Our beta releases are as “just works” as we could make them. We are well aware of the limitations, that is why we call them beta, and also why advertising towards non-technical users is currently limited.
Making it a paid product is only going to make people expectations even higher, and make them entitled for a good level of support, which we can’t provide with the current small team and unfinished codebase.
programming is not wizardry. And knowing programming is not an excuse to ignore user interface design, documentation writing, user support, and so on, which are all just as valuable skills.
Saying otherwise is encouraging an elitism that isn’t acceptable. Programrers should learn some of these skills, or work with other people who do know some things about it.
You might be interested in this: I have a video scheduled for 24 October in which I took a default Haiku installation and systematically stripped out developer packages to see how much disk sppace one could save with a “Haiku Lite”. Python, Perl, OpenJDK, even GCC, all gone. The installation worked out about .5 GB smaller and it just keeps on running. Reboots with no problem.
That was an experiment, a learning opportunity, not a serious proposal. Still, it’s an interesting data point, I think.
You can simply uncheck all of these from the “optional packages” list in Installer when doing an install from the distributed images. In fact they are not activated in the live environment when you boot a freshly downloaded image (this reduces its merory usage a bit and allows installer to run on some more low-memory configurations).
You can maybe remove a few more things, but I think not so much, unless you also remove webpositive and the data translators
I know, but removing packages from an existing installation makes for better video. Will he accidentally delete something he didn’t mean to? Will he brick his installation? 
And removing WebPositive was never on the cards.
Seeing as you can reboot to an earlier state… 
I live dangerously*. After every step in this experiment, I ran Filwip to get an accurate disk usage reading.
*Well, not really, it was on a VM especially created for this.
Actually previous states can with time eat more space than few devel packages, especially if you have LibreOffice and Iceweasel installed. A simple to use limiter of some kind would be welcome for end user. Filwip is not a good option for this as it wipes the directory completely.
Huge apps with huge deps with frequent updates (or rebuilds) are problematic indeed. Though not sure what to do about that, maybe let haikudepot have an option to clean an app from all states?
Unfortunately, you can’t really do that because at the moment you’re choosing a previous state the network won’t be up so the missing package can’t be downloaded again. Anyway, depending of state age, it may not be able to find it either. Also if you are deleting states, they have to be contiguous. Deleting a state or a series of states in the middle is like removing steps in the middle of a stairway. Therefore, you have to remove a continuity of states starting by the oldest.
A good start would be to monitor space used by states and warn the user if that is exceeding a certain amount. Amount of space left on disk and place taken by installed packages could be taken in consideration. You don’t want states to take more than remaining space left and having states taking twice the place of installed packages seems ridiculous. Of course, the warning would have to be optional; some have huge disks and don’t care.
How hard would it be to have FilWip (or a new app based on it) wipe everything but the last two saved states?
Keeping last two blindly isn’t an option either. Install a package then aquarter later another, you end up with two contiguous states with nothing on it but an activated packages file. So, if uou have only that, it won’t make a difference at boot. You need at least one or two states containing haiku packages.
Only a small part of our comunity do package building with haikuporter (I guess), so sometimes when I build the full frameworks for KDE/Qt and KDE’s gear I get over 200 states in one day, mostly I clean up after that for everything older then one week (manually deleteing state and transactions directories) to keep the adminstrative directory clean enough for smooth startups).
So one must know how to handle the states.
Absolutely, however I fully recognize the brilliant work you are doing In engineering we say “Anything sufficiently complex is indistinguishable from magic (to the consumer)”. An engineer finding their work simple to them, would be a worry. I tried bits of programming and it is hard work to do well, so we must agree we are using “wizardry” differently, and since this is mainly developers here, I need to word it as “wizardry to the consumer” 
Come on PulkoMandy, accept what a brilliant job you are doing!
Maybe so, but for me, part of the job is making sure the concepts are well understood by other developers and also by users. So, in a way, “it looks like wizardry” is a failure at that, it means something is obscure, dfficult to understand, too different from what you’re used to. So it isn’t what I aim for
I see. So you’d have to have something more like Apple’s Time Machine, with all non-updated packages hardlinked, for that to work.
I don’t think we even have hard links, do we?
Well I am going to keep singing your praises anyway: for how brilliant you are stopping people listening to my praise LOL 
You are spot on as usual of course, and this is a very much a developers forum, so I am out of order and I am sure you are Virgo 
Agree to disagree, no developer here
Yes. Very interested Michel, though I think I wasn’t clear about my
Needs planning and streaming such users through “Non-developer oriented” streams, along with a “curated” release of Haiku that “just works”
idea. What I envisage is well outside the current Haiku development infrastructure with little impact on current development or developers, unless they engaged in doing it. (That is the different Streaming) In fact it would probably require close working with Mozilla or Brave or whatever fully featured Browser was chosen as THE basis, and a paradigm shift from the current developers. (That is part of the huge Planning) It is hard to explain without some people thinking it is being critical or saying things are being changed, but it only because Haiku seems to capable of it (from what I am actually running) that I am even imagining it. It is not some pie in the sky dream. It is could be done if people wanted to, and that is up to them.
Imagine the equivalent of a ChromeBook (now you can get cheap stuff on AliExpress and the like that make Chromebooks look very under-power so hardware and performance in not an issue (yes drivers are, but if built on specified hardware it would be selected to be compatible).
Unlike most developers here, most users do most of their stuff online in a web-browser or or web apps, and this is partly because personal software on the platforms costs and is a hassle for them. The array of existing apps already ported to Haiku would provide users the ability to pull some of their work back off the cloud (Google) and enjoy better speed and more control, without loosing any functionality.
While a mockup could be created with existing ported browsers (Both Iceweasle and Floorp look very promising until you try download ) working with the Browser developers would provide huge benefits, integration and performance gains. Given Mozilla is under pressure from Google and MSoft, it may be very keen on pushing its “own” platform to avoid the squeese skuldugery that Google and MS are engaging. There are other Web Browsers. If you view the Web Browser as an eco-system within an eco-system most of what people currently do can be done in the browser and the OS provide the platform. That is what I was meaning by a cut down system. By cut down I didn’t mean Haiku itself, I meant the package of software included, the main being a stable fully capable web browser plus a curated selection of software aimed at people who chat, watch online videos and take photos. That is the curated bit, curating the software to be shipped with the HaikuWebBook, not curating Haiku itself.
For that market the interface has to “match the market” and so the theming and desktop would need some developing for that streamed offering. While it is a dirty word on this forum, a sort of Distro. The difference is that if Haiku did the development it would be inhouse and operationally consistent with the existing mainline Haiku that exists as it is.
The finished HaikuWebBook Distro could be made available as a subscription, because that market expects to pay and it would have been worked on purely to support that market. If supplied on hardware it would be in the cost. Anyone wanting to access everything Haiku existing could do so as it is: no changes. Anyone with the HaikuWebBook could use it like a standard Haiku instal and load what they want from HaikuDepot (with no guarantee it will “just work” or special support. Any net profit would flow back in to support Haiku development (Your coffers).
Now there are many risks, problems and issues as has already been raised and will be further raised I am sure. Personally I think the software and even hardware can be resolved with some time and effort. The real risk is in the organization “growth” required to handle such a project, as that is where all the fighting and accusation starts. That is why it would need serious planning to fully implement and pull off. That was my planning requirement I identified (You guys seem very capable of planning development already), and I am talking hard nosed business plans and market research that no-one reading or writing this has much interest in it seems Why? Obviously you can easily lose trying this sort of thing, if it is not something the market will buy enough to at least cover costs.
Imagined time frame if everything fell into place and the project started by Christmas is three years minimum, but real developers with relevant experience would be able to post more accurate estimates I am sure.
However: If haiku Org/Inc doesn’t do it, others may. Google took Linux and made Andriod, Chrome book … Apple took BSD and made something 
. Mozilla could take Haiku and do it their way. Not likely maybe, but a possibility and opportunities arise from taking possibilities and making them work for you, if you can and if you want to. Maybe “others doing these things” is your preferred option.
My interest is simple and limited with the above being observation and speculation. Such a device would be good for my children and wife and seems like a cool thing to develop. De Googlising is all the rage, even though Google is my life line when I need to just do something and MS is more my long term perceived enemy of the people. I am hacking such a thing together for that purpose as we discuss, but that is a several orders of magnitude lower goal than the discussion above imagines.
Repeating that my ideas was separate and growth based to Haiku, Haiku org, Haiku Inc and everyone on this discussion, not trying to change anyone or anything, so please don’t respond as if this is a “hostile take-over” attempt. Your have a right to be concerned about how such a thing would effect the status quo, but it ain’t me who is doing anything other than what I said. Call the idea naive and ill conceived, because it is only a few days of musing, but please don’t flame war me.
Agree to disagree with your disagreement. Everyone (most everyone) here is trying to develop Haiku one way or anther, and I am comparing to the default Windows/Mac user. Those of us that are not wizards (said in a way that will not get PulkoMandy offside ) ourselves contribute what we can, but the deep insight is from the developers, or people who may not call themselves developers that, but in practice they are. Your name pops up often enough to make you one (who else would read this thread) .
The suggestion I just posted would change all the dynamics here, which is why I have said a separate streaming of support and everything would be needed.
The few total users on these forums are probably just enough for developers to cope with and give insight into what dumb things I will attempt