[Chicken-hackers] Re: Google maps implementation (was Map with Chicken users)

[Brandon J. Van Every]

Dan wrote:

Look if you don't understand why the Frappr website
looks like it does, 
the amount of work it takes to maintain such a site,
and why they're actually hiring people...
    

"Don't understand" ?...

The following is pretty much all that's needed:
http://www.google.com/apis/maps/documentation/async.html

The webapp must publish a suitable data.xml, in the
following format:

http://www.google.com/apis/maps/documentation/data.xml

That's it. Just copy the _javascript_ to the Chicken
users map page, and expose data.xml. Oh yeah, the
default zoom needs to be changed, and we need to get a
Google Maps API key.

"Don't understand" ?...
  

I think we're talking about 2 different things here.  You're talking about the technical underpinnings of an API.  It's probably not that tough, but it's also not Chicken.  I'm talking about commercial polish, what causes people to say, "Gee that looks good."  Frappr is a more effective looking website because it's got a bunch of people being paid $,$$$,$$$ to make it look good.  Alejandro got 80% of the job done just homebrewing it in Chicken.  To a suit, that may not be good enough, they might require the other 20% that costs all the $,$$$,$$$.  In fact they might say that Frappr is a relatively ugly website by commercial standards, and they need to get themselves in gear.

But to a *techie*, which is the core demographic that Chicken is currently capable of addressing, Alejandro has done a fantastic job.  I'm not losing sleep over whether techies will feel his work is "inadequate" in any way at all.  Techies want to see what they can do with a language.  Ok, they're not getting a demo that "it's painless to interface to other programming languages, such as the Google API" or whatever.  They *are* getting a demo that things can be accomplished in Chicken *from scratch*.  I still have ambitions of writing a 3D engine in Chicken.  That would "duplicate" lots of other open source engines out there.  But you know what?  It would be in Scheme, and architecturally I think that could have many advantages.  Particularly for scene graph traversal.

"Glue language" is just one way to look at development.  I've been primarily interested in core development, myself. 


Cheers,
Brandon Van Every