[isf-wifidog] User profiles in WifiDog

[bock at step.polymtl.ca]

> Benoit, I understand that  you prefer not to rely on some commercial
> fad, that will probably get replaced in a few months / years. But
> then the same reasoning can be applied to the Google Maps mashup I
> did last summer. As far as I know, I have not heard anyone complain
> that it was a terrible thing to use Google's API, that will most
> probably eventually feature ads and be replaced by the next best
> thing on the Web. No, I implemented it because we had dreamed of
> having a map for so long and that there was a readily available tool
> to do it. In the same manner, Google Maps is a service that users
> know how to use and like. The same idea applies to Flickr and
> del.icio.us for people that are interested in these services.

Indeed the same reasoning can be applied to google maps mashups, and it has,
frequently, including for wifidog.

We (especially Pascal) spent a lot of time in the early days playing with
mapserver for this use.  We even spec'd our server with that in mind. 
Unfortunately despite our contacts and the institutional support we had at the
time and the efforts we put into it, we never managed to secure GIS data that
would be acceptable for the common man, under a licence that couldn't be put
under our feets.  We supervised a project at concordia to try to give us the
tools we needed for coverage maps.  At WISFI in London we brainstormed to try
to find tome open tool to use gmap tiles inside a real GIS package so we could
do all the mapping vizualisation stuff we wanted to do (And just lose a layer
if for some reason gmap was no longer acceptable).  There were promising leads,
but apparently people's increasing reliance on gmap for simple needs has killed
the market and interest in those projects, which didn't make much headway since
(or so I'm told).  Every conference I go when I talk about it to GIS people, and
they are worried (haven't yet met one who told me that it's not an issue).

Obviously I can't comment on why you actually spent the time to implement it. 
However, the mashup is very different from flickr three important respects:
-Gmap was the only viable option at the time
-All the actual data in the database will not change at all should we finally
implement a real vizualisation solution, and is fully generic (and exchangeable
semantically).
-You actually contributed something usefull that works well, not endless
debates.

Off course I'm not blaming you for doing the mashup, that would be silly
(especially from the wifidog'S project point of view).  However it's undeniable
that ALL work on GIS stuff at ISF has stopped since we got the mashup.

> I hate when people argue like that, it just leaves a bad mood for
> some time around the project. The profiles will most certainly be

Me too, sorry you got caught in the crossfire.

> improved step by step. If users could simply add a link to their
> website, choose a nickname and add an avatar beside their nickname on
> the portal I would be extremely pleased for a while.

That's exactly what I had in mind, see where that got me?  It would ALREADY be
commited if it weren't for today.

> rush adding Flickr, del.icio.us or whatever. I will be really happy
> to have a look at these.

Unfortunately, it would seem that putting in any profile stuff in right now will
cause some people to demand that we immediately put flickr as a core field
(which would become very easy to do, and thus very hard for me to prevent
pending discussion of how to represent them in a generic way). Once there is
actual data in the database, transitioning to proper abstractions would become
criplingly difficult, and the onus of it would never fall on the person that
arranged to put in the problematic field.  (Not to mention that I don't want to
be the bad guy again)

I'm be willing to reconsider it is shown that there is broad political support
for actually coming up with a set of generic elements, and vetoing any elements
that do not yet have the proper abstractions.  Otherwise I'll wait till I can
commit something that is purely generic (no core elements) and flexible enough
that it doesn't risk sending us in months of demotivating online flamewars.

Good night,

See you tommorow.