 
>> I went through our brainstorming thread from last year 
>> (http://www.nabble.com/Google-Summer-of-Code-Ideas-to9378841.html) and 
>> here are some of the ideas that we could potentially put on our list for 
>> this year:
>> 
>>   - KML driver
>>   - label drawing efficiency (building on Steve W.'s bit cache idea)
>>   - autorotation for polygon features
>>   - run-time joins for rendering and classification
>>   - develop a stable C API
>>   - optimizations for tiled maps, particularly around labeling
>>     and feature continuity.
>> 
>> We need to keep in mind that SoC projects need to be self contained and 
>> feasible by a student over the course of a summer. Don't expect students 
>> to work full time for 3 months on the project... my experience from last 
>> year is that they don't (or not all of them do anyway).
>> 
>> We'll also need mentors... I was mentor for one of the GDAL SoC projects 
>> last year and did a poor job due to my lack of availability (or lack of 
>> commitment?). If we want good results then we need to dedicate some time 
>> (1-2 hours per week?) or perhaps schedule bi-weekly meetings to stay in 
>> touch with the student and make sure he's doing okay and making some 
>> progress. I like the idea of 2 mentors per student that was in Wolf's 
>> email... a primary mentor and a backup for each project... that would 
>> definitely help.
>> 
>> Daniel
>