Sorry if this turns into a long one, but in the nature of the beast - a quasi-summary of what went before - it's unavoidable as all points have to be addressed, especially for any possible new readers. There certainly seems to be a fair amount of interest in this thread, looking at the viewing figures.
oOSGearOo wrote:
Nexter wrote:
Time has moved on, it really is time to look at different paradigms and metaphors.
My thinking would be an easily portable 64-bit app that provides the user with a simple, easy to use and fully configurable desktop environment/user interface that combines some of the elements of Winstep (but in a completely new form) and adds in some new elements.
We started out on the windows side, but end up talking on the linux side, but one thing I get from this is that the whatever it is 64-bit.
Indeed - what prompted this in the first place was thinking ahead and that Winstep Xtreme etc. may not have an unlimited lifespan and that perhaps a (eventual) replacement might be of interest here generally as well as specifically to Jorge.
We ended up talking Linux side because - esp. in the absence of any response from Jorge - both you Steve/oOSGearOo and I are pretty dedicated Linux users mostly, and because of the differences in code between Windows etc. and Linux - see further down for more on this.
oOSGearOo wrote:
So we want to go for some off Winstep's elements. I hope that's cool with Jorge? But in a new way/form, with some new elements. Like?
Well, it's not using any Winstep code so I don't see a problem there, and besides it's a very different thing. For more detail, I think it's better if we wait till I can edit/abbreviate/simplify/extract my architecture/design as mentioned before.
oOSGearOo wrote:
Nexter wrote:
By portable, I mean that it could be made available for Windows, Windows tablets, Android tablets, and perhaps Mac. This could easily be achieved by making it open source but with commercial/proprietary restrictions and bringing in volunteer developers to be paid through donations/sponsorship.
Here we go back to Windows/Android.
We wondered away from Windows again to Linux, I suppose to develop for both, is easy enough, in my minds eye.
If OS=winblows then
{win code}
Else Linux then........
The difference?
This is perhaps the main reason for talking Linux more than Windblows.

(Gosh, long time since last I've seen/heard that expression!) The difference is, under Linux we have to develop a complete new Desktop Environment, for a variety of reasons, whereas Windows of course has a completely different structure to Linux and does not have a separate DE - Explorer is both window manger and DE rolled into one (and more). Hence, we cannot put a DE as such on top of it, and we certainly don't want to f*** with trying to replace Explorer - some have tried (e.g., Litestep) and failed miserably, and that was before Windows 7-10 came along. So, it will be far easier to initially develop a Linux version and then when that is pretty much done think Windows. A lot of the DE-specific code will have to be dropped out and/or modified for a Windows version, there's no way around that, and some code added. E.g.,for a start the "app" will - like Winstep - have to "hide" the Explorer desktop as completely as possible. It's far easier this way than the other way round, that is, starting with a Windows version and then moving to Linux.
oOSGearOo wrote:
Nexter wrote:
I've not spoken of a Linux port yet, because for Linux, the core of this would need to be expanded into a full Linux desktop environment and preferably even provide its own window manager. And that would definitely require a team of dedicated open source developers. But then, things seem to increasingly be moving towards a hybrid open source - proprietary model, often funded through sponsorship and/or donations as well as sales.
Well we have to decide if this is a Linux/Windows/Android project or involve all.
I like the idea of being supported by donations and sponsorship, although when I think on it going down the sponsorship route I don't like the idea off inserting names into things, perhaps having a credits box or something.
Sales I don't know if we could that kinda goes against the grain, I suppose we could design the desktop(for want of a better word), then charge for a dock or some other wiz bang item on top.
Initially, as and for the reasons outlined above, it will have to be a Linux project as I see it. (Where it's also needed much more.)
Sponsorship doesn't mean having to splash sponsors' names all over the place - there would be a separate page on a web site listing all the sponsors etc. Just about all - except commercial, i.e., paid for - Linux distros depend on sponsorship to a greater or lesser extent, plus often also donations. Have a look at some of the various distros' web site to see what I mean.
As for sales kinda going against the grain, well, not entirely. There are an increasing number of paid for Linux apps and even distros, in addition to the corporate/enterprise flavours such as Red Hat Enterprise Linux (RHEL) and SUSE. (Pity RH - and Fedora - got there first with the name, otherwise I might have called this project after one of my own past or present hats!) All in all, I certainly wouldn't have any objection to going down the commercial route, but it may be wiser (esp. in order to attract developers) to stick to a free (of cost) Linux version for private users, but asking for a donation would be fine. Selling individual components additional to the basic thing wouldn't be feasible really, especially as everything will essentially be completely integrated.
oOSGearOo wrote:
Nexter wrote:
Ultimately, my ideas spring out of one of my kind of (old in origin, about 20 yrs ago, extended/revised a good 10 yrs ago) 'thought experiments' - envisioning a whole new OS that would be scalable to anything and offer the greatest possible security and privacy.
Projects like WHOnix and QubeOS show some aspects of this, to a limited extend.
The OS I envisioned would make extensive use of machine learning (I hate to use the term 'AI' because at present - and no doubt for a very long time to come - it is a complete misnomer, Artificial Stupidity would come closer, IMO) throughout, and the installer would automatically identify the processor type and other hardware, and initially install just the kernel and hypervisor with basic control centre and then - for servers, mainframes and desktops - give the option of installing the full OS, or any other OS/OSes, in a VM or VMs.
I have liked the idea of AI for a long time one that's coming to the for in some operations/styles although iam not in favor of inserting things just to make up the bulk of things just to make it look good. I think of searches I do with shopping and the results are never JUST what you asked for.
Well, like I said, the term 'AI' as used at present is a complete misnomer. machine (auto-)learning is more accurate to a limited degree, but Artificial Stupidity for the moment seems fine by me.

That said, good algorithms have their uses but in everyday computing are far too rare. The kinds used by search engines, shopping sites etc. are next to useless. You may be looking for/at a simple sauce pan and the damn thing wants to sell you a mobile BBQ!

And I agree with you, inserting things just to bulk things up to make it look good is certainly not desirable. No, the way I envisioned machine learning there is a very practical one - e.g., algorithms to observe the way things run/operations are carried out/messages passed etc. and to suggest how the code could be improved - made more efficient, leaner, etc. etc.
oOSGearOo wrote:
DesertDwarf wrote:
Well, first, I like the idea of a discussion board to flesh out our thoughts.I think a discussion about a new product could be really cool. Even if it doesn't come out as a new product, with Jorge peeking over our shoulders (or even adding his thoughts), there might be new ideas he could implement in Winstep.
I to would love to get Jorge involved how about it? Jorge?
We already have a new home (in mind Jorge sign you up?) Iam going over all the old stuff in the thread and trying to stir the pot. One thing tho DD(and others) get involved

plz.
Agreed. Yes, please guys, and anybody else, get involved. Your input would be most welcome.
oOSGearOo wrote:
Nexter wrote:
Would be nice if we could have Jorge in on this. Although, I don't think there would be much that could be implemented in Winstep from this - it's too complex and needs multithreading.
As I said before, the thing is I would see a Windows etc. version as a successor to Winstep Xtreme, so I really would love for Jorge to be involved. Even though the basic code would have to be open source (with a BSD-like or even wholly proprietary license), Jorge could make the Windows version proprietary and commercial, with just an acknowledgement of the open source code portions used, any copyrights etc., and inclusion of its license.
So, how about it Jorge, let's have your thoughts.
oOSGearOo wrote:
Nexter wrote:
However, before deciding on matters of distro etc., I think we should be clear about exactly what we are aiming for. To make this a bit easier and a bit more structured, I could extract a basic, abbreviated architecture design from my overall architecture once I'm well enough again. Perhaps at that point we could start a forum for this on your site and first do a summary of the discussion thus far.
Yes that is a good idea. I hope your well enough soon. If you included some graphics into it[different paradigms and metaphors] so much the better.
Yeah, shouldn't be too much longer. As for including some GFX, I obviously have a pretty clear idea what it would look like, so I'll try my best but it'll take additional time.