discuss-gnustep
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Cacao Linux - GNUstep based Linux distribution


From: Sergii Stoian
Subject: Re: Cacao Linux - GNUstep based Linux distribution
Date: Wed, 24 Oct 2018 01:47:23 +0300



On Tue, Oct 23, 2018 at 2:30 PM Liam Proven <lproven@gmail.com> wrote:
On Mon, 22 Oct 2018 at 22:15, Sergii Stoian <stoyan255@gmail.com> wrote:
>
> GNUstep libraries, libdispatch, libobjc2 and  build toolchain. Non-GNUstep is my code, NeXTstep graphics (icons, login panel). I hope someday it will be refreshed with new graphics with the NeXT's style.

OK. Thanks for the answers so far.

So what I would like to ask is:
* can you itemise what you are _not_ using, and why?
* can you itemise any new components you're creating, and why?
You can find answers at github page, section "Why am I doing this?". Essential point is a "...look, feel, and design principles of NeXTSTEP ".
At past in this list I saw numerous discussions about "boring, gray, outdated GUI from 90th". I like it and don't want to waste my time for discussions.

* are your new components or code under the same license as GNUstep?
Yes.
If not, why not?

* would you be interested in working withing the GNUstep project and
passing your changes and new code upstream? If not, why not?
Patches to GNUstep libraries will be passed upstream. If may patches/changes for some reason will be unacceptable - I'll leave them in my project. Everybody will be happy. :)
Other code is my personal vision of ideal desktop. I'm writing it for my personal usage. If somebody will find it useful - fine.

> What do you mean "cosmetically"? Icons? Yes. Look and feel? Yes. But it is much more than just look - that's why I need integration with underlying system (look below my CentOS explanations) and spend a lot of time to fill Workspace and applications with uniform, solid user experience (window management, startup, shutdown, keyboard shortcuts, mouse preferences, screen and display management). More to come.

By "cosmetically" I mean that, for instance, GNUstep already has a
workspace manager, but AIUI you are not using it.
IMHO GWorkspace is the file manager not the Workspace Manager. For example, docking applications to WindowMaker Dock with drag and drop from File Viewer is not possible in GWorkspace by design. There are many such little features (window and application management, focus switches) which seriously affects the overall feeling/UX of desktop.

You say that you're using some of Window Maker but not all. Why not?
Are you passing changes upstream to Window Maker?
I've tried. But my patches do not bring new functionality to Window Maker itself. So I've stopped creating patches for WindowMaker. It works for me as it is.

> Not literally forking but current development is based on some outdated GNUstep libraries
Why?
Because changes in GNUstep code sometimes grabs my time for bug-hunting outside of my applications development.

> with some patches. Patches are aimed to be included in GNUstep later.
OK... but why later?
There are several reasons:
1. I'm not sure that I've implemented some functionality correctly enough.
2. It works for now and I have more interesting things to do right now to achieve my goals.

> The main reason is to have some stable base code and focus on applications and environment. Right now I'm the only developer and do not have time to align my code with changing GNUstep.
Is GNUstep really changing that fast?
I changes. And I have no serious reasons to stick with master/HEAD branch of GNUstep right now.

> After I reach the version 1.0 (it's really close now) I plan to return to GNUstep libraries (I have patches more ideas not implemented yet for GNUstep: screnn resolution handling, high DPI and so on).

OK. How do you intend to "return"?  Take a newer snapshot and work
from that? (This is how Ubuntu  works with respect to Debian.
Basically, every 6 months, they take a snapshot of Debian "sid" and
then work from that.)
It means update patches to new versions of GNUstep libraries. Test it and send it via pull request for discussion.

Is it part of your intention that your work all ends up in GNUstep?
If not, why not?
My intention quite clearly stated at github page. I hope changes to GNUstep libraries will be accepted. Other code - I don't think about it but why not?

I'm not trying to challenge you here but I do not really understand
what you are doing.
Thank you. I really do understand what am I doing for about 3 years.


You asked:

> I didn't quite catch what's Cacao Linux different from NEXTSPACE (https://github.com/trunkmaster/nextspace)?
> Do you plan to include some special frameworks, applications? Is it just Debian specific GNUstep build?

I would say that there are 2 main differences.

[1] Gregory wants to use plain code straight from GNUstep and make a
Debian-based distro from it. In contrast you seem to be doing your own
thing, which happens to be mainly based on GNUstep.

[2] Gregory wants a general-purpose distro as a showcase for GNUstep.
Not something that inherits from GNUstep and does more, something that
shows off what GNUstep is and can do.

I hope I understand his intentions correctly.
I understand. Gregory has already replied me.

Do you see the difference?

> Not that much. It's integrated with some system libraries and tools: Xorg (Xkb, XRandR), D-Bus, NetworkManager, PulseAudio, UDisks, UPower. It's not CentOS specific nowadays.

Aren't all those things in (for instance) Debian and openSUSE as well?
I don't know.

People have strong loyalties to their own distros.
Also, some distros are more current, or have better 3rd party support.
Often commercial applications has support for Ubuntu and RHEL/CentOS. I've chose CentOS.

Today, Ubuntu is more or less the default Linux for most people. There
is a lot of 3rd party software that only targets Ubuntu. It's a fairly
modern, fairly fast-moving distro with a regular update cycle which is
clear and well understood. It's also quite desktop-focus.
Ubuntu has it's own mature desktop. Why it should be replaced?

CentOS is a slow-moving, server-centric distro, owned by RH for
reasons that are not clear to me even as a former RH employee. The
"official" free RH distro is Fedora. CentOS is a sort of unsupported
branch of the much slower-moving RHEL.
And it is great. I have Ubuntu installed aside of CentOS on my laptop. Sometimes Ubuntu losts WiFi, CentOS - never! I like it.
I don't need to spend days trying to understand why my application behaves wrong after system package updates.

I personally don't like Fedora much and I don't like RH's package
management tools.

I prefer SUSE and Ubuntu's tools.

My first choice would be Ubuntu, or failing that Debian or Devuan. My
second choice would probably be openSUSE because they pay my bills.
:-)
I think the user shouldn't know what's inside your working environment: Cocoa or GNUstep, Linux, Mach or BSD.
As many users of NeXTSTEP just use it's applications without knowing version of the Mach kernel. ;)

> I want the NEXTSPACE will be Operating System for users and developers.

OK... So why CentOS then? Why a slow-moving server distro without a fixed clear update cycle?
Because it lets me do my work. And plenty of 3rd party applications works and supported here (Maya, Davinci Resolve, VirtualBox to name a few).
For example: I switch NEXTSPACE development from FreeBSD to CentOS 7. At that time CentOS already has systemd, Ubuntu has upstart. Now it's mainstream.
Firefox version is 60.2.2 ESR. What fast-movement do you need?
Personally, I really tired of fast movement from GNOME 2 to Unity, from Unity to GNOME 3 in Ubuntu.

> I hope so. And Gregory's kickstarter page quite confusing for that matter.

It is, yes, I agree. But remember the FOSS mantra: "release early, release often".
 You can install it right now: https://github.com/trunkmaster/nextspace/wiki/Build-and-install.

Better to get it out there than spend ages polishing it. Failing to do this is what killed Etoilé.
As for me, Etoilé is great playground for conceptual architectural masterpieces without really useful working desktop.
My intention more pragmatic: make comfort, fast and useful desktop environment even in the sake of code beauty. ;)
And I want it to be less featured by more polished in terms of UI/UX and stability. :)


--
Liam Proven - Profile: https://about.me/liamproven
Email: lproven@cix.co.uk - Google Mail/Hangouts/Plus: lproven@gmail.com
Twitter/Facebook/Flickr: lproven - Skype/LinkedIn: liamproven
UK: +44 7939-087884 - ČR (+ WhatsApp/Telegram/Signal): +420 702 829 053


--
Sergii Stoian, ProjectCenter maintainer

reply via email to

[Prev in Thread] Current Thread [Next in Thread]