Hi.

Yesterday I tried Neon User Edition from USB dongle. Good progress !

One thing still is unclear for me. I tried to find answer myself but found nothing on neon.kde.org or other Neon resources (i.e., Neon on Google+).
Neon is shipped with "bare-bone-kde" SW only, application are being added along the development of Neon itself. As I understand, user is responsible to find and install all components himself/herself, no metapackages for "software suites" (i.e., Collaboration, Office, Hacking/Development, etc.) are provided. Am I right ?

The Neon Download page says, that Neon User Edition is "Ideal for everyday users". Most of such users will struggle to play his/her FLAC music collection "out-of-the-box" (sure, "sudo apt install flac" will fix that). Or this user will have problems not having a "groupware" for collaboration "out-of-the-box". Yes, it is ease to install Kontact, but how to get Instant messaging installed ? Where are Online Accounts' setting ? Neon will fail collaborate in Windows/Unix network "out-of-the-box" (users will have to learn that they need samba, smbclient, nfs-utils/nfs-common, cups).

I know, some of such bits and pieces area also missing in other distros. Yes, I know that Neon just started to deliver good and useful software to "everyday users". My question is "What is an Neon's everyday user ?" in terms of use cases. Admin ? Developer ? My mom ? Some of "everyday users" need pre-selected choices in order to use Neon everyday. Are there any plans for it in the future ?

Disclaimer - I DO support clean installations which are free of SW user don't need, but "clean" should not be the goal for itself.

Regards
ArtūrasB.

Did you really read the pages and try it? It even bundles VLC by default and a software center (Discover) for you to install new programs with a click. Also, all Ubuntu-based operating systems offer to add support for mp3 by default during the system installation.

Completely computer-illiterate users unable to find Discover won't install the system by themselves, and the person doing it for them will teach them about how to get new programs, and set up their network, if needed.

anewuser wrote:Did you really read the pages and try it? It even bundles VLC by default and a software center (Discover) for you to install new programs with a click.

Completely computer-illiterate users unable to find Discover won't install the system by themselves, and the person doing it for them will teach them about how to get new programs, and set up their network, if needed.

Yes, I did read and install what I need/want. In general, any user (with appropriate knowledge) can install any of SW

But my post is about what's available out of the box. Or what could be available as installation choices (not only "Install proprietary software" and "download updates while installing"). samba, cups, printer drivers (full set), free codecs and similar components won't take much disk space - why not include it with installation ? Or at least why not to give such choices during installation ?

Installation of almost any modern Linux distro is so easy these days, 15-30 minutes until reboot. But making the installation complete takes couple of hours more, for not so IT savvy users - even more, because they have to "assemble" quite common things themselves.

And again - it is not about me , it is about my mom. Or my 12 years old daughter. Or my old neighbor.

Since I install my machines using debootstrap and hand pick the applications, I highly regard important metapackages to avoid too many dependencies on software, that someone thinks I might need.

So, my sugestion to the maintainers would be to make a separate "Working environment" meta package providing all the dependencies of kubuntu-desktop and a Kubuntu base installation (Office, Players, Codecs and the other good stuff, you name it) and ask if you want this "Working environment" installed during installation. You wouldn't even need to include it on the installation media, it could be downloaded like other packages during setup, or afterwards using a pre-installed link.

That would accomodate both groups of users - people like me, who like a fresh installation clean and tidy, as well as those users, wo expect to get a system where most things are already done and decided for them during installation.

I would be interested in the opinions of jriddell, apachelogger and other maintainers about that idea.

Best regards,
Martin

Cymaphore wrote:Since I install my machines using debootstrap and hand pick the applications, I highly regard important metapackages to avoid too many dependencies on software, that someone thinks I might need.

So, my sugestion to the maintainers would be to make a separate "Working environment" meta package providing all the dependencies of kubuntu-desktop and a Kubuntu base installation (Office, Players, Codecs and the other good stuff, you name it) and ask if you want this "Working environment" installed during installation. You wouldn't even need to include it on the installation media, it could be downloaded like other packages during setup, or afterwards using a pre-installed link.

That would accomodate both groups of users - people like me, who like a fresh installation clean and tidy, as well as those users, wo expect to get a system where most things are already done and decided for them during installation.

I would be interested in the opinions of jriddell, apachelogger and other maintainers about that idea.

Best regards,
Martin

It is quite close to my vision. Maybe not a single "Working environment", but few options, i.e., "Developer", "Gamer", "Office rat", sorry "Office clerk" , "Multimedia".
And yes, it would be good to hear from Neon maintainers' opinion on this.

I started with a reasonably well setup Kubuntu-16.04 then added the Neon repos ,,,,,,and as stated on the Neon home page , it ate the puppies,,,,,,,,,,and spit out a pitbull

https://lh3.googleusercontent.com/-dwm5XquRmlI/V38OWPn2c0I/AAAAAAAACCU/shyR4QrMmjsLl9QQoui6_S5mJl1N5FXLgCCo/s912/snapshot81.png

I like it

VINNY

arturasb wrote:It is quite close to my vision. Maybe not a single "Working environment", but few options, i.e., "Developer", "Gamer", "Office rat", sorry "Office clerk" , "Multimedia".

This actually was also brought up at last year's Akademy and then forgotten about (or rather shelved) as it was somewhat grander in scope. And it is more of a design element than a development matter so I doubt my views on this will be of much use

There are, as I see it, three problems with this:
a) More of anything usually becomes too much very quickly. Specifically, by offering too many choices it becomes impossible to choose as multiple options might apply or none or some or some a bit. To prevent that from happening you have to draw a line. But where do you draw the line? How do you justify one set is more useful or better than the other? It's nigh impossible to do without then exposing this option to feature creep, which primarily comes from...
b) Just because someone is roughly doing what another group of people does, doesn't mean they use the same software. In fact, quite the contrary. In development the applications, tools, even documentation utilized is vastly different across languages and disciplines. I struggle to find any one thing every developer would want. Same goes with a Gamer. Is it a casual gamer? Or maybe someone who exclusively plays CS:GO? Or even a streamer? They have partial misalignment over what they want in software, heck not even every gamer would need a working opengl stack
c) All of this needs to be tested and maintained (i.e. software that is in such a set mustn't break as part of neon's work, which directly translates to developer effort down the line). FLOSS developers have a tendency to stretch themselves too thin, we'd like to not do that.

And once you have come up with something awesome, in the end problem b) is an indefinite fight against windmills so to not end up with bloated monstrosities that once were useful application sets.

This eventually lead us from "let's have different versions per group of users" to a very simple mantra for picking applications: "People are smart. They can open discover. They can type what they want in discover. They can click an install button.". Which is what drove the decision what to put on the User Edition.

SOOO
All that being said, it can be a good thing. It usually will be a marginally useful thing because of problem b). If one is vigilant about problem a) it won't grow out of control. If someone other than Jonathan or I does it, even c) is a non-concern.
If you are interested in looking into implementing this, I'd very much invite you to bring it up in the visual design forum viewforum.php?f=285 for some design input first and poke us on IRC to discuss tech details. This shouldn't be too hard to do on a technical side. I rather expect this to be a simple as having a package that provides a "fake" application one can find in discover.

apachelogger wrote:a) More of anything usually becomes too much very quickly.

When ever in the history of mankind did that happen?

apachelogger wrote:But where do you draw the line? How do you justify one set is more useful or better than the other? It's nigh impossible to do without then exposing this option to feature creep, which primarily comes from...

We differenciate between two groups of users here: Those, who want to cherrypick their applications and those who are basically happy if something is nicely preselected for them. Providing too many options and asking too complex questions will make it useless for the lesser group, while the first group will still do it by itself anyways. And that's where I think should be drawn the clear line.

One basic blank installation, as is,

And the option (as a simple question) to install by download a set of most common applications for everyday use resulting in a Kubuntu-like installation.

apachelogger wrote:"People are smart. They can open discover. They can type what they want in discover. They can click an install button."

...taking advantage of one of the most advanced package systems to date. You make some very good points here.

apachelogger wrote:a) More of anything usually becomes too much very quickly. Specifically, by offering too many

As I wrote earlier - predefined choices, 3-5 not more. Community&users could vote for metapackage's candidates. And of course, let's limit SW to what KDE and related projects can offer, i.e., digiKam (with all its dependencies + recommended packags) should be in "Photography", but GIMP should not - users will install it themselves separately. Sure, we want to avoid "Paradox of choice" (https://www.ted.com/talks/barry_schwartz_on_the_paradox_of_choice).

apachelogger wrote:But where do you draw the line? How do you justify one set is more useful or better than the other? It's nigh impossible to do without then exposing this option to feature creep, which primarily comes from...

Again - let' stick to KDE (or other selected/voted) Apps. Remember - primary goal of such packaging is to give users functional desktop since first moments of use. We cannot give the billion of choices. We want functional desktop to start working with.

apachelogger wrote:FLOSS developers have a tendency to stretch themselves too thin, we'd like to not do that.

This is common part of human nature

apachelogger wrote:And once you have come up with something awesome, in the end problem b) is an indefinite fight against windmills so to not end up with bloated monstrosities that once were useful application sets.

Look, these 3-5 "user packages" still leaves the most universal choice - "don't like whats served ? install bare-bone and add whats needed or pick the closest (to your needs) package and fine tune by adding/removing apps/libs". "User packages" would let lazy or non technical people start quick and adjust as it goes later.

apachelogger wrote:"People are smart. They can open discover. They can type what they want in discover. They can click an install button.".

I general I agree, but "the common sense is not so common" . As per my example - my mom, your grandpa, hard working neighbor, pupils in the school. They are smart in their areas and at specific level, which usually differs from group to group and among people in same group. For them it might be hard to understand why Amarok refuses to play their beloved FLAC collection.

Workaround for such people would be OS telling that something is missing and suggesting to fill the gap right away. I.e., an user browse his home network with fresh Neon install. He/she clicks on a NFS share and Neon suggests to install NFS support. Same with other not installed plugins, libs, addons and apps. In fact, this would be the preferred method of fulfilling gaps

apachelogger wrote: If you are interested in looking into implementing this, I'd very much invite you to bring it up in the visual design forum viewforum.php?f=285 for some design input first and poke us on IRC to discuss tech details. This shouldn't be too hard to do on a technical side. I rather expect this to be a simple as having a package that provides a "fake" application one can find in discover.

I have very (if not extremely) inflexible professional and personal calendar which makes my participation in IRC almost impossible. Said that, I could read documentation and then start looking into implementing some of those "user packages". Any link to start with ?

https://www.freedesktop.org/software/appstream/docs/ would be the primary tech involved I'd imagine. But as I said first take the idea to the visual design team for input.

apachelogger wrote:https://www.freedesktop.org/software/appstream/docs/ would be the primary tech involved I'd imagine.

Thanks.

apachelogger wrote:But as I said first take the idea to the visual design team for input.

Got it.