Re: online desktop APIs

[Havoc Pennington <hp redhat com>]

Hi,

BJörn Lindqvist wrote:
> I don't understand. What is Mugshot then? What is it not? To me, it is
> not clear what the intended audience is, if there is one.

I talked about this in http://log.ometer.com/2007-04.html#3
(scroll down to "Target Audience") I even have a "not" section about 
people I don't think would be interested.

> Then lets start small. For example, it would be useful and fun to have
> GNibbles upload your high score to the Mugshot server. Then you could
> compare your high score with other Gnibbles players. If that works well
> enough you could abstract it out into a general UploadYourHighscore
> API suitable for all high score tracking games, Gnometris, Mines etc.
>
> Then you'd also have to solve obvious security flaws like how to
> prevent people from cheating on their high score.

I think that's a good goal (though I don't know how to prevent high 
score cheating, short of DRM-like measures, if high scores are only 
shown for your circle of friends it probably doesn't matter)

> How about FreeCiv? Last I played it, it had big problems with server
> outages and to few players available to find a game. Maybe Mugshot
> could help with both finding players and be a stable server to play
> on? Or is that out of scope for Mugshot?

I don't think it's out of scope.

> You have basically answered every single question I have asked about
> Mugshot with "Sounds good, we will have that." :)

What I would say is "sounds good, we could have that" - I'm not saying 
that anyone I know of is implementing this stuff, just that I think they 
are worthwhile directions for design and/or implementation if people 
were interested.

> Are there some cool
> and "online" things Mugshot is not?

There are lots of things it is not right now, but the point is we aren't 
arbitrarily limiting what we could do or would support someone in doing.

Some things I'd say we don't want to do:
 - we don't want to replicate existing widely-used stuff, such as
   becoming a self-contained social network site, or becoming a photo
   organizer site, or adding our own online word processor; mugshot
   is more about being a "glue" or "meta" service to connect things
 - we don't care about features for certain audiences, two examples
   perhaps are enterprises and "unix die hards who only use terminals"
 - we don't want to do things that will bring our servers to their knees
 - we don't want to do things that suck from a design standpoint

> I think there is more to it than that. For example, in Sweden the most
> popular community site is www.lunarstorm.com. Naturally, to attract
> Swedish "community people" you would need to hook Mugshot up to
> that. For news feeds you'd like to subscribe to sites such as
> www.aftonbladet.se or www.idg.se instead of www.cnn.com.

That's why we want to support a long list of different existing services 
and just "glue" them, rather than trying to replace what people already use.

Ultimately to support more services, we'll need help from people who use 
those services and want them supported.

> There are a lot of those kind of problems if you want to attract
> "everyone."

Of course. So we start with some people and some stuff, and evolve it. 
Same as GNOME; I remember when GNOME was a gray rectangle that pretty 
much crashed on startup.</in my day>

> And that is only the technical part of it. For example, I
> don't think anyone could ever convince me to use MySpace no matter how good
> their site is technically.

The beauty is that you don't have to. We should avoid any path that 
demands that people switch what they use.

Havoc