Sorry for the double post on this one (for those of you on both lists), but thought it should go to everyone involved here, as it greatly concerns our little "love triangle" between OBOS, Glass Elevator, and BeUnited. > I think that this post gives me room to finally ask > a question I have been wondering about.... There's probably a lot of folks that have been wondering these sorts of things. > The potential problem I see here is if apps are just > thrown into preference panels and shipped with the > system by default, there could easily end up a > situation that does not have a BeOS-polished look and > feel.... Keep that thought a tic... > BeOS (in my opinion) should ship with as many apps as > are available and useful, but, those apps should also > be focused on what they do (no competing apps for > themes bundled with the system, for example). Now, > perhaps if you want to put them in a 3rd party folder, > that is fine. Other than configuration applications, OpenBeOS is not about applications. It's about providing the ability for others to build applications. For a preferences panel for media or screen, yes, it should be provided by the OpenBeOS team. But for things such as theme'ing, the basis should be placed there for others to build their own themes applications (and it's associated limitations that are in place as well). If OpenBeOS developers had to worry about all the "little things", we'd never see a finished R2. However, I will consent that most of the "little things" are very likely to be done by the very same people as OpenBeOS members. Later, that will not, nor should not need to be, the case - provided OBOS is remotely successful. On a slight tangent, I'd also like to say that drivers should be limited in scope by the OpenBeOS team as well. This is not to say that it shouldn't be worried about, but I wouldn't want to see all the devs working on indefinites such as constant driver development. There's more important things to get done. A reference platform or three should be "supported" by OpenBeOS... more if time is available. More than that is a very good thing, but you all are limited in time and energy. Set a reference hardware platform (or a few more), and let others fill in the rest, unless, of course there's room to do more within OBOS, have at it. Done with that thought. > Just make sure that there is one and only one standard > for, in this case, theming, installed as the system > standard. There should be just the basis for theming in place. Distributors would be the ones implementing which applications go out with a distro. This is Open Source, there will be distributors, and if OpenBeOS is even remotely successful, there will be more than one, whether we like it or not. Which means things will be different per distribution. There will be that fragmentation. You cannot stop it with OSS. BeOS stayed the same because it was closed. MS controls windows because it is closed. Linus doesn't really have much to say about what goes into a RedHat or SuSe distribution, does he? The only thing we can do as a group to prevent this in a terribly bad way is to have an "official distribution" or certified program of sorts. This I have discussed with many people in various areas, and this is where we (at least I hope it's "we") see BeUnited fitting into the OBOS/GE/BeUnited movement. First, as a "standard setter" for things like the UI interface and theming guidelines. Second, as an "official distributor" of OpenBeOS, and possibly a certifier for applications to become "certified OBOS apps" or whatever. Any other distributor that wished to be "certified" with OBOS would only need to distribute with certified applications and meet the guidelines set forth at BeUnited (in the name of OBOS developers, obviously). I know, you might be thinking - conflict of interest, if BeUnited is the official distro, and then must certify other distributions. I suppose. But if BeUnited and OBOS are essentially the same group, and BeUnited controls the certification process, we can control the fragmenting of the OS into various camps of entirely different OSes all claiming to be OBOS. OBOS itself could not do this, as it would be branded a mockery of OSS, creating an OS that claims to be OpenSource, yet directing how it should look and feel and what should be the applications included as part of the OS (which, as I have already stated, shouldn't be done by the OS people, IMHO). But including applications is a necessity. BeUnited would "bundle" applications with the distribution, obviously. Among other things, I've been trying to also get ahold of some applications that are no longer being actively developed and take over development of those under BeUnited's name (but not necessarily as open source or free). BeUnited would also distribute the "most pure" distribution of OBOS... following exactly to specs and guidelines, and as close as possible to what the developers of OBOS envision. The OS would then be distributed through BUGs and others in order to keep the costs of distribution down. All monies received for the distro would be then paid to those that have "shares" in the distro (which is detailed and too long to discuss here, but it involves the developers and participants, not "shareholders" as the name suggests). > When you start putting items in the system folder, then > you start getting into an area where the app must be a > polished, finished product that represents the entire > system's way of doing things instead of just a 3rd > parties prototype of the idea. Here, you must be very > careful not to tread on how the OS should handle things > like preferences. That is only controlled by the people making the distribution. Thus, OBOS' name could be dragged down if anyone can do whatever they want. An official distro is needed, as well as a certification process. But unless it is a essential app to the operation of the OS, it should be a third party that does it. > If there is not one good, thinking, intuitive group > behind the decisions as to what ships with each version > of the OS, then we would eventually have 2 or 6 apps > that set "be themes" the way that 2 or 6 different > people want the interface to look like. That would be a > nightmare, in my opinion, to see apps just thrown into > the OpenBeOS download without thinking about how those > apps look with a newly installed BeOS system. 2 or 6 different apps on 3 or 12 different distros, with 33 different default settings. Success with OSS is a mess, if you ask me. > I guess my main point is to remember and realize the > difference between having a 3rd party folder included > with the OS and having already installed apps and > preference panels that represent BeOS as an entirety > as a system. Let's just make sure that when the time > comes for putting together the many versions of > OpenBeOS that are to come that it is just as > professional in look, feel, and functionality with > the apps that ship with it as it has ever been. You can only do that so far without the certification process and an official distro that acts as the "benchmark". > That way, OpenBeOS will not become a bloated, 40,000 > ways to do the same thing that breaks standards with > itself and becomes confusing as to which "theming" > application is the one to use as the system default. Exactly my point. If 4 of the 6 theming apps are approved by the OBOS team and "certified" by BeUnited, there's a much greater chance of maintaining control over the various distributions (if the OBOS devs don't think a theming app is "good enough" or think it "breaks too many rules to be stable" - why would a company use them in a distribution. Only then would "rogue" distributions be the bad eggs. > So, what do you all think? I hope that I've explained > what I believe accurately here. If not, I may try to > clarify. Does such a quality assurance group exist for > OpenBeOS that decides what / when will be included? > I'm sure the developers themselves will do a fine job > at most of this anyway, just wondering what others may > think. OpenBeOS really cannot decide that. It is the nature of OpenSource. It has to be an entity that certifies and creates these rules and guidelines. Sure, this could fully be done by the OBOS team, and have an "OBOS certified program". But then BeUnited brings in a certain "commercial" factor that OBOS cannot. Closed source apps could be distributed with the OS, CD's could be pressed and sold, groups of developers and users could help decide on guidelines collectively, people get paid (however small in the beginnings) for their contributions, etc. Does this give you a good idea of what myself, Helmar, Michael and many others have been discussing behind the scenes (not to be secretive, just to keep it where it belongs in order to keep things productive)? Hope so. :) If anyone has a problem or question or comments with all of this, we should take this matter up off list (or rather, on an appropriate list), as this mailing list is not exactly the proper place for it. ;) Deej