Citát Jesse Ross <[EMAIL PROTECTED]>: > >> While that may be the case, is it laid out clearly on the front page of > >> the Étoilé website? I don't think it is, and that should be addressed, > >> since it's far more likely someone will be confused if there's no > >> explanation on our website, than in a seminar where it can be explained > >> to them immediately. I also think it should be said on the Étoilé > >> website that Étoilé is not an operating system, or even necessarily tied > >> to an operating system. It should probably even be referred to as a > >> desktop environment on front the webpage at least once, and then go on > >> to explain why it's unique/different. > >> > > > > To be hones, it was not very clear to me too about how should I put the > > Etoile > > environment in a larger context. I have made few diagrams of possible > > scenarios. Attached is a .pdf and OO document. > > > > Can someone correct the diagrams and add comments/questions there? > > > It's really a matter of "what are the goals of Etoile" and "who is its > audience"? >
Good question...However, I think that the answer is yet to be built....on this list. > Is our audience new users? If so, the we need to provide 4, the complete > solution -- probably via an installable LiveCD. > Why? I can imagine having "windows live CD" too. > Do we expect our audience to install from scratch or to have a system they > want to integrate Etoile into? If it's the former, we can do a lot more > with how the system works because we're not forced into making sure we > play nice with other environments. Installing from scratch also has a > higher barrier to entry -- look at Enlightenment and the effort it's had > to make to get KDE and GNOME users. > Yes, you are true, that we can do a lot more if we have our own system. However, Etoile OS would consume more resources than portable Etoile environment. Concerning the "nice play with others". I think, that we should start by implementing scenario 2, that is, ignore cooperation, focus on ourselves. Then we should compare features of the etoile world and the host world and create bridge where possible. > Additionally, are we experimenting, or are we making a "product" that we > will support? If the former, then we'll probably go with 2, as that's less > to maintain. If the latter, then we'll probably concentrate more on > integrating it with existing technologies and go with 3. > I would say both :) We are experimentig with creating a new product. Yes, 2 is liss to maintain, but who will use it? Only those who have everything inside Etoile. Can etoile provide everything standard user would expect from a desktop environment? Do not take me wrong, but in the near future not. > This is an open source project, so if we reach a critical enough mass or > provide something that no other environment provides, then 3 will happen > at some point eventually, because we can't provide _everything_ that a > user will need -- just look at the efforts to tie GTK into Cocoa. > I agree. > Personally, I'm most interested in 4 because I'm a control freak and like > the idea of a complete "packaged" solution. I would say though, that > instead of exposing Core OS Apps to the end user as shown in the diagram, > we would simply provide Etoile frontends for anything OS specific (like > configuration for things like networking, displays, etc). I like that it > would allow us to work closely with an OS team and make software that is > tightly and smoothly integrated into the underlying system. That's > something that you just don't get from most open source distros because > it's a lot to coordinate and maintain -- just look at the number of BSDs > vs the number of Linux distros. I'm also much more interested in the > simplicity of the OS X way of "here's a disk, pop it in and go" vs the > infinite complexity of the Linux way of "first pick a distro, install > that, then pick a desktop environment, install that, then pick..., then > hope it all works together". Granted, the multitude of options that Linux > provides means that people are better able to tailor the system to > themselves -- but what if you're a new user and you don't know what you > want yet? Some distros do manage better than others at providing > simplicity, but I think that's a niche that can still be explored. I think > there are users that want that and I think it would be really cool to > provide that for them. > 4 is nice, but I think it is an utopia. Well, at least at the moment. Why? Look at the number of possible operating systems + environments. There are some well known, sufficient to get work done and have support. Etoile OS would be "just another OS distribution". You still would have to do something between 3 and 4. Yes, I would like to have 4 too... Fact is, that we can not integrate Etoile into existing environments, as it has different nature. Therefore the Etoile would still be kind of a cell with boudaries. The etoile environment is a cell and the bridge is a cell membrane. See images here: http://images.google.com/images?hl=en&lr=&c2coff=1&q=cell+membrane Where the proteins in the membrane are gateways for information passing and data transfer (pasteboard, DO, notifications, events, ...). We should focus to create our own cell internals and present the Etoile to the outside world as something resembling "an application". We and advanced users will know, that the Etoile is an environment, however, as there are no applications (or at least there should not be) inside the Etoile, then it will look like an app with modules. Long time ago I was playing with Smalltalk MT. It is full featured smalltalk environment, but it feels differently from the outside world. Similar to other smalltalks. From the Windows point of view, it was an application, from my point of view, it was an environment. I will use my overused argument with Java. Any GNUstep based environment should be like Java environment with integrated environment manager (desktop, workspace manager, workplace, call it as you like). Any application/framework would be just a module for that environment. Therefore the evolution and focus should be like this: 1. stand-alone environment, rely on GNUstep + create very simple bridges (well, there are already some in GNUstep) 2. create more powerful bridge from the host to Etoile heading towards state where the Etoile is master environment and host is slave. 3. pick good combination of host and etoile, remove or replace parts of the host and blend it into single OS This way or that way, I see that both branches can exist: the guest environment and the stand-alone OS. The first one can attract more developers, I think. Now excuse me more biology, but I can not help myself ;-) Another argument for the host + guest is, that it is like mammal reproduction. Host is mother, Etoile is a child in her. Mother organs do the work for the child until the child is fully developed and can live by its self. User does not lack missing fuctionality as it is provided by the host OS until it is developed in the Etoile. Conclusion? Go for the host+guest, but keep in mind and slowly develop stand-alone OS. Regards, Stefan Urbanek -- http://stefan.agentfarms.net First they ignore you, then they laugh at you, then they fight you, then you win. - Mahatma Gandhi
